Congrats, this is the tube machine you deserve! The G-15 is one of my favorites. Imagine, a machine so small, it can be your own personal computer! And that green cabinet presentation… Epic.
He covers systems nobody has nostalgia for so its never going to be as popular as people who talk about c64 or msdos games. but its interesting for geeks like me who want to learn about the roots of computing .
Dad worked at Bendix in the 60's right out of college, never even knew they made computers until you mentioned them a while back. He worked on safety triggers for nukes, but that's all he could tell me. We were only allowed to visit the plant once in his entire career.
@@Starchface Bendix also made lots of toxic waste. My dad's plant is now an official toxic waste site, last I heard the land can never be used again. Kind of mad that it was allowed to get that bad while my dad worked there.
Bendix was all over the board, they made all sorts of wild stuff! Near the end of the G15s run, Bendix sold the whole computer division to Control Data Corporation. CDC never built any G15s as far as I know, but they did sell and re-badge them as Control Data machines.
Mr Usagi, a tip for you, as I have found very few people know this these days:- I notice power-type selenium rectifiers (red fins - probably Federal type) at 21:20. These slowly turn into resistors when not used for many years. if you suddenly apply power at full voltage they may overheat and destroy themselves with a bad smell. A very bad smell. They need to be reformed, somewhat like old electrolytic capacitors. Use a variac to bring the voltage applied to them up very slowly (hours) while monitoring the current and how warm they get. If you do this they will usually then be fine and turn back into diodes. Some restorers of old electronics don't trust them and replace them with silicon diodes, but if you want an authentic museum quality machine, carefully reform them. [Note that selenium rectifiers were made in two grades - one for consumer equipment and one for industrial equipment. NEVER trust the consumer grade stuff - replace them with silicon.]
I've got to confess to a major case of envy; I've always wanted to work on a G-15. So glad to see Bob trusted it to your capable hands. Can't wait to watch all your update videos.
This is beautiful. What amazes me more than anything is just how fast computer science moved 70 years ago. From ENIAC to this in less than a decade. Being able to read the schematics, see the OS code would be fantastic. Those guys back then weren't dummies, that's for sure.
And if you go 10 years beyond this, we had the Apollo Guidance Computer using flat-pack integrated circuits! It was truly a bonkers time for computing. This machine did run some interesting programs, one of which was Intercom, a sort of virtual machine that sits on top of the original machine and allows you to write programs directly on the machine. It's not as fast as writing bare-metal programs, but it also will keep you from going insane. As we get close and closer to running, we'll get deep into the weeds on just how exactly it loads and executes programs from the drum, what the instruction set looks like and how to use it. It's going to be a lot of fun learning about this thing!
@@UsagiElectricFlat-pack IC's and core rope memory... something that as much as I understand how it technically works, I still can't wrap my head around it. Truly amazing what these guys (and gals) were able to pull off back then!
I can't tell you how excited I am for this series. I melted inside at the CLUNK of that switch. I'm envious that you have it, not so much the task of getting it running. 👍 This will be AWESOME.
Indeed, a work of art! Beautiful. I'm currently building an 8 bit computer out of TTL on breadboard - see Ben Eater - and cutting and stripping wires is a little tiresome, then I saw this and thought man how easy I have it.
The fact that there are so many identical cards (and with two machines in the area as well) makes me think that making dedicated testing jigs for them might be useful. Just a Arduino program (and presumably a bunch of Mosfets or relays) that checks the function of the cards. Oh, and good luck reforming those caps in the huge capacitor bank at the bottom.
Thank you! The current goal is to have the AC coming up by years end, and then to execute proper code before the end of next year. I think, barring any intense failures, that's a fairly realistic goal!
In the early 1970's the radio club at the college I attended had a G15 and a pair of tape drives. We nearly got it functioning, but the drum memory gave serious trouble. The project eventually failed due a scratch on the drum that wiped too many tracks. Best of luck with yours!
Congratulations. The G15 and the Dec PDP 8 were the first computers that I learned on in computer maintenance school, and the first computers that I programmed on. I will always remember having to spin the G15's drum memory shaft before powering it up so that the rubber coupling wouldn't break from the motor torque. Enjoy.
The G-15 is cool in that it had a rudimentary operating environment called Intercom which essentially create a virtual machine which allowed working with double precision decimal math. The communication channel allowed up to four mag tape units or four more G-15s to be connected in a network. There was also a refrigerator peripheral sized differential analyzer which held 100+ manually configurable analog integrators, a sort of math coprocessor. There were a lot of ground breaking ideas embodied in the G-15 and the names associated with designing the G-15 are a veritable who’s who of computing founders. It is good to see this bit of history preserved. Thank you.
When I saw the nameplate I said out loud "a Bendix? - a Bendix!" WOW!! I am excited to watch you work on this. I like how you dive deep into everything you work on and keep at it - you don't give up.
You're out there doing God's work, bringing these machines to life, where they most likely would have remained non functional. Historical preservation, in my opinion, is very important. And not just for computers. I love your channel, and I can't wait to see more.
I am really glad too that there are people out there like him who love and respect antiques like this and are bringing them back to life. It's definitely saving a piece of history. Not only will he be restoring the computer itself, but by learning how it works and documenting the project with videos he is helping to preserve knowledge.
