Exploring 1950’s Computer Logic with the Bendix G-15!
ฝัง
- เผยแพร่เมื่อ 29 พ.ค. 2024
- We’re back on the Bendix baby! In today’s episode, I want to take a look at all the unique logic cards in the system, but before we get that far, we gotta get our clean on. Hop along as we take every single card out of the Bendix G-15, clean each one painstakingly, then repopulated the whole machine, and finally take a deep dive into what each card does. We also try to answer a pretty important question - is this an NC or an ANC machine?
Check out System Source here:
museum.syssrc.com/
Here's the Bendix G-15 Theory of Operation Manual:
www.bitsavers.org/pdf/bendix/g...
If you want to support the channel please hop over to Patreon:
/ usagielectric
Also, we now have some epic shirts for sale!
my-store-11554688.creator-spr...
Come join us on Discord and Twitter!
Discord: / discord
Twitter: / usagielectric
Intro Music adapted from:
Artist: The Runaway Five
Title: The Shinra Shuffle
ocremix.org/remix/OCR01847
Thanks for watching!
Chapters
0:00 Working on the Big Blue Beautiful Behemoth
1:00 The game plan
3:25 Taking every card out
5:58 If you’ve got time to lean, you’ve got time to clean
7:56 Door latches and modifications
9:18 Putting it all back together
11:04 D2 “Diode 2” module
11:56 Card edges and replacement diodes
12:48 D1 “Diode 1” module
13:35 D3 “Diode 3” module
14:20 The BI and FF cards
15:36 BI “Buffer Inverter” module
16:39 FF “Flip Flop” module
17:12 And aside about diode clamping
18:05 DC “Diode Clamp” module
18:35 CC “Clock Clamp” module
19:42 CF1 “Cathode Follower 1” module
20:54 CF3 “Cathode Follower 3” module
21:23 CF2 “Cathode Follower 2” module
21:54 RA “Read Amplifier” module
22:44 WA “Write Amplifier” module
23:42 RC “Read Clock” module
24:44 Alphanumeric (ANC) vs. Numeric (NC)
28:44 Scouring schematics for the answer
32:49 What did we learn?
34:57 Happy kitty!
Here are all the modifications:
Blue
B3T - C7K
C8K - K5P
C7N - A54P
C7V - C8V - A54A
K34J - Loose wire
B54P - Two white wires that disappear into loom
Red
C7C - E48R
C7P - F51P
C7R - B36F
D16D - C54R
D16E - D35V
D16F - J15B
D16S - D17R - E14B
D17E - A28T
B41k - K47V (Broken connection)
B41V - H50P
Yellow
K7D - D30S - วิทยาศาสตร์และเทคโนโลยี
This is the first computer that I learned to program. This was my inspiration to learn about computers. I had a long career designing computer hardware and software. I will never forget as a freshman seeing this computer function. I was so amazed at watching a typewriter print by itself. At that time, I was a freshman in electrical engineering at the University of Delaware. I programmed in a language called Intercom 500. It looked like today's machine language. UofD within a year replaced it with an IBM 1620 where I extensively used Fortran for any programming needs for my entire college career. In my last year of college, UofD had an SDS 9300 scientific computer which was infinitely faster than the G-15. SDS, Scientific Data Systems. Thanks for this video, it was great. I never did see the inside of that computer. By the way, I seemed to think that the typewriter was an IBM Selectric with a rotation ball but it is possible I am wrong. Maybe you could see if that was an optional typewriter for this computer. I do remember that it had a rotating drum for memory. My first experience with the G-15 was in the fall of 1961,
The tape reader unit shall always look like a coin slot mechanism to me... "Please insert 25c for the first 100 clock cycles!" 🤣
i think it every time i see it
I am known as the guy who has the ultimate answers to questions that are no longer asked. Not only do I have a "Standard issue" SloSyn paper tape reader that reads at 125 bytes per second, but I also have a super speed asynchronous paper tape reader with a built in data buffer that reads paper tape as fast as you can pull it through the reader's head and ships the data out serially at up to 115k baud. Of course shorty after it came to market paper tape was going the way of the Dodo. Just one of my ultimate answers to irrelevant questions. 🙂
That sounds like a concept that IBM would operate.. pay per punchcard..
Please deposit six million dollars for the first three minutes...
I'm gonna have to start carrying around a sack of quarters with me every time I want to use it, lol.
That machine was built 30 years after the advent of broadcast radio, and the start of the electronics revolution. Meanwhile, we're now 70 years since that machine was built...
