Since a few people have asked how the program counter can be stored in the ROM. It was a special register physically on the ROM chip, not a read-only address in ROM itself. Ray's paper states it was resettable and steppable and could accept an address input. So, I'm 95% sure that allowed for full program flow control. @phirenz makes a good argument for why this may not be quite so simple though....and may have actually required additional ICs to actually implement. The SAAB Viggen is getting a lot of mentions in the comments. It had a highly advanced flight computer composed of many discrete ICs. Not a microprocessor but very impressive nontheless. I am considering doing a video about it in the future. The 'MiG-25' at 6:11 is actually a MiG-31. Thanks @AvArIeNmArKu4 An ‘assembler’ translates to machine code, not a compiler. Thanks @Mat-Ellis Swing wings are a subset of variable wing geometry. Variable wing geometries may be implemented in other configurations and the field is not 'obsolete' as I stated. Thanks @nobodynoone2500 31:48 bottom line: "U if P > L" should be "U if P > U". Thanks @MatthijsvanDuin (I'm really mad I got that one wrong!) 10:00 - Lateral Axis, not longitudinal axis. So, the aircraft is pitch stable. Thanks @UnitSe7en 45:00 - The hardware prototype ran at the same speed as the final product. And yes, I know higher aspect ratio doesn't equal more drag (it's the opposite for subsonic flight). I was trying to simplify and probably simplified a bit too much...I said I always screw up supersonic flight discussions.
[ retired Engineer ] Caught out on a technicality... When I bit-banged on 4004,8008/80/88 and 6502 a microprocessor _was_ a single chip, not a collection of six. One of the Engineers that worked on it wrote a book that was published in 2017! PS I'll still give the vid a thumb though, you made a half decent vid.
The F-14 could engage 6 targets simultaneously but it could track many more. The F-14 could track 24 targets and fire on any 6 with the half dozen Phoenix missiles it carried. Very interesting video I really enjoyed it and recommended it to a few friends.
The old chip designers were crazy. No simulations, no HDL, no DRC/LVS. They would cut photomasks by hand. Imagine seeing some bug in silicon and coming up with the concept of setup and hold time to explain it.
There's a great amount of info and real world experience in "Commodore: A Company on the Edge" detailing the process they used to design the 6502, hand cutting rubylith to form the masks for the different chip layers. Absolutely mind boggling how hardcore chip designers were back in the day. Even more impressive when they then go on to explain how their chips "mostly worked" or in some cases "worked 100%" on the first batch with those old processes.
Having flown the F-14A "Tomcat/Turkey" and then going on after my time in the Navy to work in the computer world... I can't count the number of times I had to bite my lip during conversations about CPUs and sit and wonder "when" until the Osborne, XT, Kaypro and Compaqs came online. We knew the processor was there however we really didn't get into in any depth. We were just dumb Aviators... NFOs. Coming from the A-4 "Skyhawk/Scooter", my initial instincts were to turn it off and control sweep manually. But, I came to trust it and in ACM school learned when to turn it off so as to fool the adversary as to your airspeed, turn it back on to get into position to get a splash. The real beauty of Bob Kress' overall designed was that he used this computer control and the airframe design to do something completely new... it didn't roll at the nose, but at the HUD. So when engaged in ACM, the plane rolled at your eyes, not your feet - Evolutionary. I never met Ray, but I did get to meet Bob in the mid-80s in Bethpage, NY. Was that a Data General Nova in your video (yellow/brown switches)? Worked on those too. I watched this with great interest.
It does kind of piss me off that bits and pieces of this architecture were not released to Intel, DEC, etc to jumpstart the microcomputing industry. They could have done that without revealing the specifics that made this so powerful in this application.
Does the F/A-18C do the same as in rolling at the eye level? I always wondered why it felt strange when playing it in DCS and rolling, and this would explain that if it does.
@@mikedavis7065 DEC already had processor kits consisting of several ICs at the time, at least PDP-8. Also, you don't need 20-bit CPU to perform 20-bit calculations, it's misconception.
The F-14 was the equivalent of the US Navy's current NGAD (next generation air dominance) program for 6th generation fighters. It was large, expensive, and had computing capabilities that people didn't realize were possible.
I admit I'm pretty ignorant of recent military computing, but I don't see how NGAD would be using leading edge computing technology as significantly as the F-14 was. At best they're probably using some Intel FPGA's in impressive ways. I'm not sure they have the hardware volume to create an ASIC on a leading edge node like Intel 4 (7nm)
@@1funnygameMy understanding is that the biggest advances in modern military computing are on the software and integration side rather than the hardware, which tends to be a few generations behind.
While mainframes at the time and later military microcomputers had capabilities, I'd be surprised if this system even had an MMU. I suspect the first integrated capability system remains the 601.
I entered the USAF in 1978 and was a aircraft mechanic on C5’S they had a top secret (or at minimum secret) system called “ MADAR” (Malfunction - Data - And - Recording) it basically diagnosed the plane in real time and when called up the mechanic on board could figure out what he needed to repair the aircraft it also could be downloaded to put in the aircraft records for future use. Pretty cool stuff for that day!
This is a Lockheed feature in general; it was also available in Civilian aircraft like the L-1011 Tristar. The milspec one was probably fancier but Lockheed was ahead of its time in this regard.
I wonder if that eventually evolved into the self-diagnosing FBW computers that are used today on commercial airliners (probably not, I expect they were developed independantly but it's a fun thought).
@@Alexander-the-ok MADAR itself didn't, but Lockheed was about 20 years ahead of the competition with it. The systems that came about in the late 80s/early 90s brought similar functionality to other manufacturers' aircraft. I don't know how much lineage current Lockheed aircraft systems have to MADAR but it wouldn't surprise me if there's still some DNA.
MADAR that was a hell of a system, until you have to reload the program from tape with the engines running and flight crew staring at you. Bubble memory lives!
I grew up around that airplane. It was the topic of conversation for decades because my old man was a test pilot who flew f-4's then moved to f-14s. He retired as the director of flight systems. For his masters he wrote a program that let them load and analyze flight data beyond the memory capacity of the old mainframes. Nasa was building the x-29 at the time and everyone in flight test needed more capabilities to analyze flight data.
I worked on the 4004 back in the day for Intel. It was not a stand alone chip either It required a clock chip and a couple level translators. It was a PMOS chip and did not work with standard logic. The IC's required were the 4001, 4002 and 4024 as I recall.
@@tankermottindI used to be into building desktops back in the day, and I remember needing to keep an eye on northbridge and southbridge temps when overclocking. Have we really made those obsolete in the 15 years since I last looked?
@@mrb692the northbridge is mostly integrated into the CPU nowadays, although the most recent innovation is having it be a separate chiplet in the package on a lower density manufacturing node (more nanometers) compared to the CPU cores. The memory interface part of it was integrated in Athlon 64/Opteron days and was one reason they dominated over their Intel contemporaries. The Southbridge mostly still exists, except the expansion cards and mass storage (which used to run PCI and IDE/SATA through the Southbridge) now connect direct to CPU PCI-Express lanes (optionally through discrete PCI-E switches to distribute that resource between occupied slots). Southbridge (or "chipset controller", "platform controller hub", usually connected over PCI-E instead of a dedicated bus) handles slower stuff, like networking, SATA, slower USB (the bit rates of USB 3.2 and 4 have reached the point that CPUs have started to integrate some USB controllers for lower latency access). Mind you, 4 lanes of PCIe 4.0 to the Southbridge is still close to 64 Gb/s, so it's no slouch... But USB 4 is hitting 40 and 80 Gb/s now.
@@mrb692 yeah its pretty much all integrated on the cpu now. Even memory controllers sit on the cpu die now. Its neccessary since data moves so fast now, cutting the "middle men" on the motherboard was a huge leap forward in modern computing.
Note on the glove vanes. They were designed to preserve agility at high supersonic speeds and would have had little effect at lower speeds. It was later found there was almost never a need for dogfighting at supersonic velocities; combat quickly became subsonic where the glove vanes gave no advantage As a result, glove vanes were disabled and/or deleted from later production runs to save weight and maintenance.
I remember it could perform 100k operations per second, which was unbelievable. I believe it could track 9 bogies and shoot at 6 simultaneously. Pretty crazy stuff back then.
do not be confused, the CADC is the flight computer, the ability to track and shoot 6 at the same time was due to another immensely complex part of the plane, the AN/AWG-9 radar.
Top Gun was popular because it was the first movie where the air to air scenes didn’t look like tacked on scenes filmed in the wrong aspect ratio. Everything flowed together, and it was a good yarn of a story.
for real, i cant think of any other film off hand that was remotely as seamless. if you were a fighter jet fan nothing else compared. the original was a bit homoerotic though, ive not seen the new one lmao
Imagine this is the exact field my undergraduate education was happened to be in, and they failed to give us the proper prospect of what we were actually learning over there breaks my heart, imagine a one hour presentation equal to this could have turned my entire understanding of my field around, and then... it was heartwarming to see even after 20 years of shelving my education i still fully understood all of this... thank you for the effort, both thorough research and admirable presentation... you sir, are a wonderful teacher...
@@_________________404 at our place computer science was divided into two undergraduate branches, hardware and software... i was a Hardware student about the end of last ice age...
It is sad that history often becomes corrupted into fiction. Important developments are forgotten and trivial things are elevated to legendary status. The F-14's multi-chip microprocessor is one of those important developments that has largely been forgotten. Thanks for posting this valuable piece of history.
@@thomaswilliam630 the other thing was the "hot" chick might be the most manly ugly hot chick of any big budget movie either. No wonder they cut her from the remake, it looks like she's been on testosterone for the last 40 years.
Professor Frank: "I predict that in 100 years, computers will be twice as powerful, 10,000 times larger and so expensive that only the five richest kings of Europe will own them.”
Fascinating! In the early 80s, I worked in Garret Airesearch's environmental test lab. I received my first Confidential clearance so that I could work on the SCADC program. I can't remember the F-14 system ever being mentioned back then, but that could be just due to oldish age. The SCADC was a standardized air data computer that was intended to replace older units on A-6, A-7, and A-4 aircraft . The same electronics components but packaged to fit in the existing electronics bays of older aircraft. There were four total configurations, the fourth being sort of universal fit. Seeing this video, I can see the similarities. Environmental testing was to MIL-STD-510...vibration, temperature, altitude, humidity, etc. Fun time for me working with centrifuge, explosive shock testing, acoustic shock, explosive atmosphere. I left GA shortly after completion of these tests, going on to work at McDonnel Douglas on C-17 (FCSS and Flight Test and Escape Systems (ACES II Advanced Recovery Sequencer, a compact digital computer itself).
In 1975 I was trying to work on the f-111, as a result I also did work on f-14s that came into the air base their systems are much alike because Grumman was a partner in the building of the f-111. You might be surprised to know that the F4 was originally designated the f-110 . As a result I did work on several f-14s and was familiar with their systems. Because it had to integrate a lot of mechanically driven actuators and control systems, it did require a lot and I do mean a lot of Maintenance time to keep it in the air. We complain today about the F-22 never being deployed with more than five planes at a time, and you are generally lucky if you can keep 50% of the F-22 flying at any given time. Much was the same about the f-14. It was an extremely difficult plane to keep flying. Especially in its first iteration. Iran was extremely short-sighted attempting to use that as their primary fighter. I'm sure that some mechanism of the are data computers were reduced in function for their application. However, I also worked with many Iranians. They were extremely good at understanding Electronics to the highest degree of the day. After having worked on a few of the f-14s I understand exactly why. When Carter cut our nation's throat and increase the instability in the Middle East by forcing the Shah out of Iran. It quickly rendered the vast majority of the Iranian Air Force, notably the f-14s out of existence until they can figure out how to get parts for them. They quickly learned how to build their own parts for those aircraft. That's why they still had several of them flying as of just a few years ago, I don't know if any are flying today. The earliest iteration of the F-14 was nothing to behold except when it was brand spanking new and everything worked on it properly. That was not always the case, as a matter of fact, it was rarely the case. I do not know if your remark about the f11 being difficult to fly while changing its wing sweep is true, it was never true on the f-111f model that I was tasked with. I also never had a problem flying the F-111 simulator, as I was the first one that the simulator Commander had ever seen fly it for the first time and not cause it to crash. I am not certain that we had chipsets in any of our flight data computers. We may have had them in the nav computer because even though it took the old school tape to load Mission parameters. The F model could be tasked with running a alternate routes to and from the bombing point, and never exceed 400 ft off the deck while doing so. You rarely would do a mission and approach a Target in excess of Mach 1 because you would leave a clear Sonic Trail. However you could do so quite easily. I never got into an F-14 simulator during my time in the Air Force because we didn't have any. But as an avionics Tech I was able to show them how to do things like engage terrain following radar while in articulated Terrain. That procedure leads me to the concept that we may have had some sort of microprocessor in the F-111. I cannot be certain, but it is possible. I also have a good feeling that given the restrictions that you had to mechanically put into your movement of the wing at times and other clues, that there was no chip on the flight data system, except in regards to navigation. I did see several iterations of the navigation computer in other devices, some still in the 1970s as a matter of fact. The last iteration of the F-14 was the most capable and had they upgraded it 2 a more powerful set of engines, something closer to what the f11f model, which had the same engines as the early F-16 and f-15, it may have been a much better aircraft. The unique issue with the high bypass turbofan hybrid engines, was that if you pulled back power on it too far and weren't paying close attention, you would scrub off speed at 50 knots per second. That is slowing down at a really fast rate and will you would get yourself in trouble, perhaps that's why you made the claim that the F-111 was hard to fly in transition. I never found such to be the case. You just had to remember to always run with a little bit more throttle than you thought was necessary. It was easy to adjust the speed of the f11 merely by sweeping the wings forward. With five stages of afterburner, you didn't even want to be on stage 1 while actually performing a mission. The f11 I would contend still to this day would be impervious to even the most modern anti-aircraft systems. There just isn't enough time to see it recognize it and shoot at it before it's already past you. Being able to Chase and shoot it down would be a Fool's errand. Especially if you were in any kind of articulated terrain, the ability to actually stay with it would be difficult in the extreme. Aircraft of North Vietnam even found the a model impossible to follow. That was however a much earlier time frame.
I don't know what happened there, ,and how. But the Soviet Union took possession of the F-111 back in the days of the Vietnam War. There is a story on this topic.