Oh, good news. I went and looked at the tech manuals and pictures and the G-15 seems to be made of very reliable parts. I was worried that the rectifiers might be selenium or copper-oxide, but they seem to be very early germanium diodes. Goody. The filter caps may be fine and just need a slow reforming. A infrared thermometer would be ideal for watching for very leaky ones. The board capacitors seem to be mostly very high quality hermetically-sealed ones. You rarely see these go bad. Those early diodes were pretty good too. Long ago I restored a CDC 160-A which has thousands of germanium transistors and diodes. Only found one leaky diode and no bad transistors. Hope your luck runs along those lines. The tubes are mostly the relatively common 5965 or 12AV7 dual triode. It might even run with a 12AT7 in non-critical circuits. Your main tube problem might be intermittent or corroded tube sockets. You may need to find a 50-amp Variac in order to bring up the AC voltages slowly. Cheers, George
As you revealed the computer, admittedly, an Anglo-Saxon explative or two slipped out as my jaw dropped. Congratulations, man! This is going to be the best project you've done yet 🤟
I live 20 minutes away from System Source in Baltimore and had no idea the place existed!!! Made a trip up there yesterday for my birthday and got to meet Bob and Brendan and get a tour. Will definitely be going back there. Thanks for sharing!
Depends, some of those might be of such high quality that with a little reforming they might work well enough. If you have to replace them that can be a significant cost, depending on financial wealth.
@@maxs.3238 I meant the big row of filter caps in the bottom. Personally i usually prefer outright replacement with high quality reliable caps. But this is a museum piece. So an attempt at saving the filter caps might be needed. Or maybe restuffing them. On A museum piece look is important. The other caps same thing. Replace, restuff, or test and leave in there. They should be analysed and tested and then a choice needs to be made.
@@bzuidgeest A lot of this depends on what happens if a cap shorts out. I'm assuming bad things. I'm pretty sure they would be electrolytic. For long term reliability, they will need to be replaced. Buying replacements will be $$$, so restuffing might be the way to go. I'm just glad it's not my problem.
You are definitely brave. I saw what was involved with getting an IBM 1401 going at the California Computer History Museum. Now it did involve more than just the computer but they had a team of volunteers some of whom had worked on 1401s before. It still took them quite a while to get it running. Maybe on the plus side, the 1401 used germanium transistors which were a bit failure prone. Maybe the tubes might be a bit better. I really hope it goes well for you but you are definitely taking on a real challenge. I hope no irreplaceable parts keep you from finishing this project. Really want to see this work. I'm guessing you want that even more.
I love the aesthetics of the G15 and the attention to the needs of servicing engineers! Reminiscent of old power station attention to appearance which is maybe no coincidence given it’s ample reliance on one!
I just wanted to say thank you for preserving all these historical pieces. The computer is perhaps the most life changing invention of the modern age, and yet it seems we're all too keen on discarding our history as soon as the next generation of devices becomes available. It's collections like yours that preserve these machines and the history of early computing for generations to come.
As a professional stagehand who constantly deals with casters and caster replacements, I would strongly suggest next time you go to make an adapter plate for casters to not weld the caster directly on. Instead, make a second plate that matches the casters you want to use. It makes it so much easier to change out casters later because casters do wear out more often than people realize. Also, it's really not safe to weld zinc plated caster plates; There's a lot of toxic gases created.
Amazing! It almost looks like a prop from a Steampunk film. Optical paper tape readers are quite a lot older than this. The first Colossus computer, which was built in 1942, used photocells to read paper tape containing intercepted Nazi ciphers, comparing the code with an internally generated decrypt code. Previous hardware had used mechanical readers, which was much slower and a lot less reliable.
Scrolling to find this comment. I went to Bletchley last year and got to see the wonderful Colossus rebuild actually running, including the tape. If memory serves they invented this specifically because mechanical readers were to slow to do the decoding task in the required time. But then again, Colossus would have been still very much classified in the 1950ies, so Bendix might have had to invent it again. Although Wikipedia says that Harry Huskey, designer of the G15, had worked with Alan Turing on the ill-fated ACE computer before! So it might be conceivable that it could have been Turing himself who "leaked" this little trick!
This will be an epic restore. Bob is a super awesome guy, I love visiting System Source, he is such a wealth of information, and he loves to talk about his collection, and is a wonderful guy to just listen to! I look forward to seeing this work, I thought this machine looked familiar when you first showed it, when you said it was from Bob, it all clicked.
The nice thing about having a bunch of ancient computers in need of fixing on a youtube channel dedicated to fixing and showcasing ancient computers is seemingly self evident.😊
This is SO COOL!!!! This hits some of my favorite era of electronics and computing! Between the EM bits, and the sheer approachable nature of how everything is built, it's gonna be EPIC to see this whole project go! THANK YOU!!!!
Great video! Glad to help with the casters in my tired stupor! Hopefully us over at VCF in NJ can get a G-15 working so there can be 3 total in the US!!
I don't know anything about vintage computers, but the great beauty of this behemoth is undeniable. The combination of computer tech and the physicality of those buttons, switches and dials are immensely satisfying.
Love that power regulation with built in variac and rheostats. 50 amp service!! Wow. Looks like a bunch of selenium rectifiers in the tape punch/reader.
Usagi Electric - taking unboxing videos to a new level 😊 Looks like another super fascinating project you’re kicking off here. I know nothing about Bendix machines. I was watching a video from the Computer Conservation Society about the Elliot 401 from the early 1950s- that also used a drum for main memory and also used a track on that drum for generation clock signals. I guess it was a good option on systems of this vintage.
So cool. NASA used these as ground support equipment for some of the earliest space launches. If you take the Space Force Missile museum tour you will see one in the blockhouse at Launch Complex 26.
I'm happy to report that I'm someone who owns a tractor and loves vintage computers. It takes a special set of skills to work on a thousand pound computer! We are lucky to have you!
I'm so glad to hear you have a friend with one of these, because that will be a yuge help when figuring out those added-on bodge wires. All good wishes.