It's an awesome time capsule to see just how far we've come!
i find it hilarious that their idea of "buffering" was just "amplify and crank that shit up to wumbo so that anything less than a lightning strike won't degrade the signal"
this is absolutely fascinating, subbed for more
The "amplify the hell out of it and then clamp it" is a pretty effective way to reduce rise and fall times, which would be important since the circuit uses a passive pull-up instead of the push-pull amplification that solid-state computers tended to use.
My high school was donated one of these for us to explore computers and programming. I am now retired after 40 years as programmer/analyst for mainframes, minis and microcomputers. What a happy circumstance for me!
That's awesome that you got to use one for real! I can't wait for the day I get to start learning how to program the hard way on this thing.
It will be a challenge. Best of luck to you.
Carl Sandburg HS in Orland Park, IL by any chance?
@@hieronymus9 nope. In northern California, 1962.
On the bright side, removing and reinstalling every card has probably 'wiped' years of oxide from the connectors, eliminating one source of faults. (we had some edge-card machines from the 70s, when in doubt 'reseat the boards' lol )
The best way to clean the exterior is to retr0bright it! Just kidding. Don’t you love that initial detective work that draws you to the internal minutiae of a seemingly incomprehensible machine. And when you finally unravel the mystery it’s immensely satisfying. Superb cleaning and first inspection job.
I think there's a bit more going on in the buffer-inverter than just 3 inverting amplifiers. The second and third tubes share a cathode resistor, forming a Schmitt trigger. That would help with producing sharp output transitions.
An old computer with thick cabling and connectors with hundreds of pins in them. 😊! Will ❤ seeing that old Bendix running again!
With regards to the typewriter it appears to be 'universal', there was a printed label on it saying only to connect to either NC1 or ANC2/ANC3 machines. Which makes me think that it does support numeric-only machines natively, and perhaps the relay box is only used for ANC1 machines.
Ah, I think I caused a bit of confusion with that shot.
The original, early NC only typewriters had only something like 24 solenoids underneath and didn't need an external relay box. However, any ANC compatible typewriter has a solenoid for every single key and therefore needs a relay box to connect it to the appropriate machine. I believe, in order to connect this typewriter to an NC only machine like we have, we need the NC-1 relay box, whereas instead we have the ANC2 relay box.
So, the issue isn't the typewriter itself, it's the relay box that goes between them.
@@UsagiElectric Your terminal (console) is connected via relays!
The sounds so counter productive, they moved from relay computers to valves for improved speed, but then feed it into a bank of relays.
@@UsagiElectric That makes sense to me now, as I made the same comment before reading down further :).
Hopefully you have the correct relay box, and hopefully this machine has not been reconfigured to be like a ANC 1 which the typewriter says not to use with
@@paulstubbs7678 That's the way this one works. I believe, from watching this video and few others, the relays are used to store / convert to the two fast 5 bit codes into a single 10 bit code used by the typewriter
@@paulstubbs7678 My guess would be that input/output to a type writer is quite slow, så relays are sufficient, and they were probably much cheaper than tubes at the time. But I can't imagine the noise. Looking forward to that video!
Old style resistors brings back memories when I savaged junked tv and radios for my electronics projects.
What a herculean task; you're amazing! BTW I love that crinkly paint finish that you just don't see anymore.
Kinda hoping that paint texture (in '50s colours, naturally!) makes a comeback.
I can still buy it in the store, so it isn't gone.
Those 'crinkles' are used to mimic the fine texture fine cast parts would have. It als makes surfaces quite grippy. The pitted appearance of Hammerite paint also mimics a cast look.
Seems like in the past making things look cast wad a whole fad just like it is today with carbon.
The paint looks like the pattern was somehow rolled or stamped into it, it's so even.
Thank you so much!
That crinkle paint looks really gorgeous, hopefully we can get it cleaned up and looking spotless in the next episode!
Some of Beta Iota Tau at Rose-Hulman found a G-15 in a basement (I forget which building, this was late 1970s!).
And we actually got it working!
I think, the conversion of this machine to alphanumeric was started, but never finished before putting it out of service.
Maybe just bringing the signal out to the barrel connector would do, but I also would expect that the machine would have different (paper tape) software to support alphanumeric.
i was going to say the same thing that i believe that conversion was performed but not fully finished as it seems like the parts are there i would consider finishing it as well 😊
0:48 Seeing an Xbox 360 in a vintage computer museum hit me in a place I didn't know could hurt
By now that actually qualifies as vintage. Tech moves fast. Even a PS4 is ten years old by now. And that's a later generation console then the Xbox 360. 10 years is an eternity in tech. The bendix is like a dinosaur on the tech time scale.