@goddoesnotexist5688 You can fly a jet in a sandstorm?! If I saw that in a movie I would be certain that was made up. I vaguely remember hearing about an airliner that had its engines fail because of a volcano spewing ash: en.wikipedia.org/wiki/British_Airways_Flight_009# Apparently there was another such incident: en.wikipedia.org/wiki/KLM_Flight_867 Not sure if volcanic ash is worse than sand though.
One thing on F-14 there is absolutely no plausible way for Iran to fabricate are the wing-box forgings. If USA had to remake those wingboxes it would be doable but take some time. They were forged on what was, & possibly still is, the largest forging press by size in existence. Simple drop-forging vs the complexity of hydraulic forging, its a multi storey forging press in Ohio that is one of a kind in the world, it is still used for extremely large & specialized forging applications. Majority of Soviet Unions heavy forging was done in Ukraine where all it's military shipbuilding & largest aircraft were designed & built. I would like to see drawings of the wing-box assembly, perhaps its pivot-pins used high-shock bushings & pins of forging-die alloys like A2 D2 O3 & Inconell, these are all very challenging to cut & press concentric into a forging without inducing stresses that could lead to "egging" or contact galling, bushings might delay the inevitable but everything gets work-hardened & microcracks over use time including a forged wingbox, I cannot imagine these Iranian F-14 wingboxes are still functional unless they are never flown, flown like a Piper Cub with no stresses on the wing-box, or possibly modified or replacements retrofitted as 1 piece to operate only in a fixed position
Worth noting that when transitioning to supersonic flight, the center of pressure moves back on every aircraft. This is not only due to the F-14's swing wings but the interaction of pressure with supersonic shocks.
What's so interesting to me is how the forefront of so many technologies came together in military applications in the cold war, and not just in the US. In Sweden Ericsson (of later cellphone fame) cut their teeth on the Saab Viggen's onboard computer.
Our government doesn't significantly invest in non-military technology sadly. A whole breakthrough in technology that could benefit society left to rot in secret in a useless fighter jet.
@@hueyiroquois3839 I just went to their website to check, and on the main splash page it's spelled twice, once as SAAB, then again as Saab. So clearly they don't care, and neither should you
I tought DATASAAB was a part of Saab before going independent and them at some point being swallowed by Ericsson when they wanted to grow from telecom into computers. Anyway. Data Saab. CK37. First flight computer with ICs. Very early 60’s. Also a 20-bit computer. Already a veteran by the time the F14 came out. Also lasted into the 2000s on the Viggen. Not microprocessor, but several years before, and in a small european country. Heck. The Brits make movies about how they discovered a way to crack the Enigma code, in Sweden, some guy broke it 2 years previously while riding his bike to work.
This content is much appreciated , this officially pushes my level of F-14 geek to that next level that my old ride or die will never surpass . Cheers M8t!!!
A 4 bit CPU is pretty limiting and that was the only thing they were getting in that package in 1971. They were mostly used in scientific calculators where they performed with mediocrity. But when my dad was doing his master's, those calculators kept him from having to go over to the mainframe and futz around with punch cards so much.
high aspect ratio doesn't create more drag, it actually creates less at subsonic speeds, which is why gliders have high AR. At supersonic speeds you have shocks and delta becomes favorable I really liked the video though, keep up the good work!
I think there's some subtlety to this. High aspect ratio wings *do* create more drag overall, as a simple matter of increasing the frontal area of the aircraft. What they excel at is improving the lift:drag ratio. All other things being equal, a subsonic aircraft with shorter wings will experience less drag, but it will *also* be less efficient and have less endurance than the same aircraft with much longer wings. If we imagine that we change nothing about the aircraft aside from the length of the wings (i.e. we keep the same engine, propeller, fuel, etc.), then the aircraft with shorter wings will be able to achieve a higher top speed, but the aircraft with longer wings will be more fuel-efficient and have a higher service ceiling. Compare the Lockheed U-2 and the Corvair F-106. They both use the same engine, a Pratt & Whitney J75, and they have comparable gross weights. The most obvious difference is that the F-106 employs a delta wing, and the U-2 employs a high-aspect wing. The F-106 has a top speed of 1325kt, while the U-2 has a top speed of 412kt. The F-106 has a L:D ratio of 12.1 at subsonic speeds, while the U-2 has a L:D ratio of 25.6. The F-106 has a service ceiling of 57,000 feet, while the U-2 has a service ceiling which is still classified but is *at least* 80,000 feet. The F-106 has a range of 500nm, while the U-2 has a range of over 6,000nm, a difference which cannot be adequately explained by the fact that the U-2 can carry ~10,000 more pounds of fuel.
@@tissuepaper9962 frontal area alone is misleading, especially at lower airspeeds. And you can massively increase Aspect Ratio of a given design, without changing the Frontal area at all, or even reduce the frontal area at the same time.
@@tissuepaper9962 in aircraft aerodynamics, we typically consider the projected or wetted surface area, not the frontal area. You can keep this area the same while increasing aspect ratio to reduce your drag.
@@tissuepaper9962 High aspect ratio wings only create more drag overall supersonic (and very high transonic). The reduction in induced drag when subsonic *more* than makes up for the increased frontal area, and thus a high aspect wing is lower drag when subsonic. Consider: a modern airliner at basically all phases of flight always is making lift equal to weight, so lift is fundamentally fixed. If you improve the lift to drag ratio, you aren't making more lift, because your lift still has to equal weight, so what are you doing? Decreasing drag. During cruise, there's fundamentally no difference between "improving lift to drag ratio" and "reducing drag", because lift is unchanging.
Yeah.. There were parts of that movie that were cringe. The flight nerds loved it because of the air action with the music. Every guy identified with going up to 30k feet with his homeboy and kicking that smug prick Iceman and his RIO's ass. - That's two "O"s in Goose, boys.
28:04 that Horner scheme rearrangement actually increases calculation precision and significantly decreases the number of multiplications required. It is still valid and truly useful algorithm till today. Can be even used to make manually evaluating polynomials much easier.
Some friendly corrections: (may border on the seriously pedantic, your call) 5:03 It has been known since the 1950's that the most highly manuverable craft are often somewhat inherently unstable. Nothing controversial to anyone knowlegable there. 7:38 Variable wing geometry is still in active development and is not obsolete. Swing wings are obsolete. There is a distinction with a difference. 10:18 IIRC, The computer controlled the reaction of the rest of the flight control surfaces to the extension and retraction more than the speed of actuation of the swingwings and vanes. It was required to control these so that the compensation worked correctly, as you noted earlier where manual actuation was still CADC controlled. 17:45 The stance that a "rEaL" microprocessor must have onboard rom, ram, IO, is farsical. My 6-month old top of the line pc has seperate IO chips, Ram chips, Rom chip, and a co-processor cars with all 4 onboard. So only SoC would count, which is not in the same realm. I think you just wasted time here. Just muddled the whole topic here, although explaing the rom etc was nice. 23:45 Not a correction, just think you should do a video on the "Four Phase" as I have never heard of it and can find little info. There are already many great and exhaustive videos on the 4004. 30:22 I don't get why that would be harder to program for, and you made no justification for this statement. 40:33 Machine code translates to binary with no real compilation, it's not a 'language' per se, but a 1:1 human readable formatting of binary commands. Which is to say they totally wrote it in assembly and that was just typed in as binary. Your image at 48:00 shows that in the context of the 4004. 45:30 Punch cards, and even magnetic tape were available then, but tape was the most common interchange format. At the time, data interchange between dissimalr systems was quite rare. 52:40 I know where there are more about in the USA, but for security reasons, I'll leave it at that. An enterprising hobbiest might be able to get their hands on at least a partial set. If I find some i'd invite der8auer to the states to see them and make a video (export restrictions invariably exist). I'm confident the iranians have already just used an FPGA to copy them, as most countries do with stolen digital tech (see: Russia and China using FPGA copies of us missle guidance computers, there are videos on this). 54:22 There were more. Some are still classified, which should make obvious their purpose. Others were just corporate dead-ends, although not all technologically so. All that said, you did an excelent job for a non-subject matter expert. Big respect. Got my sub.
Thanks, yes as you may have gathered by my tone, I do indeed think defining a microprocessor as essentially a SOC is indeed farcical: unfortunately there seem to be many that take that stance (for some reason). I didn't really enjoy having to be pedantic there but I was trying to make clear that 'if the 4004 is a microprocessor, then the MP944 is too'. Anything I say about coding in assembly/machine code should probably come with a disclaimer that I am about 40 years removed from that world. Btw, it was me saying supersonic aircraft are 'ugly' rather than 'unstable' which (rightly) made people mad!
The F-14 is not a inherently unstable aircraft, in fact, quite the opposite. Maybe if you compare it to something like a Cessna 172. The F-14 is especially stable to the point that even the famous flat spin is difficult to actually achieve and enter. Less than 10 airframes were lost in them. The F-14 actually is designed for stability because it lacks fly by wire or computer control over the primary maneuvering surfaces. As the wings sweep back, the center of lift moves aft, increasing the distance between the CG and the CoL. This was actually the reason for the glove vanes on the early F-14A's. They kept the CoL further forward, reduced mach tuck drag, and allowed for 7.5G maneuvering at any speed the F-14 could achieve. I'd argue that modern wing geometry hasn't made variable swept wings obsolete, but certainly has bridged the gap between them. A true delta will always have less drag than the hybrid delta of something like the F-22. A true straight wing will always have better leverage against the CG for maneuvering and provide more lift than a hybrid delta. The variable wing sweep allows for a greater optimization across the entire flight envelop. Aircraft like the F-14 had an impressive turn radius, turn rate, and speed relative to the thrust. I say that because the TF-30 equipped F-14A weighed significantly more than the F-15 did and had less thrust, and yet the top speed difference between them is not that great. The question is whether or not the weight and complexity or a variable wing sweep justify the performance gained over a hybrid delta, and that is where a variable wing sweep tends to fall behind. The other aspect is the difficulty in stealth. No one has made a modern variable wing sweep aircraft using modern low observability or composite technology. It may very well still present enough advantages, for the right aircraft, to justify its weight... but no one has really done it.
@@Whiskey11Gamingwrt swing wing, I think the real reason the Navy and Air Force don't use it anymore is because they can acheive "good enough" without having to maintain a swing wing system. Plus it adds weight, reducing fuel consumption. I think they found it just isn't worth the cost.
Formally, it's not wrong to refer to an assembler as a compiler. The definition of a compiler is basically just a translator for programs. An assembler qualifies, even if it's a lot simpler than we usually think of. As long as we're being pedantic. :)
@@Noisy_Cricket I concur, basically today they can beat the aerodynamic limitations with the force of sufficiently powerful contemporary engines to achieve acceptable results.
Hi! Still watching the video for the first time, wanted to thank you for your conscientiousness in using real captions instead of the auto-generated ones. That said, there's a few typos, maybe have someone read it over? Thanks!
I never thought I would casually watch an almost one-hour-long video about a 1970s processor. Loved the video! Really well explained and interestingly made
I found this entire video fascinating. Your explanations of definitions, concepts and use cases were excellent. My degree is in Digital Microprocessor and Telecommunications Technology, and included not just the basics of all electronics and simple digital circuits, but also integration into larger and larger I.Cs. Really took me back to those days of writing out machine code with pencil and paper, then hand-coding and old 6502 (8-bit) microprocessor in hexadecimal. And to think, they were handling 20 bits, 20 years earlier than that. Awesome stuff!
Well, they were handing many more bits even earlier than that, just not with microprocessors. The mid-60s CDC 6000 series were 60-bit machines. And the IBM 701, released in 1952, used 36-bit words.
I cant remember the last time I watched something this long, understanding next to nothing, still being captivated to the very end, even rewinding to rewatch parts I understood a little of. I absolutely loved this content! Instant subscriber. There is very little content on yt that deals with aeroplanes on this technical level. I would love to see what this channel would do with the subject of ”Centralkalkylator 37” the flightcomputer in the Saab Viggen. Big thanks!
17:50: As I recall, there are a number of "microprocessors" that are an aggregate of several chips. The 8080 "microprocessor" being one that comes to mind. Intel quite clearly refers to it as a "microprocessor" on their datasheets, but requires 8224 and 8228 to build a system around. Nomenclature wars anyone?
Generally, a microprocessor contains the ALU, registers, and the control logic to fetch instructions and send data along the correct path. I/O is usually included on modern microprocessors too, but generally interrupt controllers and such are not. A microcontroller is designed to be an all in one system containing ROM and RAM, along with all I/O. By putting these in one chip you can put computers in smaller devices at the cost of versatility of the system. There can be large debate as to which extensions would be required, such as floating point and multiplication for a microprocessor. The 8086 up to I think the 80286 required separate floating point co processors or FPUs such as the 8087. And x86 processors are what we still use today in most desktops. There were versions of the 386 I believe with and without the x87 floating point extensions. To say that we went from assembling computers transistor by transistor to microprocessors is very misleading as stated in the video. Computers before microprocessors were assembled by logic gates, as the 7400 series chips were released in the '60s. Then computers were assembled with more complex chips containing ROM, RAM, control, registers, and full ALUs in single chips. Just how we don't see the reason to combine our CPUs with SSDs and RAM now in our computers it wasn't very worth it to do it with the mainframe computers at the time. The technology jump from having separate control, registers, and ALU chips to one chip is smaller than most people make it out to be. There were much bigger jumps such as when the original 7400 chip was released, putting multiple transistors in one chip. There were also large technology leaps such as the PLA and other predecessors to the FPGA, allowing us to "program" a chip to behave like any logic chip we want to.
You are pretty much correct there. Actually, you could build simple microcomputer systems without the 8228 System Controller, though you lose easy access to certain signals such as whether you're accessing memory or I/O address space. (That removes a useful but not necessary feature: you can always use memory mapped I/O.) But yes, without the 8224 Clock Generator, you'd need other equivalent external circuitry to generate the unusual asymmetric ϕ1 and ϕ2 clocks that the 8080 requires.
I was in the Army and was a mechanic for the 1980's light tank and I was impressed by the components, open one up and it was filled with multiple universal cards, so we could essentially lose a few cards(around 4 if I remember correctly) before the whole box needed replacing.