(Apologies for being a wet blanket) A while ago I was watching a machinist video (I watch a lot of them). Lathes, Mills, Surface grinders etc are really heavy. He was talking about castors. He pointed out that while you might *like* to say that with 4 castors each castor is taking 1/4 load, it's likely not the case, unless your floor is a surface plate. As everyone knows, 3 points of contact is all you can really guarantee, which is why photographers use TRI-pods. But it gets worse; call the corners A-B-C-D, going round You might hope that even if you don't get 4 point contact, you might at least "fall back" to a 3 point configuration (say A-B-C). But (assuming a roughly even load) it's sort-of obvious that this A-B-C contact could easily be tipped to become a A-D-C contact. A-C is a "high diagonal" (either A or C must be high) and the load can pivot on this A-C axis. Which means, at least momentarily, the A-C castors are taking *all* the load. Summary: your castors (and floor) need to be spec'd for at least 50% of the total item weight.
I don't know why I'm excited for this series but I am :D My country was behind the iron courtine in this era so this is something very exotic to me. This behemoth you have was built before even first computer in Poland was designed not even talking about mass production.
The Electric company is probably looking at your account like a kid in a candy store. Omg this is going to the most epic restoration ever. They only made 400 of these, and there’s only a handful of them left in the world so yes, it will probably be the only fully functioning one unless like the Smithsonian has one or something.
I’m not sure if you already have a solution in mind for providing a 120V 50A circuit, but I have a suggestion. Rather than heavily loading one phase of your electrical service, you should try to find a cheap used 240X480V to 120/240V transformer. The primary can be jumpered for 240V, and the secondary jumpered for 120V. You’d be looking for a 7.5-10 kVA class transformer. Dry type or a potted enclosure that is meant to be wall mounted. This would allow you to run a 30A 240V circuit to the transformer to step down to 120V. From there hit a 50A fused disconnect switch or single breaker in an enclosure (it could be a 2-pole breaker, just leave the second pole disconnected). If the computer has its own 50A breaker, then you could just leave the external one out and hardwire the transformer secondary straight into the computer. I would use a 240V 30A twist lock plug and receptacle on the primary side of the transformer so it’s not permanently hardwired to your building. Slap it on a rolling cart! Make sure to ground one leg of the 120V secondary to be the neutral, the ungrounded leg will be the hot conductor to be switched or fused. This is all under the assumption that you service is 120/240V split phase. If it’s 208V, then obviously the transformer will need to be different. Some transformers have taps, for example a 90% tap for 240V would be 216V. If you connected that to 208V, the secondary side would be at 115V which would probably be fine if the computer can handle it.
Best of luck reviving the G-15D. I tried to get one working as an undergraduate. One of the primary problems was that the diodes developed high reverse leakage. A goodly number of them had to be replaced. I did get it working and compiled a number of Fortran programs on it.
I have lived within 30 minutes or so from that museum for most of my life. Why am I just hearing about them now? I know where I am going on my next day off!
Aren’t caps from that era filled with nasty PCB’s? You’re probably aware if so but if not check. You don’t want to mess with those without taking the proper precautions. Great project! Can’t wait to see it running!
It's something like 20 years younger than the ban on the use of PCBs (polychlorinated biphenyl) so it would make sense that they'd be used extensively.
@@oliverer3 PCBs are mainly but in the AC capacitors, such as the run capacitors on motors or AC line conditioners, but could be in smaller oil filled capacitors. If a capacitor says Pyranol on the label, there is a good chance it is PCB-bearing oil filled. Electrolytic capacitors have polysaccharide dielectric material with additives. I don’t know what the additives are, but I wouldn’t count on the additives being harmless.
Thank you! All of the caps on the DC side are high-quality Sprague electrolytic. These are the same ones that were used in HP test equipment in the 50s, and I've found them to almost all still be good and safe to use. The only thing that is a mystery is if the drum motor has a start capacitor and what it is. That particular one may need to be replaced if it's a PCB filled cap.
@@UsagiElectric The G-15 is a beautiful machine and it is great to see it in such capable hands for its restoration. It would be interesting to see if any metal paper tape cassettes with canned programs can be found, especially ones for POGO and ALGO. In my scrap electronics adventures, I’ve never seen anything like those Bendix cassettes in 50+ years. I took a quick read through of the programmer’s reference manual, and boy would writing just about anything from scratch be a tedious undertaking. The system reference manual also contains a flow chart which refers to a self-test program to be run at power-on after the initial voltage checks. Sometimes small-ish capacitors can be suspect. I ran across a 4.7 uF axial lead capacitor with cardboard sleeve and red rubber end caps in a 1946 National HRO receiver’s audio stage. The sleeve was wet with oil and there was oil on the bottom cover of the chassis. Usually a cap like that would have been wax paper and foil. I happened to have a WW-II Sprague fungus proof cap in a plastic shell which was about the same size, so I swapped that in.
The G15 was my first personal/home computer. You did an excellent introduction and if you would like to connect to discuss the repair, operation, and programming of the G15, I encourage you to connect with me (I can share the challenge areas I navigated such as the master timing circuitry). Sadly, my G15 died (its an interesting story) and was disassembled in 1979 but I still have a few parts. In addition to the G15 & console, I had 2 MTUs. FYI that small corner is going to get REAL hot ---- too hot to keep the system running smoothly. As the plate voltages drift due to excess heat the system will misbehave. I look forward to future videos & possibly connecting. You've made me quite nostalgic for a simpler time.
Man, you are my hero! When I saw the plastic wrap inside the big box, I knew, "it's a Bendix!", due to the shape. Thank you very much! I'll see every part of this saga!
So cool that you're working on bringing this machine back to life! My uncle has told me stories about his days programming a Bendix G15 back before computing life was easy. Looking forward to seeing you at VCFSW!
When I was a high school senior (1965), we received a G15 for a new computer class. We were told it was the very first computer in the nation specifically for student use. I don't remember much about programming it any more, but have always triggered on them wherever I ran into an article, or picture. There is one at the Smithsonian National Museum of American History, but I don't know if it is operational. What surprised me was that it has a mag-tape drive attached. I really look forward to watching your progress on getting this system up and running!
I am very glad you are documenting your journey, I love watching how the systems work! You and other creators have inspired me to restore my own devices!