I just found out today that there is already a successor to the PS4...!
The read amplifier looks simple enough to me. The clock pulse on the right resets the flip flop and the read clock will sample the read head signal to determine if the flip flop should be set. It's roughly equivalent to a D-type flip flop. It's to avoid stray pluses on the read head causing a ripple effect though the machine.
That makes sense!
Sometimes I look at a circuit and my brain just goes "not today, son." Thanks for the kick in the right direction!
This is amazing. That typewriter interface with the relays and 10/5 bit conversion, mind blown...
Thanks!
And that relay box is wild indeed. When we get deeper into bringing the machine up, we'll dig into the typewriter and relay box in much more detail and really start learn how it works inside for sure!
Yes, I understood right away some kind of memory would be needed, but I hadn't thought of using latching relays for it. I wonder how noisy it gets...
Oh! This is a question I can actually answer the weird indexing on the schematics pertains to how they are indexed in the real world for the technicians. The schematics had to be professionally printed. (Mimeograph is a few years out still) they index that way because multiple copies were likely stored in a physical index so that if you were going on a service call you could pull just the copies you needed (and then return them.) There was also likely a printed book but that was more or less likely indexed based off the drawings id (as that would have been produced by the designers)
Oooh, color me interested, but I'm not quite following yet.
The schematics were indexed to what? Why would multiple copies of the same schematic have a different index, or if they didn't, why not just use the schematic number as the index? It's super confusing, but sounds like you might be able to bring me up to speed!
@@UsagiElectric so I'm not an archivist (but I do design and draw maps for a large MSO). Back in the day field engineering departments used large indexed map drawing storage systems for keeping documentation typically in A0 format ( that's like 33 x 46 ) they have to be rolled or folded and stored in large hanging folders. This takes alot of physical space. There might be 2 or 3 copies of each drawing. Because the drawings weren't likely to be mimeographed at that size or predate mimeographs. The drawings had to be indexed by whatever system was in place by that specific archivist, and that might differ from the design number. Kind of like all libraries use a similar system for organization of books. Its possible some Bendix design numbers at one point coincided with the field index but at this point the design index and field index appear to be different. Field drawings often also have a grid index to show what drawing is next, in the case of telecom maps the drawing numbers line up on a grid so you can tell what map is to the 8 directions surrounding it. I have seen large circuit drawings done this way also. Obviously somebody who's familiar with bindex processes can tell you more specifics but I'm 90% sure that's what's going on here.
Watching you clean that amazing machine does make me feel like i should clean some electronics too. Don't have vintage things around right now so a microwave will have to do. My friend said it looks filthy... Also the fridge..
Classic. A lot of early computers used diodes for the logic, and the tubes or transistors were only there to boost the signal or boost and invert the signal. I mean there was an entire class of logic chips that people tend forget about (or just not know about entirely), and that's DTL, or diode transistor logic, the precursor to TTL.
I might be wrong, but i think the Apollo program computer in the lander was built using dual 3 input nor gates, using DTL. Some of the first digital IC's made. Again, could be confusing my history, but it's interesting nonetheless.
I clean 45 year old machines daily at work and your technique is spot on! I combine the paintbrush with a hepa vac and use a little contact cleaner around the electronic bits and stubborn gun from wearable rubber components, doesn't seem to take off paint! Otherwise I am new here and I love the content, I'm sure the original engineers would love to see their work getting this much appreciation. Thank you!
Great video. The crinkle paint you have to be extremely careful with. Some become gooey over time and using the wrong product could remove the paint. A test with Dawn dishwashing liquid, water, a non-abrasive cloth, and testing on an inconspicuous place would be a great idea.
At a planetarium I worked out in the late 80's, had a Spitz 512 digital planetarium projector and the projector and the control console used the crinkle paint and we were warned by Spitz that the crinkle paint could be a problem, as they had issues with one particular paint manufacturer's product during production. They recommended only soap and water and gentle pressure cleaning. There was no Dawn Ultra back then, just the regular old Dawn, which may still be manufactured. I remember seeing it at a dollar store a few years ago.
If the signals exist in the machine, maybe you can bring them out to the cannon connector. Good luck!
That's definitely one potential path we're thinking about!
