I had no idea just how thoroughly and meticulously, or difficult and lengthy this project is and has become. This is 100s of thousands of dollars of combined effort, skill, resource, and time. Pure unadulterated mutual love and concerted effort. I hope this AGC restoration gets the public acknowledgement that it deserves. It goes to show that the Apollo project was an effort rivaling any of 7 wonders of the world. I would be interested to see a comparison to the Russian technology of the same period.
Nicolaj Baramsky There’s several documentaries on the design and building of the AGC, filmed on location during the fabrication. I’d like to this in parallel and a combined montage.
Russian technology of the period was a music box with switches controlled by cams on a rotating drum. That's how our astronauts described it after seeing it aboard Soyuz in the Apollo-Soyuz mission in 1975.
I was four years old when Apollo 11 happened. I remember running outside that night to look at the moon and see if I could catch a glimpse of the astronauts as they did their thing. I finally met Armstrong a few years before he died - he was one of my heroes and one of the reasons I got into engineering in the first place!
Right. Monday. Let's get started work after a long weekend of children's birthday party. Motivated and ready. No distractions. Ooo! Another CuriousMarc AGC video!
It would be very interesting to see the explanation in a separate video. It's kind of besides the whole point to skip it. Thank you for your very, very exiting videoes. Such a team! Best regards,
Also - this entire team is so cool! Can y'all do a group photo at some point - maybe a round of brief (re)introductions as to who everyone is? Thanks again for making these videos; really really excellent!
The curiously named fixed fixed memory refers to page 0 in fixed memory (indeed from core rope) that does not need to be moved into view within the very restricted address space of the AGC by using the memory bank register (which is all the other pages, or the majority of fixed memory). Hence it’s doubly fixed. It would have been better named base bank or base page fixed memory, but it’s much less fun.
As a computer engineer, I've always appreciated the AGC. I am very glad you guys are working on one and sharing your experience with us. I also appreciate the very technical information as well.
I love watching this kind of Digital Archaeology, because that's what this is. Digging into our past through empirical observation, analysis of known information, and re-learning things that have long been forgotten. Totally love the direction this channel has been going. You guys are true specialists, and its awesome that someone is taking the time to preserve and resurrect such important parts of our history - so that we can continue to learn from it and avoid repeating past mistakes. Cheers.
The amazing thing is not the AGC itself, but all the engineering and work that went into creating it, double that, they had only the technology of the 1960s at hand!
I know so little about how computers actually work, yet I fund these videos accessible and easy to understand. The passion shown by the experts, their incredible knowledge, the ability to trouble shoot problems. Fascinating viewing. Thank you very much
Any time I start to feel too smart I just sit down and watch one of these videos. I know a little bit about consumer computer repair, but this is well above my pay grade! Thanks for taking the time to lay a little knowledge and understanding on me. It's kind of like visiting a museum and getting a personal guided tour.
I just now found this, amazing what you guys have done, Mike's abilities don't cease to amaze me, this guy is a machine!, Man I I'm a hardware engineer and been working on assembler and digital circuits most of my life and after seeing what Mike and all of you are doing here I feel like a moron, fantastic series, thank you so much for done that, too bad I was so late to the party
Marc, your channel is absolutely behind compare, you have no equal peers online, with such unpretentious, hardcore, unique and EXTREMELY interesting videos with VERY hard to come by machines and friends with in depth knowledge of them. Wow! God bless you and thank you so much. Matthew.
Wow guys I love it, kept up with most stuff until now but the rope memory is definitely a struggle for me. Reminds me of a bunch of my friends huddled around an 8080 development board trying to get some interface working. Wonderful to see real fundamental knowledge ( and old test gear ) at work
6:45 I'm so glad even you recognise understanding some parts in your videos can be challenging. I've followed the whole AGC series without any serious, in-depth computing knowledge at all, go figure my struggle at some points (though not as many as I expected at first). So finally I'm not feeling like the only idiot watching this, yay!
For me this is building up more hype than the game of thrones finale did. Even at 36 this is making me want to get into more electronics stuff. Thanks a ton.
I get excited to watch what you guys get up to with this project, What an incredible job you're doing. Thanks for these videos.... I can't wait till the next one.