You did an excellent job describing this fascinating element of the F-14 I have never seen investigated in detail like this. I was a little kid when the top gun movie came out and was absolutely enthralled with the F-14 because of that film. Approaching 50 now and I completely forget at times how far the hardware for computing has changed and how revolutionary an aircraft with chips like these were. I started in college with electrical engineering but changed to a degree in information systems, but your talk about this brought flash backs of my coming to the conclusion I was never going to be an electrical engineer 😂. Really nice video on an interesting subject.
14:00 an FPGA is not a microprocessor - it's a collection of logic that can be configured for different arbitary logic functions, possibly including one or more microprocessors
I think he was pointing out how loose the term was back and any sort of chip that could do arbitrary logic whether that be in hardware or software could have been considered a micro processor by old parlance
This is very interesting. I was a structural design engineer on the F-14 program for a few years in the early 1980s. Even as I contributed my small part, it felt like i was chipping rocks compared to the stuff the guys in other parts of our building, and in other buildings, were doing.
Still, it's a long way from actually chipping rocks! That's what civil engineers do. (Mechanical engineers build weapons; civil engineers build targets.)
Excellent video. I've greatly enjoyed these long-form deep dives into engineering relating to aerospace. As an aside, 2024 is supposedly the last year for the F-14. Iran is retiring them in favour of Su-35s (or a similar Flanker; I forget exactly).
This is the best Ive found so far for computer tech back in the 60s, 70s, 80s. The low level programing needed to control swing wing on the F14 Tomcat.
They started teaching computer architecture in A school in the 80s. They had Altair looking computers where you would actually hand enter opcodes by hitting switches (they were Octal - so base 8) and pressing a button to enter the value into a register or memory. You had to basically hand enter the program one opcode after another. The press another button to execute the program and you crossed your fingers it would end up in the correct configuration after running your program :D. I, of course, had already taught myself Assembler using DOS 3.2 and the DEBUG program when I was 14, so that unit was a breeze for me. A lot of people struggled with that unit, but I was lucky. Ironically I taught myself Assembler because I wanted to get closer to the "metal". When I discovered PCs I was in an electronics phase, so for me it was the EE part that was interesting. I eventually became a full time programmer / software engineer :D
@6:26 Being from that era and from the US, I can speak to why it was so popular: it was so well received because, in addition to being just a spectacularly performed/shot/edited/etc. movie, it also served as a desperately needed reassurance 5 very long years before the Soviet Union would collapse. Every single day that the USSR was in existence came the renewed threat of nuclear annihilation. Many people very reasonably assumed that the only way we'd know the winner of WWIII would be for it to play out, which of course meant that very few of us would be alive to witness it, and many of those would find it very difficult to want to go on. Imagine the threat of being forced to live the IRL version of a Fallout game, where all the enemies are invisible, and the majority of them eat you by slowly and torturously melting you from the inside out. That was our day-to-day. Every day we knew that the sirens might come up, and we'd have very little time to say our goodbyes let alone make sure we could all be together with our loved ones at the end. Top Gun showed another way. It showed that we could fight the war in miniature, and if the USSR wanted anything worth conquering, it would have to take it the old fashioned way. And this was not without its merits. You can't grow crops on irradiated soil, and when Top Gun came out, only a month or so prior, Chernobyl had given the world but especially the Soviets a hard look at what the consequences of a full scale conflict would be. And Completely unbeknownst to all of us, the military technological capabilities of the Soviet union had been dropping behind that of the US for several years, and the USSR had been working very hard to hide that fact. When that very brave Russian officer defected with his aircraft on September 6, 1976 to the Western allies, the United States was able to dissect the MiG and learn its weaknesses. From that came the F14, and as far as I know, by the time top gun came out, the Russians knew that the movie represented no idle threat. Where the communists had kicked our asses in Vietnam and Korea due to our lack of dog fighting skills, the F14 was specifically designed to be able to get to conflict very quickly using the minimal amount of fuel, and then have the aerodynamic capability to outperform and outturn it's opponents in a dogfight, with more fuel to work with in an engagement. And that's not all. Not too long after the F14 came out, the United States brought out the F-16, and then the F-18. The Russians simply couldn't keep up, and they realized that these aircraft and the force stacking capability they represented made an impenetrable wall against their conventional military tactical ideas. That's what top gun showed the American public: that the Russians could be beat, and that maybe we might all live to see the coming millennium. VERY excellent video, and please believe me I don't blame you for not understanding why it was so popular. I'm really glad that you (likely) never had to live with that daily dread.
See, here is the thing though... I was born in 1988 and I saw this movie in 1994-1995 at a babysitter... I DISTINCTLY remember the movie and it began what has been a nearly 30 year obsession with the F-14 Tomcat. I was extremely disappointed that when I graduated from High School, the F-14 had been retired. There were no threats of nuclear annihilation in the post Soviet era when I saw the movie, and yet, I still love the movie today because it, and The Final Countdown, are two movies just chock full of F-14 porn... this super sexy high speed extra lethal display of US Military Might. I only began to really appreciate the things the F-14 accomplished as an adult... the F-14 was the F-35 of the 1970's... it defined what it meant to be a fourth generation fighter, and for nearly two decades it was the only fighter in US service which could fire multiple radar guided missiles at different targets and have it hit.
Or because this UA-camr is tone deaf and missed out on the jingoistic and nationalism dripping out of Top Gun during the Reagan era and its unique concept and cinematography. Everybody wants to think they’re the good guys. Just look at us justifying several generations of ethnic cleansing in Palestine.
"When that very brave Russian officer defected with his aircraft on September 6, 1976 to the Western allies, the United States was able to dissect the MiG and learn its weaknesses. From that came the F14.... " Uh,.... Tomcats flew CAP over the evacuation of Saigon in 1975. Their development started long before Belenko flew his Mig-25 to Japan.
@@thatguy6054if anything, the Foxbat was one of the reasons the USAF got spooked and spurred them to make the F-15 The Belenko incident basically washed away all notion that the MiG-25 was a very capable aircraft, where infact it was a VERY, VERY chunky interceptor that was powered by engines that were derived from a cruise missile.
Thank you youtube for recomending me this MASTERPIECE of a channel. There is not a single other channel that goes into as much detail on stuff like this besides yours. I have looked for ages to find information like this, and have never found it. The video on the AGC was also so unique. This is so amazingly accessible and well written, bravo :D
Great video. In the past I have been involved in two small processor designs.. a silly one where I had an evaluation copy of a system level design tool and a 150 cell CPLD and an EPROM.. and another that sold in the millions , the de emphasis digital audio filter in a Philips CD player chip. It had to do several multiplications and some additions all in 24 clock cycles. It was bit serial and used modified Booths algorithm multiplication .. two bits per clock cycle on 18 bit data.. 18 bits because it began life in a Digital Compact Cassette player..
Wow. What a dense summary. Well done. It’s very difficult to communicate a complex concept that you have very little information on. I appreciate the skill shown here.
Most of this went well over my head but watching was well worth the time regardless. Your commentary only gets better with each video and I appreciate your willingness to maintain a degree of silence with limited music and other background noise. You also refrained from saying the F-4 is ugly even while it featured in the video :)
The flaps and slats were interconnected and never moved individually, if any one of them got out of position by more than 2 degrees from the others, the system would lock out in that position. There is a set of position sensors at the end of each flap and slat, when the aircraft was put under high G the computer would deploy the flaps and slats a few degrees to increase turn ability. The inboard flaps only functioned when the flap selector handle was actuated, they were either up or down. If the flaps were selected down the wings had to be in the forward 20-degree position for the inboards to work, they were hydraulic actuated and would blow back up if the airplane went over a certain airspeed. If flaps were deployed at all the wings would only sweep to about 50 degrees. When wings were forward, roll was assisted by spoilers when the wings swept back past a certain point the spoilers failed down and they were locked out, roll was by the stabilators only at that point. Grumman developed the F10F Jaguar in the 50's with swing wings, I think they were manual only, but the company was a prime subcontractor on the F-111, bringing their swing-wing experience to the project, the CADC was probably developed for it and refined for the F-14, remember that the SECDEF at that time wanted the Navy and Air Force to have the same fighter/ bomber since the F-4 worked for both services. The F-111 sucked for the Navy for a number of reasons and Grumman left the program and developed when the Navy initiated the FTX program. Politics killed the cat.
Great video! I really wish we would have a program today that could be used to verify that every transistor in the system is actually working correctly. Even if running that program once took a whole day, it would still be highly valuable to verify the hardware that you have, either to verify it after building a new system or to diagnose unstable system. Currently the state of art seems to be "randomly swap parts until it starts working again for the programs that you actually use daily".
Programs can be written and the trick is to be able to prove that a "single failure" is picked up. Overall its a simple concept. If a * b = c and if c is the incorrect answer then you have a problem. How many a * b or a/b do you have to do to catch all single failures. Now multiple failures are a different story. Ray Holt
@@firstmicro The problem is that you cannot know by definition which part of the hardware is potentially broken. If you *compare* any numbers to each other, what if the comparision operation in the CPU is broken? Modern CPUs have multiple execution units per core so even if you successfully execute 5 comparisions correctly with known values, you cannot know if the 6th execution would actually work correctly or not. And since the actual execution units are selected by proprieatry algorithms inside the CPU, you cannot assume you have a safe way to run any kind of test code on all units. We would need some hardware support to be able to test all the transistors in the CPU. For exaple, ability to select which execution units are in use (basically disable everything but one) is needed to check if those units work correctly. And to make things harder, the L1 or L2 cache could be broken and there you have the problem that e.g. 16-way associate cache can have the cached value in different locations in the cache and any random execution of the code may or may use the part of the cache that's broken. It's turtles all the way down and at best you can give some kind of statistical estimate how correctly the hardware is working.
Stunning! Early computer history suffers from being stored on media (paper and restricted scanned images) not widely viewed by current students so these videos expose new generations to how computer fundamentals came about. Trouble shooting these types of systems is very tricky (and not always successful). I worked on directly connected adjacent parts to the F14 and to circuit board construction is just jaw dropping. The printed wiring boards are super high density and battle hardened. The difficulties in programming are unbelievable. The original engineers were amazing.
For someone who doesn’t like Top Gun, you sure picked an interesting font 😅 This video is great! This topic just came up during a family dinner. My brother and I were huge nerds for the F-14 and other aircraft of that era. Thank you for this!
Ok. I have an F-14 adjacent and somewhat funny sort of story. I was an enlisted onboard the USS America CV-66 from 1975-1978. I worked in CVIC, the intelligence center. We had cypher locks on the main door entry and one of us enlisted guys would have to do a four hour watch checking IDs and "buzzing" people in through the door, unless they had approved access and knew the combination. CVIC had a main entry that was connected to a large debriefing room for the pilots. Other functions were behind another cypher locked door. During Med Cruises F-14 pilots would come down off the flight deck for debriefing while in full gear and we'd buzz them in without having them show us their military ID. Invariably our boss would periodically insist that we had to follow the guidelines and make everyone show us their military ID. The look on the faces of these pilots when told we needed to see their ID wasn't so much as awkward as just plain pissed. They would be carrying their helmets and possibly some other gear. They would then struggle to dig out their ID from whatever inner pocket it was in. Eventually, they'd complain to their CO who'd complain to our CO and we'd be allowed to relax the requirement for pilots. And of course after three or four months the instructions from on-high would be that everyone had to show ID. Needless to say, I always thought it was a really stupid requirement. These guys are flying F-14s, over the Mediterranean Sea. Exactly how likely was it that their plane, mid-air, would be commandeered, they'd be stripped of their uniforms, tossed out at speed (I guess), the enemy would dress in their uniform and gear, then land an F-14 on a US Carrier, then come down to CVIC and expect access. I always figured that if a Russian could do that then we'd already lost the Cold War anyway and we should give them access for their debrief.
One thing, I think, you missed that was controlled by the flight computer were the variable ducted engine intakes that changed their profiles as the engines went from subsonic flight into mach speeds. I seem to remember one of the F14 WSO/RIOs talking about this...and in the early engine mods could cause a flameout under the right conditions.
Those were controlled by a different computer. Probably because the engines have a different manufacturer than the airframe, and existed independent of the F-14. So, use existing means to control the ramps, or add more cost and complication to a CADC.
@@pgtmr2713 The F-14's air intake geometry is built into the engine nacelles and air intakes... the control of the actual variable intake geometry was a Grumman design specifically to allow the TF30's to operate in the designed speed regimes. When the GE F110 was installed in the F-14B and F-14D, the scheduling of the intake ramps was changed to optimize performance in the transonic region. The end result was the TF30 actually produced more thrust at low altitude and at high mach (not necessarily both). Air stacking in the intake allowed the TF30 F-14A to go to Mach 2.34, the GE F110's were capped at 1.88.
Interesting video. Reminds me of the seventies when I started designing embedded industrial control systems. Back then we made the distinction between mini computers and micro computers. A one chip CPU made it much easier to build a system compared to building all with TTL circuits.
Great video, brought back memories. I graduated with a computer science degree in 1978, and in 1979-1981 I worked for Marconi Avionics, Rochester, Kent. I worked on the first computer based air data systems for Altitude and 2 dimensional speed for helicopters. There were still a team of watchmakers who serviced the mechanical computers for the Vulcan bombers.
The glove veins were not even built on versions of the f14 past 1980s. The reason was they did not have any large effect on maneuverability for the cost, so they stopped making them.
Found your channel through the algorithm. Great stuff. Thank you for creating. I have only watched a few videos on processors, but I intend to binge when I have to time. :)
While in the US Navy for 20 years, 1970 to 1990, one of my main jobs, was to work on F-14 flight control actuators. One thing I was taught to do at a special instruction course was the wing sweep actuators and glove vane maint.. I could check and test the sweep actuators but was not allowed to go into them. They had to be sent to the Depot in Norfolk , Virginia for a complete overhaul. I was asked to do it several times, but I was smart enough to refuse. If the ship was in port ( USS Nimitz) I could get the paper work and hand carry to the depot that was a couple of miles down the road. I would want to see what the technicians there would find, particularly if we had a run of bad ones. Some times it would be bad aircraft wiring, a bad cannon plug or even that ADC, and not the actuator. It made for some good arguments. I spent 10 years in Squadrons, and 10 years in Aircraft Intermediate Maintenance Departments (AIMD) I also spent 5 years after I retired doing F-14 structural repair at the Norfolk , Virginia Naval Aviation Depot, and I've just retired from 20 years at the Coast Guard Depot at Elizabeth City, North Carolina.