There's a lot of modifications in there, and some of the modification wires are on those specific ANC related modules. I think the first step is to confirm what is wired to where, locate that within the schematic and start to figure out what was modified. If it was specifically for the ANC upgrade, then we can finish the upgrade and make this an ANC compatible machine. However, it's possible they were using the same ANC related tube modules for something else.
That's something we'll get stuck into in the future!
@@UsagiElectric wonder how you add a wire to a wire loom without removing all the strings making it into a loom!
@@mohinderkaur6671I would guess carefully, maybe with something like fish tape
Looks like the read amplifier does pulse/edge detection on V1A. The trim pot sets the transient sharpness treshold.
V1B gates this signal to the flip-flop on negative read clock, and then I guess the flip-flop is then reset by the clock.
In short, it detects any asynchronous pulse by the transient edge, and converts it to a digital pulse synchronous with the two clocks.
WOW. 4:30 That's a lot of cards, but they look awesome laid out like that.
I was a Field Engineer with Bendix in the 1970s. I was in the airborne sonar division at Sylmar, California. We had a computer system used to run Acceptance Test Procedures on sonar that was mounted in SH-3 Helicopters. The computer was a beast. It weighed 55 pounds and was quite large. Back then there were components that were so large we had to solder them with a soldering gun instead of a soldering iron. Ah, those were the days.
Usagi Roadmap:
- Zuse Z3
- Programmable Loom
- Abacus
- Stone age pebbles
- The turing machine that runs the Universe
😀
What I find most mind-boggling is that the processing power of these massive machines is literally dwarfed by anything that is the size of a grain of sand and costs a few cents today. That's probably the most important reason to keep these machines conserved and running: To teach our kids how computers started and were developed.
Children from today take computers absolutely for granted, and it's kind of important that these machines were designed to save lives, namely, to win wars and protect us against threats from enemies.
These machines were post WWII and weren’t military. The G-15 was targeted at scientific and industrial tasks. WWII era machines were used in code breaking, firing table calculations and for the calculations needed for the A bomb, they were few and pretty specialised.
@@stevetodd7383 Even post WWII, they did serve an important purpose in the Cold War. Maybe not directly as military machines, but they were probably used somewhere in the space race, if they were used for scientific tasks.
Being generous with the classic CPU hardware, your typical Casio smart watch (not those Android OS version, rather I meant classic ones - that was cheating. LOL) has more computing performance than the Bendix G-15 in this video.
“Win wars”, “protect us”, “enemies”. Wow. I wish you a long and happy life mate
@@lsfornells We are so much better off today. There are no more wars, no need for protection, and all the world are friends.
Cleaning the wrinkle-finish paint on the outside should be easy: You can use a mild detergent and a toothbrush to work the dirt loose, then wipe it clean with a soft, lint-free cloth. Thanks for the fascinating look at this beautiful machine!
I've always loved this machine, and the more I learn about it's designer, Harry Huskey, the more impressed I am. He was arguably the most influential pioneer that few have ever heard of. I'd be lying if I said I wasn't a little envious of your opportunity to work with it. I look forward to future installments.
I'm not super familiar with vacuum tube Electronics, but from the schematic you showed for the read amplifier card, it looks like it's an edge triggered monostable. Meaning that on a rising or falling Edge, depending upon how it's wired up, that circuit will generate a single pulse on the output of a fixed time with. That's why you see the Transformer coupled input driving a flip-flop. The Transformer coupled input makes it so that you will only act upon a rising or falling Edge, again depending upon how it's wired up, and then the flip flop is made so that you can with a single input trigger a single pulse on the output..... at least, that's what it looks like to me. For what it's worth.
u-z is only one such notation, there were more than a few, and had to do entirely with mapping onto given keypunch and teleprinters in a particular setup. ;)
Fun Fact: multimeter leads;
A female 4pin molex pin can be used a to connect to the end of a standard multimeter probe so you can make custom multimeter ends (I made a pair of alligator clips for mine)
(@4:13) You doing the fast-forward there if you pulling out the circuit cards reminds me of the ST:TNG episode where Data gets “drunk” and pulls out all the isolinear chips from the main engine computer. LOL. 😅
I find the best way to clean industrial crinkle and hammered paint is just to get in there with soapy water and a brush. Kinda like cleaning textured vinyl and such. Anything like polishes and such just leaves a white residue that you end up scrubbing out with soap and water or naptha and wishing you had just done it that way in the first place!