I'm actually pretty excited by this thing. I love how the power on process is a whole launch sequence. I'd love to spend a day just... toning out all of the connections to the boards to make a map of it.
That G-15 General Purpose Digital Computer is serial number 238. How do I know that? My father J. H. Harrison, and then my family, owned that baby from 1967 until 2021. It was in my family's basement in a suburb of Washington DC and was last operated in the early 1970s. My dad purchased it in 1967 from Philips Laboratories in Briarcliff Manor, New York. He purchased this unit and some others, and many G-15 related items, when the G-15 was being phased out by businesses and schools. My brother contends that my dad was one of the world's first computer hobbyists. As you mentioned, 238 is now owned by System Source Computer Museum in Hunt Valley, Maryland. My dad's siblings are big fans of them and their museum that can rival any computer museum around. Good luck and have fun with the restoration project. We will be watching the progress.
You're absolutely right, 238 on the nose! I've heard a lot of about the Harrison family in relation to the machines, it's an honor to finally hear from y'all! If you don't mind, shoot me an email at "Nakazoto at gmail" or join our Discord chat server ( discord.gg/p7UsfHD ), I'd love to hear some stories of this machine and the other G15s y'all had. And also, hopefully you can solve a mystery on this machine that's got us scratching our heads!
Worked on the Navy's NTDS systems in the 70s. Paper tape, flex writers, 800 bpi tape drives, drum drives. 32 k of 8 bit core memory. Those were the days. Had to be very careful with memory ussage.
HEAT! At 38A, 120V, that's 4.5kW so that's like having at least three typical 120V room heaters going flat out at once - toasty! Good luck though, it's going to be fun winter project, time to duct that air around the house... and to the neighbours even!
This is just an amazing machine. It's the same age as me! That backplane reminds me of the DECSystem-10 KL-10 backplane I worked on back in the 1970's - complicated. But vacuum tubes and a drum memory? This is gonna be quite a challenge. Good luck! I'm looking forward to watching your progress.
Knowing you're likely younger than Bob means that my hypothesis is true that there will be young folks like myself interested and invested in obscure old machines even as these things reach 100+ years old. And having access to stuff like these videos (hopefully archived) will help too. Having it in a museum will hopefully prevent this stuff from being lost to time.
Congrats, this is the tube machine you deserve! The G-15 is one of my favorites. Imagine, a machine so small, it can be your own personal computer! And that green cabinet presentation… Epic.
Tektronix teal!
Oh Sir Curious do we smell a crossover video? x
@@geemcd I'm absolutely sure there will be one from VCF Southwest :)
Thought of you Marc immediatly!
Looking forward to hearing your thoughts on the machine in person at VCF next week!
Back in 1949 Popular Science said that in the future computers would weigh no more than 1.5 tons. You just proved them right.
Meanwhile there is a pentium 3 somewhere making "your motherboard so fat!" Jokes. 😂
One point five tons and waddya get? Another gray coder and a forerunner of the PET...
@@denniseldridge2936 your motherboard so fat, the front side bus is an actual bus. 🤣
@@rmcdudmk212 she got fat by eating pentium 3 candy bars
@@deltaray3 She had too many bytes.
This is hands down the most underrated vintage computing channel on UA-cam.
It's highly rated i think. It's just not very large. But that is normal with such niche content.
He covers systems nobody has nostalgia for so its never going to be as popular as people who talk about c64 or msdos games. but its interesting for geeks like me who want to learn about the roots of computing .
It's cool to see each retro UA-camr kinda has their own "era" of system they prefer, let's me enjoy them all equally
Thank you!
The content is definitely niche, but the community that pops up around these systems is truly stunning!
Agree. Between Usagi Electric and Curious Marc, I definitely get my big-iron computing and test-equipment fix.
Dad worked at Bendix in the 60's right out of college, never even knew they made computers until you mentioned them a while back. He worked on safety triggers for nukes, but that's all he could tell me. We were only allowed to visit the plant once in his entire career.
Nuclear bombs, computers, brake pads. Is there anything Bendix _doesn't_ make?
Would it be possible to visit the factory where they make the un-safety triggers for nukes? Asking for a friend...
@@Starchface Bendix also made lots of toxic waste. My dad's plant is now an official toxic waste site, last I heard the land can never be used again. Kind of mad that it was allowed to get that bad while my dad worked there.
Bendix was all over the board, they made all sorts of wild stuff!
Near the end of the G15s run, Bendix sold the whole computer division to Control Data Corporation. CDC never built any G15s as far as I know, but they did sell and re-badge them as Control Data machines.
_[Fallout theme plays]_
Mr Usagi, a tip for you, as I have found very few people know this these days:-
I notice power-type selenium rectifiers (red fins - probably Federal type) at 21:20. These slowly turn into resistors when not used for many years. if you suddenly apply power at full voltage they may overheat and destroy themselves with a bad smell. A very bad smell. They need to be reformed, somewhat like old electrolytic capacitors. Use a variac to bring the voltage applied to them up very slowly (hours) while monitoring the current and how warm they get. If you do this they will usually then be fine and turn back into diodes. Some restorers of old electronics don't trust them and replace them with silicon diodes, but if you want an authentic museum quality machine, carefully reform them.
[Note that selenium rectifiers were made in two grades - one for consumer equipment and one for industrial equipment. NEVER trust the consumer grade stuff - replace them with silicon.]
I've got to confess to a major case of envy; I've always wanted to work on a G-15. So glad to see Bob trusted it to your capable hands. Can't wait to watch all your update videos.
I'm kind of sad that I eventually have to give it back, but the best place for it will be at a museum where people can appreciate it in person!
This is beautiful. What amazes me more than anything is just how fast computer science moved 70 years ago. From ENIAC to this in less than a decade. Being able to read the schematics, see the OS code would be fantastic. Those guys back then weren't dummies, that's for sure.