I actually heard about this growing up, looked online for more information, didn’t find anything, and assumed I just was mistaken. Thanks for the interesting video!
While filming the USS Lexington with my drone(and at every Museum ship for that matter lol) I toured the museum for a whole day. If memory serves, there is a big display talking about this subject but in a manner for all to understand. Its really cool having this video to really flesh out the subject. I remember they had other examples of early computers and so on. Support your museum ships people! ;)
I wanted to add that the CADC hardware prototype ran at full-speed of 375KHz. It was not subjected to mil-spec. The purpose was to simulate the hardware and the programming.
Your click picture was taken on my ship the USS America CV66. Carrier Air wing commander Mike Snodgrass was performing a fly by. Thanks for the memory.
If anyone thinks you're a weirdo Alexander, it's because you're making 1 hour long videos about the F-14 CADC, not because you were in a UAS! Great video, can't imagine the effort that's gone into this for our benefit.
Ha was going to caption that bit as ‘a weirdo but not that type of weirdo’. To clarify, I hd an awesome time in the UAS and am proud of what I did during that time. I just don’t want to come across as one of those people that pretends they were in the actual military.
I like how someone had an idea, used the materials available and created a proof of concept. Now that 'it' works and is something useful, our ability to refine it changed the world.
I worked on the F14 in 1994. There was the Z8080 embedded in a center console black box. I had an old programming manual from the 70s and was able to work on it.
Great video. A couple of niggles (I know you love them): iirc the program which converts assembly language into machine code is called an assembler and not a compiler. I’m sure I’m like the 100th pedant to point this out. And between the 4004 and 6502/Z80 were the 9900 (no registers, first consumer “16-bit” chip), the 6800 (inspired the 6502) and the 8080, among others.
In the 60's, IBM was working on Josephson circuits, multiple superfast processors in a liquid-nitrogen bath (to lower electrical resistance to near-zero). Those Josephson processors/circuits were microcircuitry prototyping.
What is a microprocessor? In reality it really does not matter, however, companies and marketing people love to label what they have. When I started the design of the F14 CADC the implementation of the “processor” was unknown. As we tended towards the new “integrated circuit (IC)” technology names such as IC processor or LSI (large scale integration) was used. Micro was also in there somewhere as the computer design evolution progressed from large, to mini, to micro. It was not a big issue to anyone. Ultimately, as the CADC project was completed and I wrote my paper I chose the name “microprocessor”. There was no focus on single-chip, multi-chip, program counter location, etc. In fact, even Turing was set aside as the main focus was to do the job irrespective of # of chips and what function was where, and even how data was transferred. The selection of any cleaver design techniques (pipelining, co-processing, in-flight self-test, dual redundancy), algorithms, and package selections we made to reach the speed and accuracy necessary to fly the F-14 as needed. It was unfortunate (for me and history) that the 1970 presentation of my paper and any patents were strongly refused by the US Navy thus 30 years of silence started. Around three years later the Intel marketing group started labeling their work. First came the World’s first microprocessor, then the World’s first commercial microprocessor, then Intel’s first microprocessor, then Intel’s single-chip microprocessor …. Today, everybody wants their design to be a microprocessor, otherwise, how are they going to out market Intel? In 1973, Hank Smith, Microprocessor Marketing Manager, Intel Corp made this statement at the IEEE 1973 WESCON Professional Program Session 11. “A CPU uses P-channel MOS and is contained in 1,2,3 or4 LSI standard dual-in-line packages from 16-42 pins per package”. Even at this time the word CPU was sometimes used for the entire processor and not microprocessor. After this, history started dissection the word CPU, microprocessor, single chip, etc to fit the functionality, design, and packaging of each product. Thus, here we are in 2024 (54 years later) still trying to make it all fit into convenient definitions. Ray Holt
This just makes one think about the advancements that are out there at the moment that we the general public don’t have access too. Hard to believe how ahead of their times those engineers were with what they could accomplish
The use of microcomputers in the F-14 to command aircraft flight controls closely follows the terrain following radar controlled flight of the F-111 fighter bomber, which Defense secretary McNamara infamously foisted on the Navy as a do-everything fighter/bomber. As a first generation swing Wing supersonic fighter the F-111 had huge teething problems and fell far short of the Navy's requirements, however, Grumman was well aware of every outstanding attribute or shortcoming of the F-111 in designing the F-14 soon after as a superior replacement. And even before the F-111, the Apollo moon rockets used some of the first integrated circuits - I believe they were Fairchild microprocessors? - in their rocket guidance command and control. Even more astounding were the analog computers used in World War II, especially the big firing control units that estimated range & velocity for the huge Naval guns on the US battleships, and the far smaller Fire Control computers for the remote gun turrets on the B-29 Superfortress bombers, which apparently were accurate to twice the distance of purely human aimed gun turrets. It is amazing to think that the Concorde, the 747, the X-15, the SR-71 and Appollo rockets we're all designed by slide rule in the days before engineers had electronic calculators. And finally, it is a shame that another defense secretary mafia, this time under dick cheney, pulled the rug out from under the mighty F-14, instead of upgrading it with the newer engines and avionics they deserved, forcing the Navy to adopt the smaller, less capable but less expensive F/A18 Hornets
The MP944's architecture was ahead of its time in terms of parallel processing and pipelining. It's amazing that they were implementing these concepts back in 1970 for the F-14's CADC.
Dang, I never would've imagined anyone would think military aircraft were scary/ugly. They make me proud of human ingenuity, and that's why they're beautiful to me - they're feats of man's wrestling with the forces of nature, and one another.
Sir, with great respect, while your comment about high aspect wing ratios creating more drag was not necessarily incorrect, high aspect ratio wings actually have a higher lift to drag ratio. That is to say, of a high aspect ratio wing and a low aspect ratio wing of equal lift coefficients, the high aspect ratio wing will usually produce less drag (case in point the high aspect ratio wings of the U2 or a modern airliner being designed for maximum efficiency while lifting a heavy aircraft). It would be truer to say the Tomcat in the "T-pose" would create higher drag because it has an increased wingspan, to gain additional lift at low speeds. Love the video though!
I am endlessly fascinated by mechanical computers for some reason. They’re just such a seemingly crazy blend of technology and discipline in design and construction.
I did come across that as part of the research for this. It was the first use of the term 'microprocessor'...though I don't really think it falls under the generally accepted definition of one today (though, of course that is very much open to debate).
Wow - I learned from you that the F14's swing wing didn't require straight and level flight to change - that's impressive! I'd definitely previously made the same assumption you formerly had, and your video was perfect to undeniably emphasise the impressiveness of this by video of the F14 rapidly changing the swing wing at 8:26 and 8:40 (and introduced a bit at 7:40 too). Awesome! Thanks for your video =)
Excellent video (speaking a SoC designer). If I can just point out one detail: the tool to convert assembly language to binary code is called an assembler, not a compiler. A compiler is for high-level languages like C and it produces assembly language. The confusion in recent times comes from the fact that the assembler has been rolled into the compiler. If you program at low level, you notice that when write assembly language within C (by using dedicated macros such as "asm()" or by including .s files in your projects).
"You couldn't for example build a practical desktop computer with a 4 bit processor". I can't see why not. The "4 bit" in the 4004 was the ALU and data path width between the four chips. The memory addresses were actually 12 bits and that's the far more important number. It's still a bit small, but all the "8 bit" microprocessors had 16 bit memory addresses and people did quite a lot with 8080/z80 CP/M machines, the Apple ][, the C64, the Atari 400/800. Those had 2^16 memory locations each of which was 8 bits in size, while the 4004's 4096 memory locations were only 4 bits in size. But a 4 bit computer with 17 or 18 or 20 bit addresses would be just fine, as long as it ran at sufficiently high MHz. Note also that the z80 has a 4 bit ALU, two 4 bit buses, and separate latches for the hi and lo 4 bit ALU results, which are produced in different clock cycles. What surprises me about the 4004, looking back, is how many registers it has. There are 16 of them! So you can do quite a lot of calculations without hitting RAM at all. This is quite a stark contrast to for example the 6800 and 6502 which had three registers each for program variables (plus PC, stack pointer, and flags). Altogether the 4004 had 17 4-bit registers, 4 12-bit registers (PC plus 3 return stack levels), and a carry flag, for a total of 117 bits. The 6502 has 5 8-bit registers (A, X, Y, S, P) and a 16 bit PC, for a total of 56 bits, less than half as many register bits as the 4004!
You say addressing width is far more important, but bus width bring speed of operation. I'd say they're just different, except that when it comes to powerful computing like GPUs, it's often this data bus width that is the limiting factor of the architecture.
"You say addressing width is far more important, but bus width bring speed of operation." Yes, that's exactly what I said. Addressing width is fundamental. It determines the complexity of programs you can run. BUs width is merely speed, which can be compensated in any number of ways. “There is only one mistake that can be made in computer design that is difficult to recover from-not having enough address bits for memory addressing and memory management.” Bell and Strecker, 1976. @@VndNvwYvvSvv
It had good amount of registers but it was needed for addressing, fetching, arithmethic operations and narrow bus. The CPU lacked logical instructions and when I researched it it seemed it's hard to write more complex programs due to these limitations. It was really calculator on a chip designed for decadic calculations. Even the 4002 RAM chips were designed to store mantissa and exponent with sign in separate locations and you had to use different instructions to access them.
Wicked video. I am a child of the 80s, born the year after Top Gun released, so it was a big part of my childhood and I still rewatch it occasionally even now. I've always loved the F-14 as a result, and it was the main reason I clicked on the thumbnail for this video. My other minor interest in computing being the other reason. Never knew how much computing was going on inside those things. Makes a cool plane even cooler imo. Always a shame that folks who make such incredible discoveries and put so much committed work into their projects get no recognition in their time due to classification. I get the need for it, no issue there, just it would be nice if they could be recognised for it when they did it, not decades later when the files get declassified coz they're no longer relevant. Ingenuity is humanity's finest trait imo. We can figure out such incredible solutions to the problems that stand in our way. To the point where we can shoot human beings out of our own atmosphere atop massive engines and put boots on the ground of other celestial bodies. We're an amazing species when we put our minds to advancing stuff like this.
Since a few people have asked how the program counter can be stored in the ROM. It was a special register physically on the ROM chip, not a read-only address in ROM itself. Ray's paper states it was resettable and steppable and could accept an address input. So, I'm 95% sure that allowed for full program flow control. @phirenz makes a good argument for why this may not be quite so simple though....and may have actually required additional ICs to actually implement.
The SAAB Viggen is getting a lot of mentions in the comments. It had a highly advanced flight computer composed of many discrete ICs. Not a microprocessor but very impressive nontheless. I am considering doing a video about it in the future.
The 'MiG-25' at 6:11 is actually a MiG-31. Thanks @AvArIeNmArKu4
An ‘assembler’ translates to machine code, not a compiler. Thanks @Mat-Ellis
Swing wings are a subset of variable wing geometry. Variable wing geometries may be implemented in other configurations and the field is not 'obsolete' as I stated. Thanks
@nobodynoone2500
31:48 bottom line: "U if P > L" should be "U if P > U". Thanks @MatthijsvanDuin (I'm really mad I got that one wrong!)
10:00 - Lateral Axis, not longitudinal axis. So, the aircraft is pitch stable. Thanks @UnitSe7en
45:00 - The hardware prototype ran at the same speed as the final product.
And yes, I know higher aspect ratio doesn't equal more drag (it's the opposite for subsonic flight). I was trying to simplify and probably simplified a bit too much...I said I always screw up supersonic flight discussions.
[ retired Engineer ]
Caught out on a technicality...
When I bit-banged on 4004,8008/80/88 and 6502 a microprocessor _was_ a single chip, not a collection of six.
One of the Engineers that worked on it wrote a book that was published in 2017!
PS I'll still give the vid a thumb though, you made a half decent vid.
I don't think you can call something a compiler if it just translates from ASM to machine code. I'd call that an assembler.
The F-14 could engage 6 targets simultaneously but it could track many more. The F-14 could track 24 targets and fire on any 6 with the half dozen Phoenix missiles it carried.
Very interesting video I really enjoyed it and recommended it to a few friends.
An assembler IS a compiler, albeit for _Assembly Language._ EG - MASM, TASM, WASM.
- ex game programmer
@@hamesparde9888It's still a compiler, just for a lower level language. I wrote compilers for 3 years (hard stuff!)
The old chip designers were crazy. No simulations, no HDL, no DRC/LVS. They would cut photomasks by hand. Imagine seeing some bug in silicon and coming up with the concept of setup and hold time to explain it.
The old stuff that gets me is magnetic core memory. Making those by hand is just nuts.
@@whyjnot420 its why memory used to be the most expensive part of any system.
@@mrrolandlawrence Yeah and if you do the math, having todays amount of memory done that way would bankrupt many a nation for a single computer.
There's a great amount of info and real world experience in "Commodore: A Company on the Edge" detailing the process they used to design the 6502, hand cutting rubylith to form the masks for the different chip layers. Absolutely mind boggling how hardcore chip designers were back in the day. Even more impressive when they then go on to explain how their chips "mostly worked" or in some cases "worked 100%" on the first batch with those old processes.
@@mrrolandlawrenceactualy it still is. Registers and caches take up a significant area on any chip that uses them, jacking up the price.
Having flown the F-14A "Tomcat/Turkey" and then going on after my time in the Navy to work in the computer world... I can't count the number of times I had to bite my lip during conversations about CPUs and sit and wonder "when" until the Osborne, XT, Kaypro and Compaqs came online. We knew the processor was there however we really didn't get into in any depth. We were just dumb Aviators... NFOs. Coming from the A-4 "Skyhawk/Scooter", my initial instincts were to turn it off and control sweep manually. But, I came to trust it and in ACM school learned when to turn it off so as to fool the adversary as to your airspeed, turn it back on to get into position to get a splash. The real beauty of Bob Kress' overall designed was that he used this computer control and the airframe design to do something completely new... it didn't roll at the nose, but at the HUD. So when engaged in ACM, the plane rolled at your eyes, not your feet - Evolutionary. I never met Ray, but I did get to meet Bob in the mid-80s in Bethpage, NY.
Was that a Data General Nova in your video (yellow/brown switches)? Worked on those too. I watched this with great interest.