After pulling out the cards, I would have taken the opportunity to do some continuity checks on the card slots, and checked them off against the wiring diagram. That is, if you hadn’t “painted yourself into a corner” there by the pinball machine! This way you can note (& fix) any bad edge-connector pins, bad solder joints, or just dirty contacts. Doing it now will save time later, or you’ll be like Rainman Ray (Rainman Ray’s Repairs YT channel) when he says, “I love my job so much, I do it twice.” 😏
Just interface the G-15 to the pinball machine. Power it off a steam engine. Steam punk tech
Its not so many card slots to check, you can start checking the slots and i can help out in a bit ;)
If he has the 'wire list' that shows all the connections between slots. (he may have them, don't know). That would detail each wire point-to-point connection, making it pretty straight-forward to ring each one out.
Trying to check using the schematics would be a lot more difficult, hunting through sheets and sheets looking for every connection.
It's so cool that you are going to try to get this machine running! The Kinkaid School in Houston acquired a Bendix G-15, serial number 186, in 1969. It was donated by a local engineering company and formed the nucleus of the first computer club. Unfortunately, over the next two years it became increasingly unreliable and the students in the computer club found themselves spending most of their time troubleshooting and repairing it. I entered ninth grade in 1971, too late to ever see the G-15 in operation. We used an ASR-33 to dial into a timesharing service instead. I was always fascinated by the old machine, which was still sitting silently in the corner of the room. I hope you get this one running!
This, along with your other videos, encouraged me to visit System Source this weekend for their repair and boot sale and open door to the museum, where I met Bob Roswell and was taken around the place. What an amazing museum he has set up! Bob is an incredibly generous person that was wonderful to visit and talk with, thanks!
That's awesome to here!
Bob is an absolute legend and I'm glad you got a chance to meet and hang out with him, if you head back up that way again, give him a high five for me!
@@UsagiElectric Thanks, will do! I have some tube goodies to bring to him for one of his racks that have some sketchy tube cover/retainers and some spare tube for that rack box. Love your videos and have been learning a lot from them!
I wonder about the wisdom of taking all the cards out at the same time. Yes they are keyed, but like the backplane, there might be cards with modifications on them. If you don't realize/miss this, you might not put it back in the same place and you might have a difficult mistake to figure out. Some of the tube modules had stickers on them that again might not make sense in any other place than the original one.
Better to take them out one by one and replace them before moving on or at least mark every single one so it goes back where it sat.
That said, it's done by now :). It will be very interesting to see it progress.
Clean always seems to make stuff run better.
the design of this machine is simply enchanting
thanks as always for sharing your cool stuff with us
What an amazing machine. I started on computers in the late 70's with the TRS-80's, and had no idea how complex computers were from the 50's. Thanks for sharing.
On the Amiga, the Bus Error signal is called BEER instead of BERR. I guess for the designers, it meant "Let's have a break and think about this. Again." So what is that POOP? "It's spilling out random sh*t at the other end"? 🤔
I still think the computer has a coin slot.
Haha.
Wonderful video, it was an honour talking with you while you were cleaning the cards
The traces on those cards are a thing of beauty!
Mylar film and plastic traces stuck to it to make the pcb mask
The shot of all the modules and boards arrayed out is great and a really rare thing to see, thanks for taking the extra time to set that up.
Regarding cleaning... i use a particular bar soap that's in the hand soap section when I need a soap that rinses clean with tap water, it's Kirk's unscented castile soap. You can either dissolve a little in one container and have a second container and sponge for rinsing, or just rub the cleaning implement (brush/sponge) on the bar itself once it's wet. Does a remarkable job removing "regular" gunk and grime.
The 1950's was a time before static was a thing.
@@tripplefives1402 It also doesn't matter with discrete components with their elements measured in millimetres vs nanometres
@@tripplefives1402 Also the amount of current normally carried is far greater than what static electricity could deliver. This is untrue for the innards of modern logic chips. Now it's possible that static could blow one of those germanium diodes. It would be a good idea, if Mr. Usagi hasn't done it already, to ohm all those diodes out in both directions. The voltage margin does make me a bit uncomfortable. Modern Schottky diodes could be substituted to come up with something more robust.
One of the great things about vacuum tube computers is how robust they are to transients and low current stuff!
static? I ber a dollar to a donut this beast would survive an EMP
Heh. Before they cared, certainly. A good pop could zap some of those diodes, esp. as old as they are now. (the zap you get touching a doorknob is enough, but I doubt that's a problem for him.)