Being such a small machine, I doubt it had an OS. You would just run your programs on the bare hardware.
And if you go 10 years beyond this, we had the Apollo Guidance Computer using flat-pack integrated circuits! It was truly a bonkers time for computing.
This machine did run some interesting programs, one of which was Intercom, a sort of virtual machine that sits on top of the original machine and allows you to write programs directly on the machine. It's not as fast as writing bare-metal programs, but it also will keep you from going insane.
As we get close and closer to running, we'll get deep into the weeds on just how exactly it loads and executes programs from the drum, what the instruction set looks like and how to use it. It's going to be a lot of fun learning about this thing!
@@UsagiElectricFlat-pack IC's and core rope memory... something that as much as I understand how it technically works, I still can't wrap my head around it. Truly amazing what these guys (and gals) were able to pull off back then!
The clock is a read-only track on the drum memory... WOW. I'd never have thought of something like that... :D This is awesome!
I can't tell you how excited I am for this series. I melted inside at the CLUNK of that switch.
I'm envious that you have it, not so much the task of getting it running. 👍
This will be AWESOME.
Oh, man, that SWITCH is unbelievably satisfying!
Thank you, I'm beyond stoked for this machine!
And yeah, that power switch is on another level for sure!
When you opened the side door and I saw all that neatly arranged wiring, I actually said "woah" out loud.
Indeed, a work of art! Beautiful. I'm currently building an 8 bit computer out of TTL on breadboard - see Ben Eater - and cutting and stripping wires is a little tiresome, then I saw this and thought man how easy I have it.
The fact that there are so many identical cards (and with two machines in the area as well) makes me think that making dedicated testing jigs for them might be useful. Just a Arduino program (and presumably a bunch of Mosfets or relays) that checks the function of the cards.
Oh, and good luck reforming those caps in the huge capacitor bank at the bottom.
Use a cheap chinesium extension cord in place of a dim bulb for reforming the caps! :-D
Coonect those capacitors across the 50 amp circuit to reform them! July 4 fireworks!
LOL, I was just discussing the idea of building some test jigs for those cards being a good idea before reading the comments.
I suspect this will be a very long series. But it will be absolutely incredible if you manage to successfully accomplish it.
Thank you!
The current goal is to have the AC coming up by years end, and then to execute proper code before the end of next year. I think, barring any intense failures, that's a fairly realistic goal!
In the early 1970's the radio club at the college I attended had a G15 and a pair of tape drives. We nearly got it functioning, but the drum memory gave serious trouble. The project eventually failed due a scratch on the drum that wiped too many tracks. Best of luck with yours!
Congratulations. The G15 and the Dec PDP 8 were the first computers that I learned on in computer maintenance school, and the first computers that I programmed on. I will always remember having to spin the G15's drum memory shaft before powering it up so that the rubber coupling wouldn't break from the motor torque. Enjoy.
The G-15 is cool in that it had a rudimentary operating environment called Intercom which essentially create a virtual machine which allowed working with double precision decimal math. The communication channel allowed up to four mag tape units or four more G-15s to be connected in a network. There was also a refrigerator peripheral sized differential analyzer which held 100+ manually configurable analog integrators, a sort of math coprocessor. There were a lot of ground breaking ideas embodied in the G-15 and the names associated with designing the G-15 are a veritable who’s who of computing founders. It is good to see this bit of history preserved. Thank you.
I love that it pre-dates the convention for hexadecimal, and the manual enumerates 0-9,u-z for the 16 digits!
When I saw the nameplate I said out loud "a Bendix? - a Bendix!" WOW!! I am excited to watch you work on this. I like how you dive deep into everything you work on and keep at it - you don't give up.
You're out there doing God's work, bringing these machines to life, where they most likely would have remained non functional. Historical preservation, in my opinion, is very important. And not just for computers. I love your channel, and I can't wait to see more.
I am really glad too that there are people out there like him who love and respect antiques like this and are bringing them back to life. It's definitely saving a piece of history. Not only will he be restoring the computer itself, but by learning how it works and documenting the project with videos he is helping to preserve knowledge.
Oh, good news. I went and looked at the tech manuals and pictures and the G-15 seems to be made of very reliable parts. I was worried that the rectifiers might be selenium or copper-oxide, but they seem to be very early germanium diodes. Goody.
The filter caps may be fine and just need a slow reforming. A infrared thermometer would be ideal for watching for very leaky ones.
The board capacitors seem to be mostly very high quality hermetically-sealed ones. You rarely see these go bad.
Those early diodes were pretty good too. Long ago I restored a CDC 160-A which has thousands of germanium transistors and diodes. Only found one leaky diode and no bad transistors. Hope your luck runs along those lines.
The tubes are mostly the relatively common 5965 or 12AV7 dual triode. It might even run with a 12AT7 in non-critical circuits. Your main tube problem might be intermittent or corroded tube sockets.
You may need to find a 50-amp Variac in order to bring up the AC voltages slowly.
Cheers,
George
The blueprints are a work of art. Thanks Bob and thanks g15 mystery man!
I absolutely love that everything is hot swappable and every single circuit is hand soldered
As you revealed the computer, admittedly, an Anglo-Saxon explative or two slipped out as my jaw dropped.
Congratulations, man! This is going to be the best project you've done yet 🤟
Nothing beats that old 40s - 50s technology, everythings big, even the wires, and it as aesthetic that is unmatched, it screams industrial and heavy.
I live 20 minutes away from System Source in Baltimore and had no idea the place existed!!! Made a trip up there yesterday for my birthday and got to meet Bob and Brendan and get a tour. Will definitely be going back there. Thanks for sharing!
I presume quite a few capacitors will need replaced by the time you're done, this looks like an awesome project
Depends, some of those might be of such high quality that with a little reforming they might work well enough. If you have to replace them that can be a significant cost, depending on financial wealth.