Awesome insight
It does kind of piss me off that bits and pieces of this architecture were not released to Intel, DEC, etc to jumpstart the microcomputing industry. They could have done that without revealing the specifics that made this so powerful in this application.
What a comment!
Does the F/A-18C do the same as in rolling at the eye level?
I always wondered why it felt strange when playing it in DCS and rolling, and this would explain that if it does.
@@mikedavis7065 DEC already had processor kits consisting of several ICs at the time, at least PDP-8. Also, you don't need 20-bit CPU to perform 20-bit calculations, it's misconception.
The F-14 was the equivalent of the US Navy's current NGAD (next generation air dominance) program for 6th generation fighters. It was large, expensive, and had computing capabilities that people didn't realize were possible.
That makes me think of the F-35, aside from the large part.
I admit I'm pretty ignorant of recent military computing, but I don't see how NGAD would be using leading edge computing technology as significantly as the F-14 was. At best they're probably using some Intel FPGA's in impressive ways. I'm not sure they have the hardware volume to create an ASIC on a leading edge node like Intel 4 (7nm)
@@1funnygame High volume isn't necessary if you accept very high unit cost and military applications are notorious for accepting them.
@@1funnygameMy understanding is that the biggest advances in modern military computing are on the software and integration side rather than the hardware, which tends to be a few generations behind.
While mainframes at the time and later military microcomputers had capabilities, I'd be surprised if this system even had an MMU. I suspect the first integrated capability system remains the 601.
I entered the USAF in 1978 and was a aircraft mechanic on C5’S they had a top secret (or at minimum secret) system called “ MADAR” (Malfunction - Data - And - Recording) it basically diagnosed the plane in real time and when called up the mechanic on board could figure out what he needed to repair the aircraft it also could be downloaded to put in the aircraft records for future use. Pretty cool stuff for that day!
This is a Lockheed feature in general; it was also available in Civilian aircraft like the L-1011 Tristar. The milspec one was probably fancier but Lockheed was ahead of its time in this regard.
My dad was a flight engineer on the C-5 and I remember him talking about MADAR all the time.
I wonder if that eventually evolved into the self-diagnosing FBW computers that are used today on commercial airliners (probably not, I expect they were developed independantly but it's a fun thought).
@@Alexander-the-ok MADAR itself didn't, but Lockheed was about 20 years ahead of the competition with it. The systems that came about in the late 80s/early 90s brought similar functionality to other manufacturers' aircraft. I don't know how much lineage current Lockheed aircraft systems have to MADAR but it wouldn't surprise me if there's still some DNA.
MADAR that was a hell of a system, until you have to reload the program from tape with the engines running and flight crew staring at you. Bubble memory lives!
I grew up around that airplane. It was the topic of conversation for decades because my old man was a test pilot who flew f-4's then moved to f-14s. He retired as the director of flight systems.
For his masters he wrote a program that let them load and analyze flight data beyond the memory capacity of the old mainframes. Nasa was building the x-29 at the time and everyone in flight test needed more capabilities to analyze flight data.
how were they storing flight data in those days , by tape ??
@@cardboardboxificationeither tape or maybe floppy disks.
But harddrives were available at that time too
@@cardboardboxification There's a humunculus in the plane that scratches it into a stone tablet.
Oh my god, that is astonishingly cool!!!
I worked on the 4004 back in the day for Intel. It was not a stand alone chip either It required a clock chip and a couple level translators. It was a PMOS chip and did not work with standard logic. The IC's required were the 4001, 4002 and 4024 as I recall.
Even processors of the 2000s required a northbridge and southbridge to handle I/O.
@@tankermottindI used to be into building desktops back in the day, and I remember needing to keep an eye on northbridge and southbridge temps when overclocking. Have we really made those obsolete in the 15 years since I last looked?
@@mrb692the northbridge is mostly integrated into the CPU nowadays, although the most recent innovation is having it be a separate chiplet in the package on a lower density manufacturing node (more nanometers) compared to the CPU cores. The memory interface part of it was integrated in Athlon 64/Opteron days and was one reason they dominated over their Intel contemporaries.
The Southbridge mostly still exists, except the expansion cards and mass storage (which used to run PCI and IDE/SATA through the Southbridge) now connect direct to CPU PCI-Express lanes (optionally through discrete PCI-E switches to distribute that resource between occupied slots). Southbridge (or "chipset controller", "platform controller hub", usually connected over PCI-E instead of a dedicated bus) handles slower stuff, like networking, SATA, slower USB (the bit rates of USB 3.2 and 4 have reached the point that CPUs have started to integrate some USB controllers for lower latency access). Mind you, 4 lanes of PCIe 4.0 to the Southbridge is still close to 64 Gb/s, so it's no slouch... But USB 4 is hitting 40 and 80 Gb/s now.
@@mrb692 yeah its pretty much all integrated on the cpu now. Even memory controllers sit on the cpu die now. Its neccessary since data moves so fast now, cutting the "middle men" on the motherboard was a huge leap forward in modern computing.
@@wizzalien7796 Whoa, that’s pretty sweet!
Note on the glove vanes. They were designed to preserve agility at high supersonic speeds and would have had little effect at lower speeds. It was later found there was almost never a need for dogfighting at supersonic velocities; combat quickly became subsonic where the glove vanes gave no advantage As a result, glove vanes were disabled and/or deleted from later production runs to save weight and maintenance.
I remember it could perform 100k operations per second, which was unbelievable. I believe it could track 9 bogies and shoot at 6 simultaneously. Pretty crazy stuff back then.
do not be confused, the CADC is the flight computer, the ability to track and shoot 6 at the same time was due to another immensely complex part of the plane, the AN/AWG-9 radar.
@@Galf506 You are correct. I was assuming that they used the same computer. Like I said it was a long time ago! lol!
@@Galf506the AWG-9 was made for the older F-111B.
not everything is about the F-14.
I'd say the F14 was the most intimidating looking fighter for its era. The F22 with 'symbiote' skin is more mean looking imho.
How do you make a jet harder to detect…. Make it chrome
It would be scary if they could get it to spend more than 5 minutes in the sky per 100 hours maint
@@meangreengoblin1besides heat signature and radar
@@keithsimpson2150it’s well worth the money and if you think that’s a problem how do you think the NGAD will be Lmao?
@@keithsimpson2150there was a proposal called st-21 super tomcat. Basically modern tomcat on steroids.
Top Gun was popular because it was the first movie where the air to air scenes didn’t look like tacked on scenes filmed in the wrong aspect ratio. Everything flowed together, and it was a good yarn of a story.
for real, i cant think of any other film off hand that was remotely as seamless. if you were a fighter jet fan nothing else compared. the original was a bit homoerotic though, ive not seen the new one lmao
the new one is not bad, they still play with balls
on the beach that is
Also the film catered to women because of the love scenes.
Imagine this is the exact field my undergraduate education was happened to be in, and they failed to give us the proper prospect of what we were actually learning over there breaks my heart, imagine a one hour presentation equal to this could have turned my entire understanding of my field around, and then... it was heartwarming to see even after 20 years of shelving my education i still fully understood all of this... thank you for the effort, both thorough research and admirable presentation... you sir, are a wonderful teacher...
Computer architecture and design?
@@_________________404 at our place computer science was divided into two undergraduate branches, hardware and software... i was a Hardware student about the end of last ice age...
@@sohrabroozbahani4700
why not pick up where you left off.
its never too late to learn new things.
@@sohrabroozbahani4700 20 years isn't that long ago... i question the truthfulness of your nonsense.
It is sad that history often becomes corrupted into fiction. Important developments are forgotten and trivial things are elevated to legendary status. The F-14's multi-chip microprocessor is one of those important developments that has largely been forgotten. Thanks for posting this valuable piece of history.
"It was in TopGun. You know, the film about volleyball" 😅 So true
Having re-watched that scene, I think I may now understand why the film was popular.
@Alexander-the-ok "playin with them boys".. weren't even trying to hide it, lol
@@Alexander-the-ok the flying scenes were the real deal in real Tomcats, which is the only good thing about the movie. Otherwise, basic Hollywood 🧀
Yep, that’s when the movie went full gay
@@thomaswilliam630 the other thing was the "hot" chick might be the most manly ugly hot chick of any big budget movie either. No wonder they cut her from the remake, it looks like she's been on testosterone for the last 40 years.
Professor Frank: "I predict that in 100 years, computers will be twice as powerful, 10,000 times larger and so expensive that only the five richest kings of Europe will own them.”
Watching this on my Apple Watch
lmao 5 richest kings in europe arent even the 5 richest dudes in europe
Prof. Frink
Fascinating! In the early 80s, I worked in Garret Airesearch's environmental test lab. I received my first Confidential clearance so that I could work on the SCADC program. I can't remember the F-14 system ever being mentioned back then, but that could be just due to oldish age. The SCADC was a standardized air data computer that was intended to replace older units on A-6, A-7, and A-4 aircraft . The same electronics components but packaged to fit in the existing electronics bays of older aircraft. There were four total configurations, the fourth being sort of universal fit. Seeing this video, I can see the similarities. Environmental testing was to MIL-STD-510...vibration, temperature, altitude, humidity, etc. Fun time for me working with centrifuge, explosive shock testing, acoustic shock, explosive atmosphere. I left GA shortly after completion of these tests, going on to work at McDonnel Douglas on C-17 (FCSS and Flight Test and Escape Systems (ACES II Advanced Recovery Sequencer, a compact digital computer itself).
Which facility did you work?
@@firstmicro Garrett-Airesearch, 190th St Torrance and LAX.
How long lasts confidential clearance?
@@orangelion03 Torrance is where I developed the CADC. I left in 1971 after the CADC went into production.
In 1975 I was trying to work on the f-111, as a result I also did work on f-14s that came into the air base their systems are much alike because Grumman was a partner in the building of the f-111. You might be surprised to know that the F4 was originally designated the f-110 .
As a result I did work on several f-14s and was familiar with their systems. Because it had to integrate a lot of mechanically driven actuators and control systems, it did require a lot and I do mean a lot of Maintenance time to keep it in the air. We complain today about the F-22 never being deployed with more than five planes at a time, and you are generally lucky if you can keep 50% of the F-22 flying at any given time. Much was the same about the f-14. It was an extremely difficult plane to keep flying. Especially in its first iteration.
Iran was extremely short-sighted attempting to use that as their primary fighter. I'm sure that some mechanism of the are data computers were reduced in function for their application. However, I also worked with many Iranians. They were extremely good at understanding Electronics to the highest degree of the day.
After having worked on a few of the f-14s I understand exactly why. When Carter cut our nation's throat and increase the instability in the Middle East by forcing the Shah out of Iran. It quickly rendered the vast majority of the Iranian Air Force, notably the f-14s out of existence until they can figure out how to get parts for them. They quickly learned how to build their own parts for those aircraft. That's why they still had several of them flying as of just a few years ago, I don't know if any are flying today. The earliest iteration of the F-14 was nothing to behold except when it was brand spanking new and everything worked on it properly. That was not always the case, as a matter of fact, it was rarely the case.
I do not know if your remark about the f11 being difficult to fly while changing its wing sweep is true, it was never true on the f-111f model that I was tasked with.
I also never had a problem flying the F-111 simulator, as I was the first one that the simulator Commander had ever seen fly it for the first time and not cause it to crash. I am not certain that we had chipsets in any of our flight data computers. We may have had them in the nav computer because even though it took the old school tape to load Mission parameters. The F model could be tasked with running a alternate routes to and from the bombing point, and never exceed 400 ft off the deck while doing so. You rarely would do a mission and approach a Target in excess of Mach 1 because you would leave a clear Sonic Trail. However you could do so quite easily.
I never got into an F-14 simulator during my time in the Air Force because we didn't have any. But as an avionics Tech I was able to show them how to do things like engage terrain following radar while in articulated Terrain. That procedure leads me to the concept that we may have had some sort of microprocessor in the F-111. I cannot be certain, but it is possible. I also have a good feeling that given the restrictions that you had to mechanically put into your movement of the wing at times and other clues, that there was no chip on the flight data system, except in regards to navigation. I did see several iterations of the navigation computer in other devices, some still in the 1970s as a matter of fact.
The last iteration of the F-14 was the most capable and had they upgraded it 2 a more powerful set of engines, something closer to what the f11f model, which had the same engines as the early F-16 and f-15, it may have been a much better aircraft. The unique issue with the high bypass turbofan hybrid engines, was that if you pulled back power on it too far and weren't paying close attention, you would scrub off speed at 50 knots per second. That is slowing down at a really fast rate and will you would get yourself in trouble, perhaps that's why you made the claim that the F-111 was hard to fly in transition. I never found such to be the case. You just had to remember to always run with a little bit more throttle than you thought was necessary. It was easy to adjust the speed of the f11 merely by sweeping the wings forward. With five stages of afterburner, you didn't even want to be on stage 1 while actually performing a mission. The f11 I would contend still to this day would be impervious to even the most modern anti-aircraft systems. There just isn't enough time to see it recognize it and shoot at it before it's already past you. Being able to Chase and shoot it down would be a Fool's errand. Especially if you were in any kind of articulated terrain, the ability to actually stay with it would be difficult in the extreme. Aircraft of North Vietnam even found the a model impossible to follow. That was however a much earlier time frame.
I don't know what happened there, ,and how. But the Soviet Union took possession of the F-111 back in the days of the Vietnam War. There is a story on this topic.
I have a good lady friend who was a mechanic on the F-111 and stationed in New Mexico. She was fond of her plane.
@goddoesnotexist5688 You can fly a jet in a sandstorm?! If I saw that in a movie I would be certain that was made up. I vaguely remember hearing about an airliner that had its engines fail because of a volcano spewing ash: en.wikipedia.org/wiki/British_Airways_Flight_009#
Apparently there was another such incident: en.wikipedia.org/wiki/KLM_Flight_867
Not sure if volcanic ash is worse than sand though.
This was a great comment, thanks for sharing!
One thing on F-14 there is absolutely no plausible way for Iran to fabricate are the wing-box forgings.
If USA had to remake those wingboxes it would be doable but take some time. They were forged on what was, & possibly still is, the largest forging press by size in existence.
Simple drop-forging vs the complexity of hydraulic forging, its a multi storey forging press in Ohio that is one of a kind in the world, it is still used for extremely large & specialized forging applications.