This kind of stuff is always neat to see......... I restore stuff too, but not old computers. My style is to restore things I can actually still enjoy using... and Old school computers aren't my fun zone. Much respect and good luck on your quest.
That crinkle paint looks like the same kind used on some typewriters in the '40s and '50s. I have a typewriter or two with that sort of paint, and I generally use Windex or Simple Green to clean those. It's worked quite nicely for me; they seem very forgiving on paint.
Yes, it's just ordinary paint that's been force-dried under heat to give that texture. Windex works fine on it.
Love the level of detail - I won't say I can follow the intricacies - but its fascinating to see these older designs up close. Thank you
All those beautiful Allen Bradley carbon composition resistors remind me of my early days working for that company.
I love that crinkle paint from that era. I was reading up on it a couple of years ago, & it turns out that it's just ordinary enamel, & the unique texture is a result of cooking it dry very quickly.
TL;DR: Windex & a stiff (nylon) brush should make short work of getting the dirt out of the crevices, followed by a thick, damp towel.
AMAZING that this machine worked at all. With SOOOOooo many ways for it to fail!! All those connector cards. Thanks for this video. This really shows how far we have come.
Watching your videos while working is very motivating: Seeing the problems you have to solve in this old beasts makes me realize this compiler error I am getting is not that bad after all.
Should imagine that the purpose of the S-R flipflop on the read amplifier is to hold the data read in in between clock pulses, the transformer would provide the amplification stage,
The first computer I ever had to service was an analogue computer, with valve (tube) amplifiers. It ran a flight simulator😊
Oooh, that's awesome! Was it the K2-W Philbrick based analogue computer? They have one of those at the museum in New Jersey at VCF East, and it's a work of art!
@@UsagiElectric It was a British built machine, so probably Singer Link Miles. It ran the Folland Gnat simulator.
After that I worked on many simulators that were hybrid digital/analogue machines and experienced the evolution to the digital age...
its amazing how much have improved from the 70s, that is a computer... i dont want to try and poke around the the ui though. It seems complicated.
Love the hand made cards that populate the machine. that is pure skill and no cheating.
Wow so with all the cards laying on the floor .it looked just like a magnified shot of a modern day chip
To clean the krinkle paint surfaces, go the dollar store and get concentrated La Awesome, dilute it 1:3 parts water, and saturate a slonge with it.
That has been the most effective for me.
Sounds like a nice idea, just be careful of the fumes (there's a lifetime maximum exposure rating on the chemical used, believe it or not) and test it first where it won't be visible if it does mar the paint.
@@SeekingTheLoveThatGodMeans7648 Citation? After looking at the SDS and checking out the components, I was not able to find any evidence of a lifetime exposure rating on anything in it. I would be shocked if anything consumer level would have such.
I very recently visited the Computer History Museum in Mountain View and it felt like watching your videos
Your video production quality never ceases to amaze me
10:00 Putting the cleaned boards back into the system reminds me of the scenes from Star Trek TNG when they used the isolinear chips in the computers. Also I'm glad to finally hear the difference between "numeric" and "alpha-numeric". I've only remember Alphanumeric being the catchphrase of Enzo on Reboot.
Haha Reboot. Classic. Goddamn megabyte
I was actually glad to see that the card edge connectors were double-sided but single-circuit, since it makes them a lot more reliable. But it still would be a daunting task to clean every one of those connectors if they get oxidized. I really like the modular design though. It seems like you could make test jigs and test all the cards on the bench without too much trouble.
Looking forward to the first power-ups! Take care with those capacitor banks, and definitely use a variac and some current limiting if possible. This will be a really fun series.
Looks to me like the typewriter can be used with ANC 2 or ANC 3 so your machine may have mods that convert the NC1 into an ANC 2 or 3... ☺
I think, you do know that you're so into a project that you are comfortable to walk around in bare or stockinged feet over potential wire shards, solder drops, sonic electronic digital ball busters! We're talking electric splinters! Electronic lego bricks!
I love to see that passion!
But, in all seriousness, it does me the power of good to see you head down (down not crashed!), deep in the game, living and breathing this stuff! I can barely wait until the final result! My father was an NCO in the British army "Royal Signals", and he even served in the Falklands, setting up telegraphy and front line communications, even when under fire :-O
Sadly, he had to leave the Army for health, but the deal he made was that he could finish his service whilst working for the British Territorial (aka Army Reserve) Army, still for the Signals corps. I remember he took me to the local HQ and showed me a machine that he promised me was a Computer, although it took up the majority of the space in the back of an Army Lorry!