@Bart Zuidgeest 1956 is deep into tar paper territory. There's nothing to reform about those. Filters maybe, paper definitely not
@@maxs.3238 I meant the big row of filter caps in the bottom.
Personally i usually prefer outright replacement with high quality reliable caps. But this is a museum piece. So an attempt at saving the filter caps might be needed. Or maybe restuffing them. On A museum piece look is important.
The other caps same thing. Replace, restuff, or test and leave in there.
They should be analysed and tested and then a choice needs to be made.
@@bzuidgeest A lot of this depends on what happens if a cap shorts out. I'm assuming bad things. I'm pretty sure they would be electrolytic. For long term reliability, they will need to be replaced. Buying replacements will be $$$, so restuffing might be the way to go. I'm just glad it's not my problem.
Looks like he might need the front end loader again for the replacement caps...
You are definitely brave. I saw what was involved with getting an IBM 1401 going at the California Computer History Museum. Now it did involve more than just the computer but they had a team of volunteers some of whom had worked on 1401s before. It still took them quite a while to get it running. Maybe on the plus side, the 1401 used germanium transistors which were a bit failure prone. Maybe the tubes might be a bit better. I really hope it goes well for you but you are definitely taking on a real challenge. I hope no irreplaceable parts keep you from finishing this project. Really want to see this work. I'm guessing you want that even more.
Awesome! That will be amazing you are ever able to get it going. Going to be a fun, but tough, puzzle, I think. Totally worth it!
I love the aesthetics of the G15 and the attention to the needs of servicing engineers! Reminiscent of old power station attention to appearance which is maybe no coincidence given it’s ample reliance on one!
Great addition to the channel. Wow - over 75 years old! Can't wait to see you get this running. Fantastic channel!
Mad props to CJ for the packing job and the Bob at System Source for sending this thing your way. Good luck, this is gonna be fun to watch!
CJ and crew absolutely killed it with the packing!
I just wanted to say thank you for preserving all these historical pieces. The computer is perhaps the most life changing invention of the modern age, and yet it seems we're all too keen on discarding our history as soon as the next generation of devices becomes available. It's collections like yours that preserve these machines and the history of early computing for generations to come.
As a professional stagehand who constantly deals with casters and caster replacements, I would strongly suggest next time you go to make an adapter plate for casters to not weld the caster directly on. Instead, make a second plate that matches the casters you want to use. It makes it so much easier to change out casters later because casters do wear out more often than people realize. Also, it's really not safe to weld zinc plated caster plates; There's a lot of toxic gases created.
Amazing! It almost looks like a prop from a Steampunk film.
Optical paper tape readers are quite a lot older than this. The first Colossus computer, which was built in 1942, used photocells to read paper tape containing intercepted Nazi ciphers, comparing the code with an internally generated decrypt code. Previous hardware had used mechanical readers, which was much slower and a lot less reliable.
Scrolling to find this comment. I went to Bletchley last year and got to see the wonderful Colossus rebuild actually running, including the tape. If memory serves they invented this specifically because mechanical readers were to slow to do the decoding task in the required time. But then again, Colossus would have been still very much classified in the 1950ies, so Bendix might have had to invent it again. Although Wikipedia says that Harry Huskey, designer of the G15, had worked with Alan Turing on the ill-fated ACE computer before! So it might be conceivable that it could have been Turing himself who "leaked" this little trick!
This is my recolection as well, having been at Bletchley Park and watching the insane rate at which Colossus could pull tape.
1956, when computing power was measured in kilowatts!
It's awesome to see the guts of some of these classic computers.
I love the smell of an old machine, they have a special scent...like it gives you the vibes of olden days
This will be an epic restore. Bob is a super awesome guy, I love visiting System Source, he is such a wealth of information, and he loves to talk about his collection, and is a wonderful guy to just listen to! I look forward to seeing this work, I thought this machine looked familiar when you first showed it, when you said it was from Bob, it all clicked.
The nice thing about having a bunch of ancient computers in need of fixing on a youtube channel dedicated to fixing and showcasing ancient computers is seemingly self evident.😊
This is SO COOL!!!! This hits some of my favorite era of electronics and computing! Between the EM bits, and the sheer approachable nature of how everything is built, it's gonna be EPIC to see this whole project go!
THANK YOU!!!!
You should slowly charge the capacitor bank through a variac at reduced voltage, then check them for shorts and leakage current.
The green/blue cringle coat paint has always been my favourite industrial colour.
Great video! Glad to help with the casters in my tired stupor! Hopefully us over at VCF in NJ can get a G-15 working so there can be 3 total in the US!!
For its age, it's an exceptionally compact and well designed machine. Fascinating stuff!
I don't know anything about vintage computers, but the great beauty of this behemoth is undeniable. The combination of computer tech and the physicality of those buttons, switches and dials are immensely satisfying.
This will be an amazing journey, thank you for committing to bring back to life this awesome machine.
Love that power regulation with built in variac and rheostats. 50 amp service!! Wow.
Looks like a bunch of selenium rectifiers in the tape punch/reader.
Pumped for this series to unfold. 😊
This will be an amazing journey, what an amazing piece of history
Beautiful. I was born in 1956. I would never have thought that something like this existed back then !
Fantastic. I am looking forward to the next episodes. Major work cut out for you. I love this channel. ❤
Usagi Electric - taking unboxing videos to a new level 😊
Looks like another super fascinating project you’re kicking off here. I know nothing about Bendix machines.
I was watching a video from the Computer Conservation Society about the Elliot 401 from the early 1950s- that also used a drum for main memory and also used a track on that drum for generation clock signals. I guess it was a good option on systems of this vintage.
So cool. NASA used these as ground support equipment for some of the earliest space launches. If you take the Space Force Missile museum tour you will see one in the blockhouse at Launch Complex 26.