Majority of Soviet Unions heavy forging was done in Ukraine where all it's military shipbuilding & largest aircraft were designed & built.
I would like to see drawings of the wing-box assembly,
perhaps its pivot-pins used high-shock bushings & pins of forging-die alloys like A2 D2 O3 & Inconell, these are all very challenging to cut & press concentric into a forging without inducing stresses that could lead to "egging" or contact galling,
bushings might delay the inevitable but everything gets work-hardened & microcracks over use time including a forged wingbox,
I cannot imagine these Iranian F-14 wingboxes are still functional unless they are never flown, flown like a Piper Cub with no stresses on the wing-box, or possibly modified or replacements retrofitted as 1 piece to operate only in a fixed position
Worth noting that when transitioning to supersonic flight, the center of pressure moves back on every aircraft. This is not only due to the F-14's swing wings but the interaction of pressure with supersonic shocks.
Something that the Concorde had to deal with by pumping fuel from it's wing tanks to the very back of the plane as ballast.
@@JPkerVideoThey couldn't very well stick trim tabs on her, she barely had enough fuel to get to New York as it was 😅
What's so interesting to me is how the forefront of so many technologies came together in military applications in the cold war, and not just in the US. In Sweden Ericsson (of later cellphone fame) cut their teeth on the Saab Viggen's onboard computer.
It's been so long since Saab Gripen came out, I wonder what they are developing now.
SAAB is an acronym, and the correct correct way to spell it is with all caps.
Our government doesn't significantly invest in non-military technology sadly. A whole breakthrough in technology that could benefit society left to rot in secret in a useless fighter jet.
@@hueyiroquois3839 I just went to their website to check, and on the main splash page it's spelled twice, once as SAAB, then again as Saab. So clearly they don't care, and neither should you
I tought DATASAAB was a part of Saab before going independent and them at some point being swallowed by Ericsson when they wanted to grow from telecom into computers.
Anyway. Data Saab. CK37. First flight computer with ICs. Very early 60’s. Also a 20-bit computer. Already a veteran by the time the F14 came out. Also lasted into the 2000s on the Viggen. Not microprocessor, but several years before, and in a small european country. Heck. The Brits make movies about how they discovered a way to crack the Enigma code, in Sweden, some guy broke it 2 years previously while riding his bike to work.
I love small UA-camrs who put in the effort and research. I know the work it takes when doing it alone. Great video, Alex! You now have another sub.
Thanks! Yep it’s a lot of work but I really enjoy making these videos.
This content is much appreciated , this officially pushes my level of F-14 geek to that next level that my old ride or die will never surpass . Cheers M8t!!!
Iphones are inferior to 70s cgi
The intel 4004 was mostly limited by the requirement that the CPU had to use that package, limiting the pin count.
A 4 bit CPU is pretty limiting and that was the only thing they were getting in that package in 1971. They were mostly used in scientific calculators where they performed with mediocrity. But when my dad was doing his master's, those calculators kept him from having to go over to the mainframe and futz around with punch cards so much.
high aspect ratio doesn't create more drag, it actually creates less at subsonic speeds, which is why gliders have high AR. At supersonic speeds you have shocks and delta becomes favorable
I really liked the video though, keep up the good work!
please upvote pilot channel's explanation. I was about to write the same thing.
I think there's some subtlety to this. High aspect ratio wings *do* create more drag overall, as a simple matter of increasing the frontal area of the aircraft. What they excel at is improving the lift:drag ratio. All other things being equal, a subsonic aircraft with shorter wings will experience less drag, but it will *also* be less efficient and have less endurance than the same aircraft with much longer wings. If we imagine that we change nothing about the aircraft aside from the length of the wings (i.e. we keep the same engine, propeller, fuel, etc.), then the aircraft with shorter wings will be able to achieve a higher top speed, but the aircraft with longer wings will be more fuel-efficient and have a higher service ceiling.
Compare the Lockheed U-2 and the Corvair F-106. They both use the same engine, a Pratt & Whitney J75, and they have comparable gross weights. The most obvious difference is that the F-106 employs a delta wing, and the U-2 employs a high-aspect wing. The F-106 has a top speed of 1325kt, while the U-2 has a top speed of 412kt. The F-106 has a L:D ratio of 12.1 at subsonic speeds, while the U-2 has a L:D ratio of 25.6. The F-106 has a service ceiling of 57,000 feet, while the U-2 has a service ceiling which is still classified but is *at least* 80,000 feet. The F-106 has a range of 500nm, while the U-2 has a range of over 6,000nm, a difference which cannot be adequately explained by the fact that the U-2 can carry ~10,000 more pounds of fuel.
@@tissuepaper9962 frontal area alone is misleading, especially at lower airspeeds. And you can massively increase Aspect Ratio of a given design, without changing the Frontal area at all, or even reduce the frontal area at the same time.
@@tissuepaper9962 in aircraft aerodynamics, we typically consider the projected or wetted surface area, not the frontal area. You can keep this area the same while increasing aspect ratio to reduce your drag.
@@tissuepaper9962 High aspect ratio wings only create more drag overall supersonic (and very high transonic). The reduction in induced drag when subsonic *more* than makes up for the increased frontal area, and thus a high aspect wing is lower drag when subsonic.
Consider: a modern airliner at basically all phases of flight always is making lift equal to weight, so lift is fundamentally fixed. If you improve the lift to drag ratio, you aren't making more lift, because your lift still has to equal weight, so what are you doing? Decreasing drag. During cruise, there's fundamentally no difference between "improving lift to drag ratio" and "reducing drag", because lift is unchanging.
Top gun:
A big part of its popularity was how well they meshed the music with the flight sequences
It was a music video with story cobbled in... who cares, Danger Zone!
Yeah.. There were parts of that movie that were cringe.
The flight nerds loved it because of the air action with the music. Every guy identified with going up to 30k feet with his homeboy and kicking that smug prick Iceman and his RIO's ass.
- That's two "O"s in Goose, boys.
The music was also what made the Iron Eagle franchise palatable enough for multiple sequels
Just takes a sharp editor with some music skills. But it was a lot more time consuming back in those days.
@@TheMDHoover Recruitment film really
28:04 that Horner scheme rearrangement actually increases calculation precision and significantly decreases the number of multiplications required. It is still valid and truly useful algorithm till today. Can be even used to make manually evaluating polynomials much easier.
Some friendly corrections: (may border on the seriously pedantic, your call)
5:03 It has been known since the 1950's that the most highly manuverable craft are often somewhat inherently unstable. Nothing controversial to anyone knowlegable there.
7:38 Variable wing geometry is still in active development and is not obsolete. Swing wings are obsolete. There is a distinction with a difference.
10:18 IIRC, The computer controlled the reaction of the rest of the flight control surfaces to the extension and retraction more than the speed of actuation of the swingwings and vanes. It was required to control these so that the compensation worked correctly, as you noted earlier where manual actuation was still CADC controlled.
17:45 The stance that a "rEaL" microprocessor must have onboard rom, ram, IO, is farsical. My 6-month old top of the line pc has seperate IO chips, Ram chips, Rom chip, and a co-processor cars with all 4 onboard. So only SoC would count, which is not in the same realm. I think you just wasted time here. Just muddled the whole topic here, although explaing the rom etc was nice.
23:45 Not a correction, just think you should do a video on the "Four Phase" as I have never heard of it and can find little info. There are already many great and exhaustive videos on the 4004.
30:22 I don't get why that would be harder to program for, and you made no justification for this statement.
40:33 Machine code translates to binary with no real compilation, it's not a 'language' per se, but a 1:1 human readable formatting of binary commands. Which is to say they totally wrote it in assembly and that was just typed in as binary. Your image at 48:00 shows that in the context of the 4004.
45:30 Punch cards, and even magnetic tape were available then, but tape was the most common interchange format. At the time, data interchange between dissimalr systems was quite rare.
52:40 I know where there are more about in the USA, but for security reasons, I'll leave it at that. An enterprising hobbiest might be able to get their hands on at least a partial set. If I find some i'd invite der8auer to the states to see them and make a video (export restrictions invariably exist). I'm confident the iranians have already just used an FPGA to copy them, as most countries do with stolen digital tech (see: Russia and China using FPGA copies of us missle guidance computers, there are videos on this).
54:22 There were more. Some are still classified, which should make obvious their purpose. Others were just corporate dead-ends, although not all technologically so.
All that said, you did an excelent job for a non-subject matter expert. Big respect. Got my sub.
Thanks, yes as you may have gathered by my tone, I do indeed think defining a microprocessor as essentially a SOC is indeed farcical: unfortunately there seem to be many that take that stance (for some reason). I didn't really enjoy having to be pedantic there but I was trying to make clear that 'if the 4004 is a microprocessor, then the MP944 is too'.
Anything I say about coding in assembly/machine code should probably come with a disclaimer that I am about 40 years removed from that world.
Btw, it was me saying supersonic aircraft are 'ugly' rather than 'unstable' which (rightly) made people mad!
The F-14 is not a inherently unstable aircraft, in fact, quite the opposite. Maybe if you compare it to something like a Cessna 172. The F-14 is especially stable to the point that even the famous flat spin is difficult to actually achieve and enter. Less than 10 airframes were lost in them. The F-14 actually is designed for stability because it lacks fly by wire or computer control over the primary maneuvering surfaces. As the wings sweep back, the center of lift moves aft, increasing the distance between the CG and the CoL. This was actually the reason for the glove vanes on the early F-14A's. They kept the CoL further forward, reduced mach tuck drag, and allowed for 7.5G maneuvering at any speed the F-14 could achieve.
I'd argue that modern wing geometry hasn't made variable swept wings obsolete, but certainly has bridged the gap between them. A true delta will always have less drag than the hybrid delta of something like the F-22. A true straight wing will always have better leverage against the CG for maneuvering and provide more lift than a hybrid delta. The variable wing sweep allows for a greater optimization across the entire flight envelop. Aircraft like the F-14 had an impressive turn radius, turn rate, and speed relative to the thrust. I say that because the TF-30 equipped F-14A weighed significantly more than the F-15 did and had less thrust, and yet the top speed difference between them is not that great.
The question is whether or not the weight and complexity or a variable wing sweep justify the performance gained over a hybrid delta, and that is where a variable wing sweep tends to fall behind. The other aspect is the difficulty in stealth. No one has made a modern variable wing sweep aircraft using modern low observability or composite technology. It may very well still present enough advantages, for the right aircraft, to justify its weight... but no one has really done it.
@@Whiskey11Gamingwrt swing wing, I think the real reason the Navy and Air Force don't use it anymore is because they can acheive "good enough" without having to maintain a swing wing system. Plus it adds weight, reducing fuel consumption. I think they found it just isn't worth the cost.
Formally, it's not wrong to refer to an assembler as a compiler. The definition of a compiler is basically just a translator for programs. An assembler qualifies, even if it's a lot simpler than we usually think of. As long as we're being pedantic. :)
@@Noisy_Cricket I concur, basically today they can beat the aerodynamic limitations with the force of sufficiently powerful contemporary engines to achieve acceptable results.
Hi! Still watching the video for the first time, wanted to thank you for your conscientiousness in using real captions instead of the auto-generated ones. That said, there's a few typos, maybe have someone read it over? Thanks!
Right, they should be in better shape now. Sorry about that.
I never thought I would casually watch an almost one-hour-long video about a 1970s processor. Loved the video! Really well explained and interestingly made
I found this entire video fascinating. Your explanations of definitions, concepts and use cases were excellent. My degree is in Digital Microprocessor and Telecommunications Technology, and included not just the basics of all electronics and simple digital circuits, but also integration into larger and larger I.Cs. Really took me back to those days of writing out machine code with pencil and paper, then hand-coding and old 6502 (8-bit) microprocessor in hexadecimal. And to think, they were handling 20 bits, 20 years earlier than that. Awesome stuff!
Well, they were handing many more bits even earlier than that, just not with microprocessors. The mid-60s CDC 6000 series were 60-bit machines. And the IBM 701, released in 1952, used 36-bit words.
I cant remember the last time I watched something this long, understanding next to nothing, still being captivated to the very end, even rewinding to rewatch parts I understood a little of. I absolutely loved this content! Instant subscriber. There is very little content on yt that deals with aeroplanes on this technical level. I would love to see what this channel would do with the subject of ”Centralkalkylator 37” the flightcomputer in the Saab Viggen.
Big thanks!
A few viewers have asked about the Viggen computer. I’ll add it to my ‘potential videos’ list!
17:50: As I recall, there are a number of "microprocessors" that are an aggregate of several chips. The 8080 "microprocessor" being one that comes to mind. Intel quite clearly refers to it as a "microprocessor" on their datasheets, but requires 8224 and 8228 to build a system around. Nomenclature wars anyone?
Generally, a microprocessor contains the ALU, registers, and the control logic to fetch instructions and send data along the correct path.
I/O is usually included on modern microprocessors too, but generally interrupt controllers and such are not. A microcontroller is designed to be an all in one system containing ROM and RAM, along with all I/O. By putting these in one chip you can put computers in smaller devices at the cost of versatility of the system. There can be large debate as to which extensions would be required, such as floating point and multiplication for a microprocessor. The 8086 up to I think the 80286 required separate floating point co processors or FPUs such as the 8087. And x86 processors are what we still use today in most desktops. There were versions of the 386 I believe with and without the x87 floating point extensions.
To say that we went from assembling computers transistor by transistor to microprocessors is very misleading as stated in the video. Computers before microprocessors were assembled by logic gates, as the 7400 series chips were released in the '60s. Then computers were assembled with more complex chips containing ROM, RAM, control, registers, and full ALUs in single chips. Just how we don't see the reason to combine our CPUs with SSDs and RAM now in our computers it wasn't very worth it to do it with the mainframe computers at the time. The technology jump from having separate control, registers, and ALU chips to one chip is smaller than most people make it out to be. There were much bigger jumps such as when the original 7400 chip was released, putting multiple transistors in one chip. There were also large technology leaps such as the PLA and other predecessors to the FPGA, allowing us to "program" a chip to behave like any logic chip we want to.
You are pretty much correct there. Actually, you could build simple microcomputer systems without the 8228 System Controller, though you lose easy access to certain signals such as whether you're accessing memory or I/O address space. (That removes a useful but not necessary feature: you can always use memory mapped I/O.) But yes, without the 8224 Clock Generator, you'd need other equivalent external circuitry to generate the unusual asymmetric ϕ1 and ϕ2 clocks that the 8080 requires.