All this time I thought Apollo 11 used a Saturn V rocket to get into space and it turns out they were using rope! The things you learn on UA-cam.... Any chance of a collaboration with a channel building a full-size Saturn V? :-) This is one of my favourite UA-cam channels. Keep up the good work!
It still blows my mind how physically difficult it was to make memory and other parts back then. It just goes to show how devoted people were to projects. Stuff got done, even if it had to be invented on the spot. And it's great to see a group of people that carry on this type of work.
same story with the F1 main engines on the Saturn 5, while the blueprints still exist, a whole lot of custom fitting had to be done by hand to actually make them work, and the notes on what was done there either was never written down, or lost. the team that looked up possibly recreating those engines actually found it easier to come up with a simplified cad model that brought the original engine's some 5000 parts down to about 55-60 pieces . simply because modern design and manufacturing techniques allow us to build parts with shapes and complexities that simply weren't remotely possible then
I'm right there with you guys in spirit sending you positive thoughts...to quote Oddball (Donald Sutherland) from Kelley's Heroes " Crazy... I mean like, so many positive waves... maybe we can't lose, you're on!"
I never thought the Apollo missions were faked, and these videos - beyond being totally awesome in their own right - are definitely confirmation. There is no way anyone would've gone to this much trouble just to shoot the whole thing on a soundstage. The point of fraud is to save time, effort, and money, after all.
@@reallyme3573 When you say, "A modern day phone with 8GB of data," you're talking about your data plan limit. What you MEAN to say, is storage space. which just means that your modern phone could fly to the moon and have plenty of room left over. Not a lot of calculations are needed, they used to do them on slide rulers, which takes a lot longer. Only a couple decades earlier, in the Manhattan Project, they were computing much more complicated equations using mechanical calculating machines larger than a typewriter before they got access to an IBM mainframe. Core memory actually was pretty fast for the 1960's compared to other forms of memory, it was just limited by the speed of the TTL logic. It was a 2MHz 16-bit computer, which is comparable to the first generation of home PCs only a decade later. I've still got a computer that isn't much faster, still doesn't have a math co-processor, and it does fractions just fine. You don't need a math co-processor to do fractions, and you never did. You just tell it to stop counting after so many decimal points, so it doesn't fill up the registers. Each sequence of the journey was pre-programmed into the ROM, and when you're programming in assembler, there's no bloat to worry about, so not much space is actually needed. It's been uploaded to GitHub, if you want to go over it: github.com/chrislgarry/Apollo-11/tree/master/Luminary099
The saga continues!! So exciting! Can we please have the 11 minute explanation in its own video? I'm curious about the fault and would like to know what he said... remember, we're nerds too!
@@mikestewart8928 It's very nice and clean indeed. The only thing I don't like is the libprussdrv library on the linux side, which is part of why I created py-uio[1], and hopefully one day I'll get around to also make a better C/C++ library. [1] github.com/mvduin/py-uio
A Mitutoyo rule for the Americans and a More and Wright Rule for the rest of us... now that IS pure quality!!! It'd be great to see a special edition with the "11 minute explanation" left in for the uber-nerds in the audience. ....... ah back in the days when "plug-ins" really did plug-in!
Marc!! Please, can you also publish the fast-forwarded talking complex bits too, in another video?! I'd really like to hear all of it!! p-l-e-a-s-e-e-e!!
Great stuff. Really enjoying this series. The DS1054 scope seems really popular, shame it's so hard to find, been looking out for one for some time now.
The mixture of gold and tin contacts in those Raytheon boxes probably isn't helping their connection issues. I've come across a lot of electronics over the years with oxidization from the differing metals. Mainly arcade games with gold sockets and tin chips.
I'm curious, what's the smallest modern computer that could run a similar program to the AGC? Could you run basic AGC functions on a arduino teensy? Or even something smaller like one of those computers on a chip, those chiplet things?