Your love for vintage computers is amazing. Thank you for sharing it with us.
kind of wishing this was a video that was 2 years old and that i could binge the whole series
I'm happy to report that I'm someone who owns a tractor and loves vintage computers. It takes a special set of skills to work on a thousand pound computer! We are lucky to have you!
I'm so glad to hear you have a friend with one of these, because that will be a yuge help when figuring out those added-on bodge wires. All good wishes.
shine on you crazy bastard, i'm on board for the whole series
Holy moly, now *THAT'S* a project! And two of them!? Fantastic! Can't wait to see how the project(s) go!
Thanks for the usual top tier content! Always makes my day to see your enthusiasm and intellectual thirst.
(Apologies for being a wet blanket)
A while ago I was watching a machinist video (I watch a lot of them). Lathes, Mills, Surface grinders etc are really heavy.
He was talking about castors. He pointed out that while you might *like* to say that with 4 castors each castor is taking 1/4 load, it's likely not the case, unless your floor is a surface plate.
As everyone knows, 3 points of contact is all you can really guarantee, which is why photographers use TRI-pods.
But it gets worse; call the corners A-B-C-D, going round
You might hope that even if you don't get 4 point contact, you might at least "fall back" to a 3 point configuration (say A-B-C). But (assuming a roughly even load) it's sort-of obvious that this A-B-C contact could easily be tipped to become a A-D-C contact. A-C is a "high diagonal" (either A or C must be high) and the load can pivot on this A-C axis.
Which means, at least momentarily, the A-C castors are taking *all* the load.
Summary: your castors (and floor) need to be spec'd for at least 50% of the total item weight.
Thankfully it's a concrete floor but those poor castors might get a bit less round over time.
I don't know why I'm excited for this series but I am :D My country was behind the iron courtine in this era so this is something very exotic to me. This behemoth you have was built before even first computer in Poland was designed not even talking about mass production.
The Electric company is probably looking at your account like a kid in a candy store.
Omg this is going to the most epic restoration ever. They only made 400 of these, and there’s only a handful of them left in the world so yes, it will probably be the only fully functioning one unless like the Smithsonian has one or something.
That's what I call equisite 50s Ero*ic! The Smell, the Sound and the Mechanics are definitely amazing when it runs!!
I’m not sure if you already have a solution in mind for providing a 120V 50A circuit, but I have a suggestion.
Rather than heavily loading one phase of your electrical service, you should try to find a cheap used 240X480V to 120/240V transformer. The primary can be jumpered for 240V, and the secondary jumpered for 120V. You’d be looking for a 7.5-10 kVA class transformer. Dry type or a potted enclosure that is meant to be wall mounted.
This would allow you to run a 30A 240V circuit to the transformer to step down to 120V. From there hit a 50A fused disconnect switch or single breaker in an enclosure (it could be a 2-pole breaker, just leave the second pole disconnected). If the computer has its own 50A breaker, then you could just leave the external one out and hardwire the transformer secondary straight into the computer. I would use a 240V 30A twist lock plug and receptacle on the primary side of the transformer so it’s not permanently hardwired to your building. Slap it on a rolling cart!
Make sure to ground one leg of the 120V secondary to be the neutral, the ungrounded leg will be the hot conductor to be switched or fused.
This is all under the assumption that you service is 120/240V split phase.
If it’s 208V, then obviously the transformer will need to be different. Some transformers have taps, for example a 90% tap for 240V would be 216V. If you connected that to 208V, the secondary side would be at 115V which would probably be fine if the computer can handle it.
Best of luck reviving the G-15D. I tried to get one working as an undergraduate. One of the primary problems was that the diodes developed high reverse leakage. A goodly number of them had to be replaced. I did get it working and compiled a number of Fortran programs on it.
My grandpa used the G15 for aircraft work at the Teterboro airport in NJ. What a beast. Hope the drum is not crashed.
When you have the 2 "beasts" running then you need to connect them over the "interweb" and have them running some sort of 2 player game. Just an idea.
5 level paper tape! Optical paper tape readers were used in both Heath Robinson and Colossus at Bletchley Park during WWII
2× G15's. You can network via Paper Tape and USPS.
Wishing you the very best and I will be following you :)
I have lived within 30 minutes or so from that museum for most of my life. Why am I just hearing about them now? I know where I am going on my next day off!
Aren’t caps from that era filled with nasty PCB’s? You’re probably aware if so but if not check. You don’t want to mess with those without taking the proper precautions. Great project! Can’t wait to see it running!
It's something like 20 years younger than the ban on the use of PCBs (polychlorinated biphenyl) so it would make sense that they'd be used extensively.
Not the electrolytics. Some paper caps were PCB oil caps. You can smell it.
@@oliverer3 PCBs are mainly but in the AC capacitors, such as the run capacitors on motors or AC line conditioners, but could be in smaller oil filled capacitors. If a capacitor says Pyranol on the label, there is a good chance it is PCB-bearing oil filled. Electrolytic capacitors have polysaccharide dielectric material with additives. I don’t know what the additives are, but I wouldn’t count on the additives being harmless.
Thank you!
All of the caps on the DC side are high-quality Sprague electrolytic. These are the same ones that were used in HP test equipment in the 50s, and I've found them to almost all still be good and safe to use. The only thing that is a mystery is if the drum motor has a start capacitor and what it is. That particular one may need to be replaced if it's a PCB filled cap.
@@UsagiElectric The G-15 is a beautiful machine and it is great to see it in such capable hands for its restoration. It would be interesting to see if any metal paper tape cassettes with canned programs can be found, especially ones for POGO and ALGO. In my scrap electronics adventures, I’ve never seen anything like those Bendix cassettes in 50+ years. I took a quick read through of the programmer’s reference manual, and boy would writing just about anything from scratch be a tedious undertaking. The system reference manual also contains a flow chart which refers to a self-test program to be run at power-on after the initial voltage checks.