But even though I never did have a clue about how one might even use such a machine, I've always been impressed with "Big Iron"!
So seeing you with these machines is a real treat!
I know what you've said about emulation, certainly of the drum memory; but I wonder if there is mileage in documenting this system *for* emulation, whilst you're working hard to keep it real?
Sad fact is, one day - there will be no more G-15s in a viable state - but wouldn't it be incredible to have a digital blue print to allow each iota of this machine to be replicated in code, to allow the dreams and the science to continue?
For me, it's all about the inspiration of young (and old) minds! The "Lest we forget" factor ... And I don't mean for this to sound like complete Hyperbole, but I believe there is a lot of value in replicating such old systems, so inspire the new generation.
I also think that there is still, somewhere, a use case for such "robust" equipment that might actually have a genuine and viable purpose for working as much more than just a museum piece!
So, that's my little story - I will try and pump my Dad for more information about that computer and share, but in the mean time - thank you so much, again!
I feel like I'm starting to understand more of the construction of this machine, this will be fun to see progress from static display to computing computer.
The best thing to do is makes some jigs and test all cards before powering up and trying to fault find on the G-15. It will likely be quicker to make jigs and know that all cards are good. The cards are so simple that if any fail, they'll take only minutes to diagnose in a jig, but in the computer it could take you days and days.
Those big round connectors used to called "bendix connectors". As an intern many years ago i was tasked with sourcing some replacements for a seismograph and it took me most of a day to figure out the military specification that they are now known by and order them from Digikey.
Its cool to see them used on an old piece of Bendix equipment.
Now I want to know what the POOP signal actually does! Excellent work BTW!
Probably it should occur in conjunction with the FLUSH and WIPE signals.
Look these immature comments do nothing to advance the content of this channel. Stop being as childish as I am.
Print Out OutPut (at a guess).
@@Pyronimous Order of operations is highly important, here!
💩
Wow! That thing is amazing! Thanks for explaining the different cards. Things are so different today.
I have a bunch of those black-brown-red germanium diodes I got in a box of old electronic parts, and I could never find any specs on them. It's good to finally know where they came from!
22:23 I've seen flippers (flipflop) used as Schmitt triggers for pulse reforming. It looks like that's what they're doing here. (maybe?)
I sometimes use that Armor All spray cleaner for auto interiors to clean equipment. Cleans well and does not have the aggressive ammonia issue in Windex. So it is much kinder to markings on equipment. There are scented versions of this cleaner to avoid.
I`m glad I found your channel. I could not wrap my head around the use of tubes. It still hurts, but not as bad any more. To think that in the 1950`s they figured this out. I can only wonder what they would think about electronics present day.
I've been watching your channel with endless curiosity!
You have to be very careful cleaning tubes - a lot of them are labelled with water-soluble paint than can wipe right off.
Yep. I've had to clean up several series-string tube radios which had TAR all over the tubes. It was necessary to use rubbing alcohol to get them clean, and it usually wiped the print right off. There were a few tube manufacturers, however, which used etched markings for tube type numbers rather than print, the most common of these being GE.
Been there, done that LOL. It makes me wonder if a cheap CNC laser engraver would be a good permanent fix to try.
@@Maxxarcade You take a big risk of cracking the glass with a laser in this case. The process was typically done using chemical etching. Sometimes on a tube which has fragile print I will use automotive clear coat to cover it.
If only a limited set of types is used in the equipment, it may not matter to lose a part number because the appearance of the electrodes will still distinguish what tube it is. To write it back on with a fine tip Sharpie or similar could be useful.
I'm definitely aware!
However, the actual tube number (6197, 6AU6, etc.) is usually put on with a different process that doesn't come off. Many of the tubes in this system have been changed out over the years and the brands are all over the place , though RCA was fairly common. Either way, the silkscreen of the manufacturers logo was so bad on most of the tubes it was flaking off by just blowing on it.
As long as the tube number is readable and the tube works, that's all that really matters!
Fantastic computer. I can't wait to see those valves glowing
"Valves?" Are you from the UK, Australia or New Zealand?
Yes, I am a world wise American.
@@handsomeman-pm9vy Yes. Named after Fleming's thermionic valve. We do use "tube" for CRTs since Braun named his the Braun tube (and they are very tubular, whereas early valves were bulbs. Tube designs came later). But whatever floats your boat is fine. It's just weird to me when someone says "It's full of tubes" and I wonder how many CRTs you could fit in a stereo system, haha.