The AGC has 15bit+parity and a complete different instruction set than any MCU. Even if the arduino is a few times faster, it would be hard to emulate the AGC, maybe it could work with an assembler transcoder, but imho it wouldn´t be worth it to make such an acrobatic hack. I think a pure software emulation would be "easily" possible on simple ARM Cortex MCUs running with at least 50MHz, costs a few bucks. They have 32bit, so you could fetch an instruction from memory and transcode or interprete it on the fly. But there would be no analog&digital circuitry included for driving gyros, thrusters, DSKY etc, just the plain software execution. Simulating them on the same MCU (by using MEMS IMUs) would be imho possible, though. Bigger processors with 100MHz can definitely do it, using only 250mW and fitting into a matchbox, compared to the 70W of the original :)
@@Tedd755 you mean a PIC10F2(XX) a 6 pin SOT-23 device? "The world's smallest microcontroller" Promotional sheet PDF: www.google.com/url?sa=t&source=web&rct=j&url=www.microchip.com/stellent/groups/SiteComm_sg/documents/Training_Tutorials/en528372.pdf But it doesnt appear to have an output for a display? Or just one i/o? (Sorry I'm a noob) Let's say you wanted to output to a little colour led/lcd? (As they dont make tiny 7 segment displays?) There already appears to be a port to arduino...so wouldnt it run on a teensy?
Great video and thank you for doing what you do. I do have a question however, why were core rope memory necessary? Why didn't they just make a circuit board that had solderable jumpers in the form of a pair of closely placed pads? Wouldn't that be easier to automate, lighter, and maybe even more robust?
Toboter XP Thanks for the reply, but I can’t imagine that to be the case. Circuit boards can be made very thin (I’ve ordered 6mil boards from manufacturers) and that the pads can be made **extremely** small, smaller than the holes that are big by necessity due to the needle that needs to go through it.
@@signalworks Today you can probably get extra small circuit boards at every corner store, but in 1960 they were a little hard to come by. Remember that the entire digital circuit industry was in its infancy.
In the 1960s boards were about 25 mils thick, and holes were big enough for through-hole components, probably 8-10 mils. ( just looked up DIP specs. The pins were speced between 15 and 22 mils wide. So about 24 mils diagonal. You must have needed a .025 hole after thru-hole plating.) The pads were on the order of 40 mils diameter. Remember DIPs had pins on 100 mil centers, and the best you could typically do was route ONE trace between the two pads for a pair of DIP pins. Also, multi-layer boards were in their infancy. It was a while before they had 4-layer boards with internal power layers, and I'm not sure 6 layer boards came before the early 1970s. 25 or 30 layer boards didn't come along until the 80s, and I'm not sure anyone other than Sperry managed to get things that dense. Rope memories were just about the densest form of nonvolatile storage possible until the early 1970s when various forms of ROM came along. Since the early ones were all mask-programmable, it's somewhat doubtful that people that decided to make a computer with only one kind of logic chip for all logic would have wanted to try to make unique mask-programmed ROM chips work for each mission.
I am not on Jedi level with eanything you guys do here.! I have som basic knowledge of some stuff. I am absolutely fascinated with all of this 👀.! I just want more .!!!
I'd love to see this up close and personal when you get everything (fingers crossed) up and running. Any chance for a public display someplace like the Computer History Museum? Would love to meet you guys and hear more about this project.
We will be showing it and hopefully running demos on July 18, 2019 at the Cradle of Aviation Museum in Garden City, Long Island, NY, and on July 20, 2019 at the MIT Museum in Cambridge, MA. Both event as part of the Lunar Landing 50th Anniversary.
Any chance that you can post the normal-speed version of the "11 minutes of explanation rope memory addressing is complicated"? I really enjoyed your "episode 4-1/2" with the technical details...some of this stuff really takes me back to cool stuff of building a simple computer (albeit simulated) in college out of individual gates - I'd love to hear all the nitty gritty about how the Apollo computers address memory!
Marc I have a question: If the Raytheon core-rope simulators were built in the 70's, what did MIT and NASA originally use during software development and testing in the 60's?