I was in the Army and was a mechanic for the 1980's light tank and I was impressed by the components, open one up and it was filled with multiple universal cards, so we could essentially lose a few cards(around 4 if I remember correctly) before the whole box needed replacing.
Another fascinating and terrifically researched piece. Great video. Thank you!
You did an excellent job describing this fascinating element of the F-14 I have never seen investigated in detail like this.
I was a little kid when the top gun movie came out and was absolutely enthralled with the F-14 because of that film. Approaching 50 now and I completely forget at times how far the hardware for computing has changed and how revolutionary an aircraft with chips like these were.
I started in college with electrical engineering but changed to a degree in information systems, but your talk about this brought flash backs of my coming to the conclusion I was never going to be an electrical engineer 😂. Really nice video on an interesting subject.
14:00 an FPGA is not a microprocessor - it's a collection of logic that can be configured for different arbitary logic functions, possibly including one or more microprocessors
Oh look! UA-cam electronics royalty spotted!
I think he was pointing out how loose the term was back and any sort of chip that could do arbitrary logic whether that be in hardware or software could have been considered a micro processor by old parlance
It is not inherently, But an FPGA can be, and often is a microprocessor. /pedant.
But you can make it into one
This is very interesting. I was a structural design engineer on the F-14 program for a few years in the early 1980s. Even as I contributed my small part, it felt like i was chipping rocks compared to the stuff the guys in other parts of our building, and in other buildings, were doing.
Still, it's a long way from actually chipping rocks! That's what civil engineers do. (Mechanical engineers build weapons; civil engineers build targets.)
Excellent video. I've greatly enjoyed these long-form deep dives into engineering relating to aerospace.
As an aside, 2024 is supposedly the last year for the F-14. Iran is retiring them in favour of Su-35s (or a similar Flanker; I forget exactly).
Oh that's interesting. It amazes me they managed to keep them flying for so long.
Before maverick steals one.
How are they going to get SU-35s when Russia only has a few dozen?
@@lazerusmfhMy thoughts exactly.
@lazerusmfh they're probably never gonna get them
I do not know how this got in my recommended feed.
But I’m very thankful for it.
Thanks for a great college on this matter!
0:01 no it's actually 2024 I think you are confused.
This is the best Ive found so far for computer tech back in the 60s, 70s, 80s. The low level programing needed to control swing wing on the F14 Tomcat.
Another reason for a nerd like me to love the Tomcat. What a fantastic video. Thank you!
They started teaching computer architecture in A school in the 80s. They had Altair looking computers where you would actually hand enter opcodes by hitting switches (they were Octal - so base 8) and pressing a button to enter the value into a register or memory. You had to basically hand enter the program one opcode after another. The press another button to execute the program and you crossed your fingers it would end up in the correct configuration after running your program :D. I, of course, had already taught myself Assembler using DOS 3.2 and the DEBUG program when I was 14, so that unit was a breeze for me. A lot of people struggled with that unit, but I was lucky. Ironically I taught myself Assembler because I wanted to get closer to the "metal". When I discovered PCs I was in an electronics phase, so for me it was the EE part that was interesting. I eventually became a full time programmer / software engineer :D
@6:26 Being from that era and from the US, I can speak to why it was so popular: it was so well received because, in addition to being just a spectacularly performed/shot/edited/etc. movie, it also served as a desperately needed reassurance 5 very long years before the Soviet Union would collapse.
Every single day that the USSR was in existence came the renewed threat of nuclear annihilation. Many people very reasonably assumed that the only way we'd know the winner of WWIII would be for it to play out, which of course meant that very few of us would be alive to witness it, and many of those would find it very difficult to want to go on. Imagine the threat of being forced to live the IRL version of a Fallout game, where all the enemies are invisible, and the majority of them eat you by slowly and torturously melting you from the inside out.
That was our day-to-day. Every day we knew that the sirens might come up, and we'd have very little time to say our goodbyes let alone make sure we could all be together with our loved ones at the end.
Top Gun showed another way. It showed that we could fight the war in miniature, and if the USSR wanted anything worth conquering, it would have to take it the old fashioned way.
And this was not without its merits. You can't grow crops on irradiated soil, and when Top Gun came out, only a month or so prior, Chernobyl had given the world but especially the Soviets a hard look at what the consequences of a full scale conflict would be.
And Completely unbeknownst to all of us, the military technological capabilities of the Soviet union had been dropping behind that of the US for several years, and the USSR had been working very hard to hide that fact.
When that very brave Russian officer defected with his aircraft on September 6, 1976 to the Western allies, the United States was able to dissect the MiG and learn its weaknesses. From that came the F14, and as far as I know, by the time top gun came out, the Russians knew that the movie represented no idle threat.
Where the communists had kicked our asses in Vietnam and Korea due to our lack of dog fighting skills, the F14 was specifically designed to be able to get to conflict very quickly using the minimal amount of fuel, and then have the aerodynamic capability to outperform and outturn it's opponents in a dogfight, with more fuel to work with in an engagement.
And that's not all. Not too long after the F14 came out, the United States brought out the F-16, and then the F-18. The Russians simply couldn't keep up, and they realized that these aircraft and the force stacking capability they represented made an impenetrable wall against their conventional military tactical ideas.
That's what top gun showed the American public: that the Russians could be beat, and that maybe we might all live to see the coming millennium.
VERY excellent video, and please believe me I don't blame you for not understanding why it was so popular. I'm really glad that you (likely) never had to live with that daily dread.
See, here is the thing though... I was born in 1988 and I saw this movie in 1994-1995 at a babysitter... I DISTINCTLY remember the movie and it began what has been a nearly 30 year obsession with the F-14 Tomcat. I was extremely disappointed that when I graduated from High School, the F-14 had been retired. There were no threats of nuclear annihilation in the post Soviet era when I saw the movie, and yet, I still love the movie today because it, and The Final Countdown, are two movies just chock full of F-14 porn... this super sexy high speed extra lethal display of US Military Might. I only began to really appreciate the things the F-14 accomplished as an adult... the F-14 was the F-35 of the 1970's... it defined what it meant to be a fourth generation fighter, and for nearly two decades it was the only fighter in US service which could fire multiple radar guided missiles at different targets and have it hit.
Or because this UA-camr is tone deaf and missed out on the jingoistic and nationalism dripping out of Top Gun during the Reagan era and its unique concept and cinematography. Everybody wants to think they’re the good guys. Just look at us justifying several generations of ethnic cleansing in Palestine.
"When that very brave Russian officer defected with his aircraft on September 6, 1976 to the Western allies, the United States was able to dissect the MiG and learn its weaknesses. From that came the F14.... "
Uh,.... Tomcats flew CAP over the evacuation of Saigon in 1975. Their development started long before Belenko flew his Mig-25 to Japan.
@@thatguy6054if anything, the Foxbat was one of the reasons the USAF got spooked and spurred them to make the F-15
The Belenko incident basically washed away all notion that the MiG-25 was a very capable aircraft, where infact it was a VERY, VERY chunky interceptor that was powered by engines that were derived from a cruise missile.
so... the movie was propaganda?
Thank you youtube for recomending me this MASTERPIECE of a channel. There is not a single other channel that goes into as much detail on stuff like this besides yours. I have looked for ages to find information like this, and have never found it. The video on the AGC was also so unique. This is so amazingly accessible and well written, bravo :D
Great video. In the past I have been involved in two small processor designs.. a silly one where I had an evaluation copy of a system level design tool and a 150 cell CPLD and an EPROM.. and another that sold in the millions , the de emphasis digital audio filter in a Philips CD player chip. It had to do several multiplications and some additions all in 24 clock cycles. It was bit serial and used modified Booths algorithm multiplication .. two bits per clock cycle on 18 bit data.. 18 bits because it began life in a Digital Compact Cassette player..
The internet is a small place! Hi Mike!
High, can you point me in the right direction to develop a chip for a small drone? It will be the basis for AI for the Ukrainian military
@@martinwhitaker5096 it sure is!
Wow. What a dense summary. Well done. It’s very difficult to communicate a complex concept that you have very little information on. I appreciate the skill shown here.
That intro to the F14 was awesome! top work.
Most of this went well over my head but watching was well worth the time regardless. Your commentary only gets better with each video and I appreciate your willingness to maintain a degree of silence with limited music and other background noise. You also refrained from saying the F-4 is ugly even while it featured in the video :)
The flaps and slats were interconnected and never moved individually, if any one of them got out of position by more than 2 degrees from the others, the system would lock out in that position. There is a set of position sensors at the end of each flap and slat, when the aircraft was put under high G the computer would deploy the flaps and slats a few degrees to increase turn ability. The inboard flaps only functioned when the flap selector handle was actuated, they were either up or down. If the flaps were selected down the wings had to be in the forward 20-degree position for the inboards to work, they were hydraulic actuated and would blow back up if the airplane went over a certain airspeed. If flaps were deployed at all the wings would only sweep to about 50 degrees. When wings were forward, roll was assisted by spoilers when the wings swept back past a certain point the spoilers failed down and they were locked out, roll was by the stabilators only at that point. Grumman developed the F10F Jaguar in the 50's with swing wings, I think they were manual only, but the company was a prime subcontractor on the F-111, bringing their swing-wing experience to the project, the CADC was probably developed for it and refined for the F-14, remember that the SECDEF at that time wanted the Navy and Air Force to have the same fighter/ bomber since the F-4 worked for both services. The F-111 sucked for the Navy for a number of reasons and Grumman left the program and developed when the Navy initiated the FTX program. Politics killed the cat.
You brought up facts ive never heard mentioned in any documentary on the F-14. Thank you!
For someone with a limited knowledge about computer systems but a big sympathy for the F14 I enjoyed this video very much and leard a lot. Thank you.
Great video! I really wish we would have a program today that could be used to verify that every transistor in the system is actually working correctly. Even if running that program once took a whole day, it would still be highly valuable to verify the hardware that you have, either to verify it after building a new system or to diagnose unstable system. Currently the state of art seems to be "randomly swap parts until it starts working again for the programs that you actually use daily".
Programs can be written and the trick is to be able to prove that a "single failure" is picked up. Overall its a simple concept. If a * b = c and if c is the incorrect answer then you have a problem. How many a * b or a/b do you have to do to catch all single failures. Now multiple failures are a different story. Ray Holt
@@firstmicro The problem is that you cannot know by definition which part of the hardware is potentially broken. If you *compare* any numbers to each other, what if the comparision operation in the CPU is broken? Modern CPUs have multiple execution units per core so even if you successfully execute 5 comparisions correctly with known values, you cannot know if the 6th execution would actually work correctly or not. And since the actual execution units are selected by proprieatry algorithms inside the CPU, you cannot assume you have a safe way to run any kind of test code on all units.
We would need some hardware support to be able to test all the transistors in the CPU. For exaple, ability to select which execution units are in use (basically disable everything but one) is needed to check if those units work correctly.
And to make things harder, the L1 or L2 cache could be broken and there you have the problem that e.g. 16-way associate cache can have the cached value in different locations in the cache and any random execution of the code may or may use the part of the cache that's broken.
It's turtles all the way down and at best you can give some kind of statistical estimate how correctly the hardware is working.
You are making some amazing videos, I’m really loving the long format content about these topics.
Stunning! Early computer history suffers from being stored on media (paper and restricted scanned images) not widely viewed by current students so these videos expose new generations to how computer fundamentals came about. Trouble shooting these types of systems is very tricky (and not always successful). I worked on directly connected adjacent parts to the F14 and to circuit board construction is just jaw dropping. The printed wiring boards are super high density and battle hardened. The difficulties in programming are unbelievable. The original engineers were amazing.
For someone who doesn’t like Top Gun, you sure picked an interesting font 😅
This video is great! This topic just came up during a family dinner. My brother and I were huge nerds for the F-14 and other aircraft of that era.
Thank you for this!
You spotted the font easter egg! Congratulations!
Ok. I have an F-14 adjacent and somewhat funny sort of story. I was an enlisted onboard the USS America CV-66 from 1975-1978. I worked in CVIC, the intelligence center. We had cypher locks on the main door entry and one of us enlisted guys would have to do a four hour watch checking IDs and "buzzing" people in through the door, unless they had approved access and knew the combination. CVIC had a main entry that was connected to a large debriefing room for the pilots. Other functions were behind another cypher locked door. During Med Cruises F-14 pilots would come down off the flight deck for debriefing while in full gear and we'd buzz them in without having them show us their military ID. Invariably our boss would periodically insist that we had to follow the guidelines and make everyone show us their military ID. The look on the faces of these pilots when told we needed to see their ID wasn't so much as awkward as just plain pissed. They would be carrying their helmets and possibly some other gear. They would then struggle to dig out their ID from whatever inner pocket it was in. Eventually, they'd complain to their CO who'd complain to our CO and we'd be allowed to relax the requirement for pilots. And of course after three or four months the instructions from on-high would be that everyone had to show ID. Needless to say, I always thought it was a really stupid requirement. These guys are flying F-14s, over the Mediterranean Sea. Exactly how likely was it that their plane, mid-air, would be commandeered, they'd be stripped of their uniforms, tossed out at speed (I guess), the enemy would dress in their uniform and gear, then land an F-14 on a US Carrier, then come down to CVIC and expect access. I always figured that if a Russian could do that then we'd already lost the Cold War anyway and we should give them access for their debrief.
One thing, I think, you missed that was controlled by the flight computer were the variable ducted engine intakes that changed their profiles as the engines went from subsonic flight into mach speeds. I seem to remember one of the F14 WSO/RIOs talking about this...and in the early engine mods could cause a flameout under the right conditions.
Yeam, he implied al lthe computer controlled was the wings and vanes which was far from factual, although I don't think he meant to.
The CADC didn’t control the engine intake ducts - they were managed by a separate In dependant system called the AICS (Air Inlet Control System)
Those were controlled by a different computer. Probably because the engines have a different manufacturer than the airframe, and existed independent of the F-14. So, use existing means to control the ramps, or add more cost and complication to a CADC.