11:04 and you just keep sneaking that shirt in there... I've started a playlist of these now!!! Ha! Seriously, you're making an already awesome video just WAY over the top for me here!
Yay, more tube computers.😆
I bet that relay box sounds awesome when in use!
This beast is impressive and I reckon it would survive a huge EMP not like our micron thin chips in modern computers. I will watch with interest for further progress.
but the power source be out even if this machine survives. Many smaller computers and phones etc will survive too especially if in protective shielding
E excellent work! Thank you so much for sharing the restoration.
Man, that's a beginning of a great new saga !! Wish you the best of luck. Will be anxiously following the developments. Godspeed !!!!
The king of the nerds is back on task with the Bendix G-15. Yes! This series is going to be epic. Well done and thank you!
I would be tempted to make a test socket for each type of board and use it to test each card separately (maybe using a Raspberry PI, although it may not provide sufficient voltage to check the diodes reverse breakdown voltage)
It's a good idea to make an automatic test and measure system, though an RPi would definitely not have enough voltage output (not to mention there would be a... significant level translation issue, even for inputs). Probably use one or more off the shelf system DC power supplies + custom circuit or one of those HP-IB controlled boxes of relays with a DVM (I forget what they are called) to send the correct input voltages and measure the outputs.
The "Read Amplifier" circuit at 22:08 looks suspiciously like a circuit I designed for a transistor computer I'm developing, so I'll give my 2 cents on what the circuit does.
In my computer I store the microcode in a "core rope"-like memory where one ferrite core represents a single bit and a each wire passing through it is a single word of storage. To read each word I put some current through the corresponding wire, but this induces only a small pulse in the output of the ferrite core, too short and too weak to be used to control the rest of the machine. So I first pass the signal through a simple common emitter amplifier, then the amplified signal is connected to the SET signal of a S-R latch, so that if the bit encoded in the ferrite core is a one, a pulse would be generated and the S-R latch would output a one, and if no pulse is generated the S-R latch would remain to zero. Before pulsing the word wire, i briefly pulse the RESET signal so that all latches start from zero.
The Bendix Read Amplifier looks basically the same, just made out of vacuum tubes; I guess they are using it to sample the output of the drum read heads to "stretch" the pulses long enough to allow the rest of the machine to use the single bits.
WOW! So much work to get this far... very impressive David! Well done, Sir!
"I just had the tunes rollin'..." What is on the Usagi Electric soundtrack?
@5:00 That would heat the house for sure. Great channel. 😊
Check some of those carbon resistors to see if they are still in specification.
Wouldn't be a bad idea to test every component possible/practical.
On that one type of card that only has one row of pins in its connector you can see that the fingers on both sides of the card are still being wiped by the connector pins, so since all the connections are doubled up, they would be very reliable.
Absolutely fascinating! Thank you so much for putting yourself to the trouble of reporting your findings on this not quite yet laptop computer. I'm subscribing, I want to see more of this. Thanks again!
Fascinating journey of discovery! The wrinkled paint finish is known as Raggrinzante, and a soft brush with mild soap and water solution should help get the dirt out those ancient creases.
Maybe by the time that this machine was built Bendix was simply making all of them "alphanumeric ready", so that if the customer desired that feature it was a lot faster/easier to complete that process by wiring up the output pins, instead of having to dive into the backplane with a post-production modification 😊
It would be cool to build a mini version using FETs
where would you get the drum from?
@mohinderkaur6671 modified hard drive maybe?
@@Rorschach1024 the platter from an old drive and it's read write heads could possibly work. I've recorded on various objects such as veggie tins, etc as hobby projects over the years. I've also, in the 80s as a youth, converted 8 track units into data drives. But what I was thinking for this was to use a strip of magnetic tape from a cassette wrapped around a cylinder and the read write heads also from the tape machine. It would be a lot of work tho... and mighty not ever work....
@@mohinderkaur6671 I'm actually thinking the simplest solution would be an endless loop cassette tape originally used with answering machines of the past and a datasette.
@@retroatx It'd have to move pretty swiftly to match the speed of that drum though.
thanks for the amazing quality video. very entertaining, and educational.
It is looking good inside. One thing you could have done with all the boards out is to check all the diodes in case any others are bad. Those cylindrical capacitors(?) seen on the left at 34:12 look nasty and will need to be replaced.
Super exciting, I can hardly wait for this machine to work.
1N270s are still readily available but BAT46 Schottky diodes are also a good replacement from what I understand.
fake 1n34a which are really schottky diodes are also fine