At MIT, they used the AGC Monitor (which I've more or less replicated), which was also known as the Core Rope Simulator. It interfaces through the AGC only through the test connector. The test connector gives enough visibility into the internal operation of the AGC that an external tool like the Monitor can detect when the AGC is addressing fixed memory. That tool can then inhibit the fixed sense amplifier strobe (via signal MNHSBF) which prevents the sense amplifiers from outputting data read from the actual stored ropes. And when the AGC goes to access the data that has been read from ropes, the tool can inject the simulated rope data onto the central write bus. This is how we've been running in all of the videos up until this one. For whatever reason, the Monitor only existed at MIT. Raytheon, Grumman, North American, and the NASA centers instead got a big piece of equipment called the Computer Test Set (CTS). Along with this was a thing called the Program Analyzer Console (PAC), which performed the same sort of rope simulation duty as the Core Rope Simulator section of the Monitor. This approach is less "accurate" or "desirable" than the rope interface box approach, because your AGC might have completely broken rope drivers or sense amplifiers and you wouldn't be able to tell. In previous videos we ran without these modules even installed at all. So these boxes let us make sure that the rope circuits actually work before installing any real rope memories. They'd also be helpful if you want to simulate a rope, but for whatever reason don't have access to the test connector -- if, for example, you're dealing with a flight AGC which has the channel 77 Restart Monitor installed and you don't want to break configuration to remove it.
Mike's long explanation is available in full in this bonus material video here: ua-cam.com/video/NNlgJ52wuGE/v-deo.html
CuriousMarc Thanks!
7:53 (a short and an open pin) "You'd think they canceled out." - That is top electronics humour just there :P
Two weeks to go until the 50th anniversary of Apollo 11. Great work guys :)
Ken must have vast amounts of patience. He had a tough task in that box.
I had no idea just how thoroughly and meticulously, or difficult and lengthy this project is and has become.
This is 100s of thousands of dollars of combined effort, skill, resource, and time.
Pure unadulterated mutual love and concerted effort. I hope this AGC restoration gets the public acknowledgement that it deserves. It goes to show that the Apollo project was an effort rivaling any of 7 wonders of the world.
I would be interested to see a comparison to the Russian technology of the same period.
Well said!!
Nicolaj Baramsky
There’s several documentaries on the design and building of the AGC, filmed on location during the fabrication. I’d like to this in parallel and a combined montage.
Russian technology of the period was a music box with switches controlled by cams on a rotating drum. That's how our astronauts described it after seeing it aboard Soyuz in the Apollo-Soyuz mission in 1975.
I was four years old when Apollo 11 happened. I remember running outside that night to look at the moon and see if I could catch a glimpse of the astronauts as they did their thing. I finally met Armstrong a few years before he died - he was one of my heroes and one of the reasons I got into engineering in the first place!
Right. Monday. Let's get started work after a long weekend of children's birthday party. Motivated and ready. No distractions.
Ooo! Another CuriousMarc AGC video!
It would be very interesting to see the explanation in a separate video. It's kind of besides the whole point to skip it. Thank you for your very, very exiting videoes. Such a team! Best regards,
Yeah, I'm really into that stuff and would love to see the full version.
I agree.....totally....in less than a heartbeat :)
Also - this entire team is so cool! Can y'all do a group photo at some point - maybe a round of brief (re)introductions as to who everyone is? Thanks again for making these videos; really really excellent!
So the fixed core rope memory is now fixed fixed core rope memory?
Its actually fixed variable core rope memory since it is a simulator. :-)
The curiously named fixed fixed memory refers to page 0 in fixed memory (indeed from core rope) that does not need to be moved into view within the very restricted address space of the AGC by using the memory bank register (which is all the other pages, or the majority of fixed memory). Hence it’s doubly fixed. It would have been better named base bank or base page fixed memory, but it’s much less fun.
As a computer engineer, I've always appreciated the AGC. I am very glad you guys are working on one and sharing your experience with us. I also appreciate the very technical information as well.
I love watching this kind of Digital Archaeology, because that's what this is. Digging into our past through empirical observation, analysis of known information, and re-learning things that have long been forgotten.
Totally love the direction this channel has been going. You guys are true specialists, and its awesome that someone is taking the time to preserve and resurrect such important parts of our history - so that we can continue to learn from it and avoid repeating past mistakes. Cheers.
Nice work!
BTW, if you punch the "scope" button on the 7854 after it powers up it will clear the "self test complete" message on the screen.
One of my favorite parts of my week!
"Two days of debugging, two capacitors is the solution". Sounds about right! :D
"A capacitor a day keeps the transients away" :D
The amazing thing is not the AGC itself, but all the engineering and work that went into creating it, double that, they had only the technology of the 1960s at hand!