@@pgtmr2713 The F-14's air intake geometry is built into the engine nacelles and air intakes... the control of the actual variable intake geometry was a Grumman design specifically to allow the TF30's to operate in the designed speed regimes. When the GE F110 was installed in the F-14B and F-14D, the scheduling of the intake ramps was changed to optimize performance in the transonic region. The end result was the TF30 actually produced more thrust at low altitude and at high mach (not necessarily both). Air stacking in the intake allowed the TF30 F-14A to go to Mach 2.34, the GE F110's were capped at 1.88.
You are thinking of the F15. The F14's version of this was entirely through discrete logic. The F15 was what brought it digital.
Interesting video. Reminds me of the seventies when I started designing embedded industrial control systems.
Back then we made the distinction between mini computers and micro computers. A one chip CPU made it much easier to build a system compared to building all with TTL circuits.
Mike, Chips like the 4004 required multiple TTL circuits. Some one-chip microcontrollers were very capable and required very few TTL circuits.
You really should send this video to Ward Carroll, former F14 Rio. He would be interested in seeing this, and might even do a crossover video.
Great video, brought back memories. I graduated with a computer science degree in 1978, and in 1979-1981 I worked for Marconi Avionics, Rochester, Kent. I worked on the first computer based air data systems for Altitude and 2 dimensional speed for helicopters. There were still a team of watchmakers who serviced the mechanical computers for the Vulcan bombers.
The glove veins were not even built on versions of the f14 past 1980s. The reason was they did not have any large effect on maneuverability for the cost, so they stopped making them.
Found your channel through the algorithm. Great stuff. Thank you for creating.
I have only watched a few videos on processors, but I intend to binge when I have to time. :)
While in the US Navy for 20 years, 1970 to 1990, one of my main jobs, was to work on F-14 flight control actuators. One thing I was taught to do at a special instruction course was the wing sweep actuators and glove vane maint.. I could check and test the sweep actuators but was not allowed to go into them. They had to be sent to the Depot in Norfolk , Virginia for a complete overhaul. I was asked to do it several times, but I was smart enough to refuse. If the ship was in port ( USS Nimitz) I could get the paper work and hand carry to the depot that was a couple of miles down the road. I would want to see what the technicians there would find, particularly if we had a run of bad ones. Some times it would be bad aircraft wiring, a bad cannon plug or even that ADC, and not the actuator. It made for some good arguments. I spent 10 years in Squadrons, and 10 years in Aircraft Intermediate Maintenance Departments (AIMD) I also spent 5 years after I retired doing F-14 structural repair at the Norfolk , Virginia Naval Aviation Depot, and I've just retired from 20 years at the Coast Guard Depot at Elizabeth City, North Carolina.
By the 80's all the CADC's would have been shipped to Grumman and most likely supported there. Ray Holt
This was fascinating. It included the two topics I am fascinated by! Jets and Microprocessors!
I actually heard about this growing up, looked online for more information, didn’t find anything, and assumed I just was mistaken. Thanks for the interesting video!
The Swedish AJ37 Viggen had a computer CK37. Viggen entered service 1971, 3 years before F-14.
Computer was built with ICs, but those parts might not qualify as a microprocessor.
While filming the USS Lexington with my drone(and at every Museum ship for that matter lol) I toured the museum for a whole day. If memory serves, there is a big display talking about this subject but in a manner for all to understand. Its really cool having this video to really flesh out the subject. I remember they had other examples of early computers and so on. Support your museum ships people! ;)
I wanted to add that the CADC hardware prototype ran at full-speed of 375KHz. It was not subjected to mil-spec. The purpose was to simulate the hardware and the programming.
The more I learn about computers the more surprised I am they actually work. Modern pcs are nothing short of a miracle of engineering.
“It was in Top Gun… you know, the film about volleyball” received the biggest belly laugh from me!
Cheers, mate!
Your click picture was taken on my ship the USS America CV66. Carrier Air wing commander Mike Snodgrass was performing a fly by. Thanks for the memory.
If anyone thinks you're a weirdo Alexander, it's because you're making 1 hour long videos about the F-14 CADC, not because you were in a UAS! Great video, can't imagine the effort that's gone into this for our benefit.
Ha was going to caption that bit as ‘a weirdo but not that type of weirdo’.
To clarify, I hd an awesome time in the UAS and am proud of what I did during that time. I just don’t want to come across as one of those people that pretends they were in the actual military.
I like how someone had an idea, used the materials available and created a proof of concept. Now that 'it' works and is something useful, our ability to refine it changed the world.
Ah nice, looking forward to this after work.
I worked on the F14 in 1994. There was the Z8080 embedded in a center console black box. I had an old programming manual from the 70s and was able to work on it.
Great video. A couple of niggles (I know you love them): iirc the program which converts assembly language into machine code is called an assembler and not a compiler. I’m sure I’m like the 100th pedant to point this out. And between the 4004 and 6502/Z80 were the 9900 (no registers, first consumer “16-bit” chip), the 6800 (inspired the 6502) and the 8080, among others.
Oh good spot! Why did I say ‘compiler’!?
Ah yes I just mentioned the 6502/z80 as they will be more familiar to viewers
Also the 8008.
In the 60's, IBM was working on Josephson circuits, multiple superfast processors in a liquid-nitrogen bath (to lower electrical resistance to near-zero). Those Josephson processors/circuits were microcircuitry prototyping.
You make amazing video's, keep it up!
Architecturally this feels more like the progenitor of the modern DSP, a truely spectacular design, thanks for sharing.
What is a microprocessor? In reality it really does not matter, however, companies and marketing people love to label what they have. When I started the design of the F14 CADC the implementation of the “processor” was unknown. As we tended towards the new “integrated circuit (IC)” technology names such as IC processor or LSI (large scale integration) was used. Micro was also in there somewhere as the computer design evolution progressed from large, to mini, to micro. It was not a big issue to anyone. Ultimately, as the CADC project was completed and I wrote my paper I chose the name “microprocessor”. There was no focus on single-chip, multi-chip, program counter location, etc. In fact, even Turing was set aside as the main focus was to do the job irrespective of # of chips and what function was where, and even how data was transferred. The selection of any cleaver design techniques (pipelining, co-processing, in-flight self-test, dual redundancy), algorithms, and package selections we made to reach the speed and accuracy necessary to fly the F-14 as needed. It was unfortunate (for me and history) that the 1970 presentation of my paper and any patents were strongly refused by the US Navy thus 30 years of silence started. Around three years later the Intel marketing group started labeling their work. First came the World’s first microprocessor, then the World’s first commercial microprocessor, then Intel’s first microprocessor, then Intel’s single-chip microprocessor …. Today, everybody wants their design to be a microprocessor, otherwise, how are they going to out market Intel?
In 1973, Hank Smith, Microprocessor Marketing Manager, Intel Corp made this statement at the IEEE 1973 WESCON Professional Program Session 11. “A CPU uses P-channel MOS and is contained in 1,2,3 or4 LSI standard dual-in-line packages from 16-42 pins per package”. Even at this time the word CPU was sometimes used for the entire processor and not microprocessor.
After this, history started dissection the word CPU, microprocessor, single chip, etc to fit the functionality, design, and packaging of each product. Thus, here we are in 2024 (54 years later) still trying to make it all fit into convenient definitions. Ray Holt
Is this actually Ray Holt?
@@joc575 It's me.
This just makes one think about the advancements that are out there at the moment that we the general public don’t have access too. Hard to believe how ahead of their times those engineers were with what they could accomplish
The use of microcomputers in the F-14 to command aircraft flight controls closely follows the terrain following radar controlled flight of the F-111 fighter bomber, which Defense secretary McNamara infamously foisted on the Navy as a do-everything fighter/bomber.
As a first generation swing Wing supersonic fighter the F-111 had huge teething problems and fell far short of the Navy's requirements, however, Grumman was well aware of every outstanding attribute or shortcoming of the F-111 in designing the F-14 soon after as a superior replacement.
And even before the F-111, the Apollo moon rockets used some of the first integrated circuits - I believe they were Fairchild microprocessors? - in their rocket guidance command and control.
Even more astounding were the analog computers used in World War II, especially the big firing control units that estimated range & velocity for the huge Naval guns on the US battleships, and the far smaller Fire Control computers for the remote gun turrets on the B-29 Superfortress bombers, which apparently were accurate to twice the distance of
purely human aimed gun turrets.
It is amazing to think that the Concorde, the 747, the X-15, the SR-71 and Appollo rockets we're all designed by slide rule in the days before engineers had electronic calculators.
And finally, it is a shame that another defense secretary mafia, this time under dick cheney, pulled the rug out from under the mighty F-14, instead of upgrading it with the newer engines and avionics they deserved, forcing the Navy to adopt the smaller, less capable but less expensive F/A18 Hornets
The MP944's architecture was ahead of its time in terms of parallel processing and pipelining. It's amazing that they were implementing these concepts back in 1970 for the F-14's CADC.
Dang, I never would've imagined anyone would think military aircraft were scary/ugly. They make me proud of human ingenuity, and that's why they're beautiful to me - they're feats of man's wrestling with the forces of nature, and one another.
You havent seen the X-32
The intro (transition to the f14) is really well done, i love it
Sir, with great respect, while your comment about high aspect wing ratios creating more drag was not necessarily incorrect, high aspect ratio wings actually have a higher lift to drag ratio. That is to say, of a high aspect ratio wing and a low aspect ratio wing of equal lift coefficients, the high aspect ratio wing will usually produce less drag (case in point the high aspect ratio wings of the U2 or a modern airliner being designed for maximum efficiency while lifting a heavy aircraft). It would be truer to say the Tomcat in the "T-pose" would create higher drag because it has an increased wingspan, to gain additional lift at low speeds. Love the video though!
I am endlessly fascinated by mechanical computers for some reason. They’re just such a seemingly crazy blend of technology and discipline in design and construction.
You may want to check out the Viatron System 21 microprocessor from 1968.
I did come across that as part of the research for this. It was the first use of the term 'microprocessor'...though I don't really think it falls under the generally accepted definition of one today (though, of course that is very much open to debate).
And four-phase al1
That 10 year period of advancement from the 60s to 70s is just mind blowing
Top gun was popular because it was Brad Pit.
Wow - I learned from you that the F14's swing wing didn't require straight and level flight to change - that's impressive! I'd definitely previously made the same assumption you formerly had, and your video was perfect to undeniably emphasise the impressiveness of this by video of the F14 rapidly changing the swing wing at 8:26 and 8:40 (and introduced a bit at 7:40 too). Awesome! Thanks for your video =)
Hey! You're back.
Excellent video (speaking a SoC designer). If I can just point out one detail: the tool to convert assembly language to binary code is called an assembler, not a compiler. A compiler is for high-level languages like C and it produces assembly language. The confusion in recent times comes from the fact that the assembler has been rolled into the compiler. If you program at low level, you notice that when write assembly language within C (by using dedicated macros such as "asm()" or by including .s files in your projects).
I wrote a paper on this in college, was very interesting.
Wow. This was a fantastic analysis and video. I love all the amazing F-14 shots too! Great work!
Now we just need to convince heatblur to simulate the entirety of the CADC into the DCS F-14 Tomcat
You don’t understand how grateful I am for this video. I’ve always wondered how these things worked down to the damn atom. Do radars next.
"You couldn't for example build a practical desktop computer with a 4 bit processor". I can't see why not. The "4 bit" in the 4004 was the ALU and data path width between the four chips. The memory addresses were actually 12 bits and that's the far more important number. It's still a bit small, but all the "8 bit" microprocessors had 16 bit memory addresses and people did quite a lot with 8080/z80 CP/M machines, the Apple ][, the C64, the Atari 400/800. Those had 2^16 memory locations each of which was 8 bits in size, while the 4004's 4096 memory locations were only 4 bits in size. But a 4 bit computer with 17 or 18 or 20 bit addresses would be just fine, as long as it ran at sufficiently high MHz. Note also that the z80 has a 4 bit ALU, two 4 bit buses, and separate latches for the hi and lo 4 bit ALU results, which are produced in different clock cycles.
What surprises me about the 4004, looking back, is how many registers it has. There are 16 of them! So you can do quite a lot of calculations without hitting RAM at all. This is quite a stark contrast to for example the 6800 and 6502 which had three registers each for program variables (plus PC, stack pointer, and flags). Altogether the 4004 had 17 4-bit registers, 4 12-bit registers (PC plus 3 return stack levels), and a carry flag, for a total of 117 bits. The 6502 has 5 8-bit registers (A, X, Y, S, P) and a 16 bit PC, for a total of 56 bits, less than half as many register bits as the 4004!
You say addressing width is far more important, but bus width bring speed of operation. I'd say they're just different, except that when it comes to powerful computing like GPUs, it's often this data bus width that is the limiting factor of the architecture.
"You say addressing width is far more important, but bus width bring speed of operation." Yes, that's exactly what I said. Addressing width is fundamental. It determines the complexity of programs you can run. BUs width is merely speed, which can be compensated in any number of ways. “There is only one mistake that can be made in computer design that is difficult to recover from-not having enough address bits for memory addressing and memory management.” Bell and Strecker, 1976. @@VndNvwYvvSvv
It had good amount of registers but it was needed for addressing, fetching, arithmethic operations and narrow bus. The CPU lacked logical instructions and when I researched it it seemed it's hard to write more complex programs due to these limitations. It was really calculator on a chip designed for decadic calculations. Even the 4002 RAM chips were designed to store mantissa and exponent with sign in separate locations and you had to use different instructions to access them.
Wicked video. I am a child of the 80s, born the year after Top Gun released, so it was a big part of my childhood and I still rewatch it occasionally even now. I've always loved the F-14 as a result, and it was the main reason I clicked on the thumbnail for this video. My other minor interest in computing being the other reason. Never knew how much computing was going on inside those things. Makes a cool plane even cooler imo.
Always a shame that folks who make such incredible discoveries and put so much committed work into their projects get no recognition in their time due to classification. I get the need for it, no issue there, just it would be nice if they could be recognised for it when they did it, not decades later when the files get declassified coz they're no longer relevant.
Ingenuity is humanity's finest trait imo. We can figure out such incredible solutions to the problems that stand in our way. To the point where we can shoot human beings out of our own atmosphere atop massive engines and put boots on the ground of other celestial bodies. We're an amazing species when we put our minds to advancing stuff like this.
I have to know, did you plan this video’s release date with CuriousMarc finishing his series on a mechanical air data computer or was it just luck?