Sometimes small-ish capacitors can be suspect. I ran across a 4.7 uF axial lead capacitor with cardboard sleeve and red rubber end caps in a 1946 National HRO receiver’s audio stage. The sleeve was wet with oil and there was oil on the bottom cover of the chassis. Usually a cap like that would have been wax paper and foil. I happened to have a WW-II Sprague fungus proof cap in a plastic shell which was about the same size, so I swapped that in.
The G15 was my first personal/home computer. You did an excellent introduction and if you would like to connect to discuss the repair, operation, and programming of the G15, I encourage you to connect with me (I can share the challenge areas I navigated such as the master timing circuitry). Sadly, my G15 died (its an interesting story) and was disassembled in 1979 but I still have a few parts. In addition to the G15 & console, I had 2 MTUs. FYI that small corner is going to get REAL hot ---- too hot to keep the system running smoothly. As the plate voltages drift due to excess heat the system will misbehave. I look forward to future videos & possibly connecting. You've made me quite nostalgic for a simpler time.
WOW! Never saw anything like this and so close! Good luck with repair! Eagerly waiting new video about g15! 👍👍👍
Dude, this is a HUGE unboxing video... looking forward to see this beauty finally work well...
Man, you are my hero! When I saw the plastic wrap inside the big box, I knew, "it's a Bendix!", due to the shape. Thank you very much! I'll see every part of this saga!
This one is another level. I never knew that rotating drum memory had been commercially used. Woah.
As I saw the thumbnail, a thought came immediately to my mind:
" No way he f*cking got a Bendix G-15"
Congrats!
Awesome growing collection. Your room will be nice and toasty in the winter with the Bendix and a few others running at the same time.
Can’t wait to see the upcoming videos on this. The G15 is one of those mythical computers that I’ve always wanted to see.
So cool that you're working on bringing this machine back to life! My uncle has told me stories about his days programming a Bendix G15 back before computing life was easy. Looking forward to seeing you at VCFSW!
When I was a high school senior (1965), we received a G15 for a new computer class. We were told it was the very first computer in the nation specifically for student use. I don't remember much about programming it any more, but have always triggered on them wherever I ran into an article, or picture. There is one at the Smithsonian National Museum of American History, but I don't know if it is operational. What surprised me was that it has a mag-tape drive attached. I really look forward to watching your progress on getting this system up and running!
Wow!! That thing is a monster! You are courageous Usagi!
This channel just gets more and more awesome. Thanks for being so ambitious!
This channel is getting better and better by the day 😮
I am very glad you are documenting your journey, I love watching how the systems work! You and other creators have inspired me to restore my own devices!
Wow, man! The Bendix is nothing short of spectacular. I especially enjoy the analog dials.
I'm actually pretty excited by this thing. I love how the power on process is a whole launch sequence. I'd love to spend a day just... toning out all of the connections to the boards to make a map of it.
That is beautiful and over 600 kilos! You would not want that to roll over your foot!
Usagi Electric: Great taste in cars and computers. Forlkift certified.
Good luck with this project! I'll be watching your progress.
That G-15 General Purpose Digital Computer is serial number 238. How do I know that? My father J. H. Harrison, and then my family, owned that baby from 1967 until 2021. It was in my family's basement in a suburb of Washington DC and was last operated in the early 1970s. My dad purchased it in 1967 from Philips Laboratories in Briarcliff Manor, New York. He purchased this unit and some others, and many G-15 related items, when the G-15 was being phased out by businesses and schools. My brother contends that my dad was one of the world's first computer hobbyists. As you mentioned, 238 is now owned by System Source Computer Museum in Hunt Valley, Maryland. My dad's siblings are big fans of them and their museum that can rival any computer museum around. Good luck and have fun with the restoration project. We will be watching the progress.
You're absolutely right, 238 on the nose!
I've heard a lot of about the Harrison family in relation to the machines, it's an honor to finally hear from y'all! If you don't mind, shoot me an email at "Nakazoto at gmail" or join our Discord chat server ( discord.gg/p7UsfHD ), I'd love to hear some stories of this machine and the other G15s y'all had. And also, hopefully you can solve a mystery on this machine that's got us scratching our heads!
Worked on the Navy's NTDS systems in the 70s. Paper tape, flex writers, 800 bpi tape drives, drum drives. 32 k of 8 bit core memory. Those were the days. Had to be very careful with memory ussage.
Looking forward to the rest of the series! My dad started off his career on learning on valve mainframes in uni in the 60s
"FLASH: There is another." -- Colossus, The Forbin Project.
HEAT! At 38A, 120V, that's 4.5kW so that's like having at least three typical 120V room heaters going flat out at once - toasty! Good luck though, it's going to be fun winter project, time to duct that air around the house... and to the neighbours even!
I am in awe of this, but also glad my only task here is to watch you figure it out. Because I wouldn't have a chance. 😂
It's going to be a blast (hopefully not literally) bringing this big beast up!
Wow! I don't think I've ever seen anything, let alone a computer, with it's own built-in variac before!
This is just an amazing machine. It's the same age as me! That backplane reminds me of the DECSystem-10 KL-10 backplane I worked on back in the 1970's - complicated. But vacuum tubes and a drum memory? This is gonna be quite a challenge. Good luck! I'm looking forward to watching your progress.
Knowing you're likely younger than Bob means that my hypothesis is true that there will be young folks like myself interested and invested in obscure old machines even as these things reach 100+ years old. And having access to stuff like these videos (hopefully archived) will help too. Having it in a museum will hopefully prevent this stuff from being lost to time.
Ya know, seeing your Lexus haul all this makes me more confident when I load up ours with old equipment.
Wow!, I wrote some of my first programs on a UC Berkley G15 back in 1961! Wish i was closer to help! Good Luck chasing all the bits around that drum!