Whoa, seems like Mike has completely memorized/understood every nook and cranny of the AGC circuitry.
you should have left that long explanation part in the video (maybe moved to the end). I would have wanted to see it :)
Put it on CuriousMarcTwo.
Or should it be 2Curious2Marc - Tokyo Drift?
I absolute agree, i wanted to hear that too.
I would like to see that sometime, too.
I too would love to see those 11 minutes of explanation. This is a tech channel after all! S’il vous plaît.
If it really was 11 minutes, some of it is missing from this video. The sped up section was only 4 minutes ish in real time.
I know so little about how computers actually work, yet I fund these videos accessible and easy to understand. The passion shown by the experts, their incredible knowledge, the ability to trouble shoot problems. Fascinating viewing. Thank you very much
That memory is CRAZY, I can`t imagine how it could be so reliable :o
This is a wonderfully fascinating series. Not only your hard work getting the AGC working again, but your video documentation is top notch!
Watching you guys problem solve is such a gratifying experience! Thank you for sharing your creativity craft and methods.
Any time I start to feel too smart I just sit down and watch one of these videos. I know a little bit about consumer computer repair, but this is well above my pay grade! Thanks for taking the time to lay a little knowledge and understanding on me. It's kind of like visiting a museum and getting a personal guided tour.
You guys are amazing, good luck and I really hope you can get this up and running in time for the anniversary!
I just now found this, amazing what you guys have done, Mike's abilities don't cease to amaze me, this guy is a machine!, Man I I'm a hardware engineer and been working on assembler and digital circuits most of my life and after seeing what Mike and all of you are doing here I feel like a moron, fantastic series, thank you so much for done that, too bad I was so late to the party
Man, I love this series! And you guys are so clever! I couldn’t have done this kind of debugging in my life ... I admire your thinking capabilities!
Can we get the eleven minute in depth explanation as a seperate video?
Yes, we want the extended cut!
You can slow it down 20x and get a slightly legible result.
your wishes just got granted!
I'm obsessed with this series. I'm on the edge of my butt for every new video. Incredible work fellas!
Marc, your channel is absolutely behind compare, you have no equal peers online, with such unpretentious, hardcore, unique and EXTREMELY interesting videos with VERY hard to come by machines and friends with in depth knowledge of them. Wow!
God bless you and thank you so much.
Matthew.
Ken is a national treasure.
Spectacular. And a pleasure to watch. You guys are absolutely brilliant.
These are the fabled men in sheds that overcome obstacles..
Wow guys I love it, kept up with most stuff until now but the rope memory is definitely a struggle for me. Reminds me of a bunch of my friends huddled around an 8080 development board trying to get some interface working. Wonderful to see real fundamental knowledge ( and old test gear ) at work
6:45 I'm so glad even you recognise understanding some parts in your videos can be challenging. I've followed the whole AGC series without any serious, in-depth computing knowledge at all, go figure my struggle at some points (though not as many as I expected at first). So finally I'm not feeling like the only idiot watching this, yay!
I understand maybe less than 1% of what you guys are doing but I’m completely hooked on this mini series. Good job guys and great content
I love how a Xerox Alto is just sitting there on a desk in the background while they work on an Apollo AGC. Paradise for geeks :)
For me this is building up more hype than the game of thrones finale did. Even at 36 this is making me want to get into more electronics stuff. Thanks a ton.
This entire process and your dedication to it is breathtaking. Great work guys!
Without a doubt, you guys exemplify the Apollo Program spirit.
Hardcore debug, analog and digital... 10 x Thumbs Up!
Finally, people using the word "glitch" when its appropriate.
I get excited to watch what you guys get up to with this project, What an incredible job you're doing. Thanks for these videos.... I can't wait till the next one.
Wonderful video as always.
You guys can do it!!!
This the most beautiful and exciting series on UA-cam for me.
Some really outstanding (reverse) software & electronics engineering !!
God I love this stuff, we all take computers for granted now a days. This is what some of the most impressive engineering I've ever seen.
These videos are like catnip - I cant resist them. Why do I find them so good? Why?
Amazing perseverance and problem-solving on display - I’m enjoying this series so much!
All this time I thought Apollo 11 used a Saturn V rocket to get into space and it turns out they were using rope! The things you learn on UA-cam....
Any chance of a collaboration with a channel building a full-size Saturn V? :-)
This is one of my favourite UA-cam channels. Keep up the good work!
It still blows my mind how physically difficult it was to make memory and other parts back then. It just goes to show how devoted people were to projects. Stuff got done, even if it had to be invented on the spot. And it's great to see a group of people that carry on this type of work.
same story with the F1 main engines on the Saturn 5, while the blueprints still exist, a whole lot of custom fitting had to be done by hand to actually make them work, and the notes on what was done there either was never written down, or lost. the team that looked up possibly recreating those engines actually found it easier to come up with a simplified cad model that brought the original engine's some 5000 parts down to about 55-60 pieces . simply because modern design and manufacturing techniques allow us to build parts with shapes and complexities that simply weren't remotely possible then
I'm right there with you guys in spirit sending you positive thoughts...to quote Oddball (Donald Sutherland) from Kelley's Heroes " Crazy... I mean like, so many positive waves... maybe we can't lose, you're on!"
Can’t say it enough! You guys rock! Thanks again, looking forward to the next installment!
Definitely got one of the tougher meals to chew on. Much respect to that man.
I love you guys! So fascinating to watch this play out.
I still have to argue with people who think that we didn't have enough computer power to get to the moon. I point them to this playlist.
I never thought the Apollo missions were faked, and these videos - beyond being totally awesome in their own right - are definitely confirmation. There is no way anyone would've gone to this much trouble just to shoot the whole thing on a soundstage. The point of fraud is to save time, effort, and money, after all.
@@reallyme3573 When you say, "A modern day phone with 8GB of data," you're talking about your data plan limit. What you MEAN to say, is storage space. which just means that your modern phone could fly to the moon and have plenty of room left over.
Not a lot of calculations are needed, they used to do them on slide rulers, which takes a lot longer. Only a couple decades earlier, in the Manhattan Project, they were computing much more complicated equations using mechanical calculating machines larger than a typewriter before they got access to an IBM mainframe. Core memory actually was pretty fast for the 1960's compared to other forms of memory, it was just limited by the speed of the TTL logic.
It was a 2MHz 16-bit computer, which is comparable to the first generation of home PCs only a decade later. I've still got a computer that isn't much faster, still doesn't have a math co-processor, and it does fractions just fine. You don't need a math co-processor to do fractions, and you never did. You just tell it to stop counting after so many decimal points, so it doesn't fill up the registers.
Each sequence of the journey was pre-programmed into the ROM, and when you're programming in assembler, there's no bloat to worry about, so not much space is actually needed. It's been uploaded to GitHub, if you want to go over it: github.com/chrislgarry/Apollo-11/tree/master/Luminary099
@@reallyme3573 the real problem is that so few people actually know how to write efficient code at the machine level these days (these guys still do),
Nice Debugging Guys! awesome to see all the old test units at play with newer tools to give the information needed to fix the problem! Excellent work!
Ha! In England if you call someone a "dipstick" it means they are stupid, an idiot! Keep up the fantastic videos.
We confirm. Dipsticks are stupid.
I'm pretty sure that's American slang originally...
Another outstanding episode, of what has to be. the best thing on UA-cam. P.S. I just ordered the T-Shirt
Just WoW, very much a time swallowing monster,, impressive for sure,, keep up the amazing dedication
Ken's a funny guy! Well done to all involved, keep it up!
The saga continues!! So exciting! Can we please have the 11 minute explanation in its own video? I'm curious about the fault and would like to know what he said... remember, we're nerds too!
Very grateful to you for these videos !
Ken is a beast! Way to go guys... I hope to see this system working someday.
Nice application of the BeagleBone! PRU is awesome
Agreed, I love the PRUs. Their crazy little assembly language is one of my favorite's I've used.
@@mikestewart8928 It's very nice and clean indeed.
The only thing I don't like is the libprussdrv library on the linux side, which is part of why I created py-uio[1], and hopefully one day I'll get around to also make a better C/C++ library.
[1] github.com/mvduin/py-uio
A Mitutoyo rule for the Americans and a More and Wright Rule for the rest of us... now that IS pure quality!!!
It'd be great to see a special edition with the "11 minute explanation" left in for the uber-nerds in the audience.
....... ah back in the days when "plug-ins" really did plug-in!
Great Job Terry You Do Fine Work
Just awesome. Can’t wait for the next episode!
This series is so awsome. I dream of working on something like this. God bless you guys.
The anticipation.... Can't wait to see more!
Its so glorious! you guys are doing some great work!
Great work as always, I would love to see this working soon. Keep it up.
Marc!! Please, can you also publish the fast-forwarded talking complex bits too, in another video?!
I'd really like to hear all of it!! p-l-e-a-s-e-e-e!!
Published now, next video ua-cam.com/video/NNlgJ52wuGE/v-deo.html
Great stuff. Really enjoying this series. The DS1054 scope seems really popular, shame it's so hard to find, been looking out for one for some time now.
The very best the internet has to offer! I am Gobsmacked!
Great work guys! Can't wait to see the next episode.
New T-Shirt Idea:
Marc &
Ken &
Carl &
Mike
These videos should have a million views
The mixture of gold and tin contacts in those Raytheon boxes probably isn't helping their connection issues. I've come across a lot of electronics over the years with oxidization from the differing metals. Mainly arcade games with gold sockets and tin chips.
Agreed. I found this weird too.
Wonderful work!
Totally awesome work guys! Smashing it out of the park!
This project is beautiful
I'm curious, what's the smallest modern computer that could run a similar program to the AGC?
Could you run basic AGC functions on a arduino teensy? Or even something smaller like one of those computers on a chip, those chiplet things?
FPGAs can simulate the entire logic of the AGC, but memory will be logic RAM and ROM, not core, or rope.
Try a PIC20F222. Six pin bit of cockroach crap in the SMD version.
The AGC has 15bit+parity and a complete different instruction set than any MCU. Even if the arduino is a few times faster, it would be hard to emulate the AGC, maybe it could work with an assembler transcoder, but imho it wouldn´t be worth it to make such an acrobatic hack. I think a pure software emulation would be "easily" possible on simple ARM Cortex MCUs running with at least 50MHz, costs a few bucks. They have 32bit, so you could fetch an instruction from memory and transcode or interprete it on the fly. But there would be no analog&digital circuitry included for driving gyros, thrusters, DSKY etc, just the plain software execution. Simulating them on the same MCU (by using MEMS IMUs) would be imho possible, though. Bigger processors with 100MHz can definitely do it, using only 250mW and fitting into a matchbox, compared to the 70W of the original :)
A quick Google search reveals someone running a Virtual AGC on a Palm Centro! That's a (theoretically) 312MHz CPU. Obviously it can run on much less.
@@Tedd755 you mean a PIC10F2(XX) a 6 pin SOT-23 device? "The world's smallest microcontroller"
Promotional sheet PDF:
www.google.com/url?sa=t&source=web&rct=j&url=www.microchip.com/stellent/groups/SiteComm_sg/documents/Training_Tutorials/en528372.pdf
But it doesnt appear to have an output for a display? Or just one i/o? (Sorry I'm a noob)
Let's say you wanted to output to a little colour led/lcd? (As they dont make tiny 7 segment displays?)
There already appears to be a port to arduino...so wouldnt it run on a teensy?
That's some super debugging right there. Huzzah!
Great video and thank you for doing what you do. I do have a question however, why were core rope memory necessary? Why didn't they just make a circuit board that had solderable jumpers in the form of a pair of closely placed pads? Wouldn't that be easier to automate, lighter, and maybe even more robust?
It would take up more space. And since this was the first micro computer ever, space was of the essence.
Toboter XP
Thanks for the reply, but I can’t imagine that to be the case. Circuit boards can be made very thin (I’ve ordered 6mil boards from manufacturers) and that the pads can be made **extremely** small, smaller than the holes that are big by necessity due to the needle that needs to go through it.
@@signalworks Today you can probably get extra small circuit boards at every corner store, but in 1960 they were a little hard to come by. Remember that the entire digital circuit industry was in its infancy.
That's why everything in the AGC is connected with wires, not traces.
In the 1960s boards were about 25 mils thick, and holes were big enough for through-hole components, probably 8-10 mils. ( just looked up DIP specs. The pins were speced between 15 and 22 mils wide. So about 24 mils diagonal. You must have needed a .025 hole after thru-hole plating.) The pads were on the order of 40 mils diameter. Remember DIPs had pins on 100 mil centers, and the best you could typically do was route ONE trace between the two pads for a pair of DIP pins. Also, multi-layer boards were in their infancy. It was a while before they had 4-layer boards with internal power layers, and I'm not sure 6 layer boards came before the early 1970s. 25 or 30 layer boards didn't come along until the 80s, and I'm not sure anyone other than Sperry managed to get things that dense.
Rope memories were just about the densest form of nonvolatile storage possible until the early 1970s when various forms of ROM came along. Since the early ones were all mask-programmable, it's somewhat doubtful that people that decided to make a computer with only one kind of logic chip for all logic would have wanted to try to make unique mask-programmed ROM chips work for each mission.
I am not on Jedi level with eanything you guys do here.! I have som basic knowledge of some stuff. I am absolutely fascinated with all of this 👀.! I just want more .!!!
Great job gentlemen. Thanks!
Hey Gentlemen, that it’s magic
So enjoyble clips.
👏👏
I'd love to see this up close and personal when you get everything (fingers crossed) up and running. Any chance for a public display someplace like the Computer History Museum? Would love to meet you guys and hear more about this project.
We will be showing it and hopefully running demos on July 18, 2019 at the Cradle of Aviation Museum in Garden City, Long Island, NY, and on July 20, 2019 at the MIT Museum in Cambridge, MA. Both event as part of the Lunar Landing 50th Anniversary.
@@CuriousMarc So, nothing on the West Coast?
@@wayneholder Not yet. It costs a fortune to move that AGC...
@@CuriousMarc If money is the issue, I'm willing to chip in some cash. And, if enough others were interested, too, perhaps something could happen.
Any chance that you can post the normal-speed version of the "11 minutes of explanation rope memory addressing is complicated"? I really enjoyed your "episode 4-1/2" with the technical details...some of this stuff really takes me back to cool stuff of building a simple computer (albeit simulated) in college out of individual gates - I'd love to hear all the nitty gritty about how the Apollo computers address memory!
Posted here ua-cam.com/video/NNlgJ52wuGE/v-deo.html
Are you going to be able to simulate the IMU at all... and thrusters ,ect.
Probably with a mod for Kerbal Space Program.
Masterful troubleshooting!
Wait... a short and and an open don't cancel each other out?
I follow this project from the beginnig and i love it!
Well done guys. It is way above my head, though I do like flashing lights, lol.
In the literal sense of an overused word: awesome.
Talk about a dream team!
Thank you for your hard work
the logic gates were often very power hunger RTL (Resistor Transistor Logic)
Marc I have a question: If the Raytheon core-rope simulators were built in the 70's, what did MIT and NASA originally use during software development and testing in the 60's?
At MIT, they used the AGC Monitor (which I've more or less replicated), which was also known as the Core Rope Simulator. It interfaces through the AGC only through the test connector. The test connector gives enough visibility into the internal operation of the AGC that an external tool like the Monitor can detect when the AGC is addressing fixed memory. That tool can then inhibit the fixed sense amplifier strobe (via signal MNHSBF) which prevents the sense amplifiers from outputting data read from the actual stored ropes. And when the AGC goes to access the data that has been read from ropes, the tool can inject the simulated rope data onto the central write bus. This is how we've been running in all of the videos up until this one.
For whatever reason, the Monitor only existed at MIT. Raytheon, Grumman, North American, and the NASA centers instead got a big piece of equipment called the Computer Test Set (CTS). Along with this was a thing called the Program Analyzer Console (PAC), which performed the same sort of rope simulation duty as the Core Rope Simulator section of the Monitor.
This approach is less "accurate" or "desirable" than the rope interface box approach, because your AGC might have completely broken rope drivers or sense amplifiers and you wouldn't be able to tell. In previous videos we ran without these modules even installed at all. So these boxes let us make sure that the rope circuits actually work before installing any real rope memories. They'd also be helpful if you want to simulate a rope, but for whatever reason don't have access to the test connector -- if, for example, you're dealing with a flight AGC which has the channel 77 Restart Monitor installed and you don't want to break configuration to remove it.
Didn't the original computer have random crashes to ?
Exciting work going on!
and what is the cause of the glitch ?