As a 31-year EDS veteran, this was a great trip down Memory Lane. Actually used one of these for quite a while at EDS. They were actually pretty reliable. Thanks for the deep dive on both the history and the machine itself.
Fascinating. I never worked for them but I knew people who did and I was familiar with their prominence in the DC area govcon scene. They had an excellent reputation and the most professional project manager I ever encountered was trained there in the mid 1990s. It seems hilarious to me, though, that they made desktop hardware for a while. It would be the modern day equivalent of someone like Booz Allen introducing a smartphone. Or imagine if Palantir came out with a smartphone - that would be even more laughable, for obvious reasons.
I was with EDS for more than 20 years - I never saw one. I supported GM and all I saw ( in pc market ) was IBM - and 1 compaq. I saw one of these at the Toledo Ham Fest ( once a year flea market - SUPERSIZED - blocks in size ). I wondered if it was my EDS or someone else who used that logo. Now I know.
David, words cannot describe how pleased i am that my uncle chose you to become the caretaker and story teller of Centurion Computer. Your excitement and enthusiasm for both the hardware and the people behind the scenes is truely amazing. Thank you for being the man that you are, and continuing share with the world, the time of Centurion Computer. Looking forward to meeting you someday. 😮
@johnnywarren6617. You may not recall I meet you helping your dad work on your dad's boat at Lake Texaoma in 1975-1976. Your dad [John] & mom [Zoe] along with Bud, where good people.
Honestly, I'm surprised they didn't just go that way anyways. The IBM PC was almost entirely built of off the shelf components anyways (78xx logic and a few PLAs), and the clone market was just starting to spring up. They could have made a killing, and maybe stayed around a bit longer.
@@BlackEpyon Well IBM actually used the original Intel reference design, and built it exactly like that, so cloning it was easy, it was an Intel computer system, and provided you bought the Intel support set with the processor, you had an automatic copyright to build the reference design. The most difficult thing was the IBM BIOS, but IBM published the BIOS code entry and exit specs, so anybody could build a clone provided they had a BIOS that followed that code entry and exit strategy, and not get a lawyer dog on them. Though many simply went and copied the IBM BIOS, and simply went and did a search and replace of IBM with their own 3 letters, and changed a few copyright strings to either spaces or null characters, and those figured they would sell a few thousand before big Blue found out, and they could close shop, move to another address, and open up as a new clone seller again. Plenty of clones like that were around, basically using the same exact board layout and components, just the silkscreen changed every few months to reflect a new company name instead. Same artwork for the copper boards all the time.
My father was a tv repair man. I remember he used to press a service button inside the set to reduce the picture to a single horizontal line. He would then adjust the color controls to get a pure white color balance.
I worked in my father's shop. The way to adjust is to go into service mode for that horizontal line. Turn all colors down until the screen is dark. Turn one color up until you can see it, then slowly turn it back down until the line disappears. Lather, rinse, repeat for the other two colors. That's all it takes to make a good grey-scale.
Specifically, that single line mode disables the vertical deflection (so, don't keep it on in that mode for any length of time; as the entire screen's worth of electron beam(s) is hitting that single line of phosphors). I'm not an expert, this is just from watching old TV repair videos on TH-cam.
Another possibility is that the single white line is used to set the convergence of the R G and B guns. When all three guns are aligned perfectly you would get a crisp white line.
@@nzbsgfan That doesn't work for the vertical deflection; in all the videos I've watched the convergence is setup using the grid test pattern. Again, though, my knowledge is only from watching TH-cam, so.... 🤔
Whoa, interesting! Were they the same style PC as this, or were they a newer type of PC? I'm sure EDS took more than one crack at making a PC clone in its history.
@@UsagiElectric It's been over 30 years, and I don't remember ever looking at the back of the PCs. But the ones that I saw looked exactly like yours in the front. My location at EDS in 1988 was sort of a "backwater", so no doubt that we got cast-offs from other locations.
I started working for EDS in the Netherlands in 1999 and remember seeing one of those exact PCs in the EDS "museum" display case at the Spijkenisse datacenter.
That's it! Those things are so weird looking. Wild that they would print the resistors onto the ceramic substrate and then solder on IC components. What a wild way to build stuff!
@@UsagiElectric Not really weird, this is still common in high reliability applications, and where you want to have precision resistors in an application, and do all the trimming and calibration during manufacture, and have zero setup needed on the line, or during repair, as each module is laser adjusted during production to be identical. Your multimeter uses this for the thick or thin film resistor array that provides the voltage selection, and also it is common in things like terminators, where you can get a precise production of a complex part with simple silk screen printing. Very common in military equipment, and also the method used to build the LNB that your satellite receiver uses, as a reproducible part that is both stable and precise, and which lasts well under extreme temperature swings. You also will find it used in pretty much every sensor in your cars, and in the voltage regulator modules on the alternator, as a way to get a robust reliable package for the electronics.
@@SeanBZA You also find them in a lot of retro kit, the Amiga 500's video output hybrid jumps immediately to mind (basically an R-2R ladder DAC with some driver transistors).
@@SomeMorganSomewhere Yes Burr Brown used this to make ultra precise units, where you would have a 24digit DAC that would not skip codes and be monotonic over the full -55 to +125C military temperature range, as the resistors all track each other precisely. Once factory calibrated, and sealed in a metal case with another thin film to do the reference voltage, they would never drift.
@@highpath4776 So, we can't yet. The Centurion was never really intended to interface with... well, anything. The PC has a pretty good serial transfer program, so we might be able to write some custom software for the Centurion's OpSys that can receive, but we have some big hurdles to overcome to do that. Most notably, the PC will send things in 512 byte chunks, but the Centurion uses 400 bytes per sector, so moving data around gets real confusing. But, that's still something that I want to try to tackle someday!
10 years at EDS. My first job at the company was working for the Tech Products Division in Warren, Michigan. I repaired these as well as IBM PC Machines. There were quite a few of them in the organization. The failure rate on these machines was high and our client, General Motors, hated them. They were happy to replace them with IBM machines. Glad to see that you could get one working. No such thing as a ram tester in those days. You would just populate one bank with 64K of known good chips and add banks until one failed, and then swap chips in the bank one at a time to find the culprit. One PC could take all day to fix, but they did give us overtime!
RAM errors are a PITA to diagnose on these early PCs. One of my XT machines was giving RAM errors, but the RAM chips themselves were perfectly fine. Turned out the issue was a single bad NAND gate that wasn't properly switching the active banks, but I had to reverse-engineer a quarter of the motherboard in order to find where the common link was. Replaced that 74x00, and it's worked perfectly ever since.
We used to have two banks of ram and eprom's as their access times were not fast enough for the 8080 (i think that was the chip but could be wrong). So one set of ram / eprom on even address and the second on the odd address, this effectively halved the speed of the bus. Added an extra step in programming the eprom's as as you had to strip the data for two eproms before you could program a software change in to them. this was done on a micro vax system. :-)
The horizontal line test is reminiscent of the "beam find" function on analog oscilloscopes. That's a freaking amazing monitor! Just for the convenience of not having to open it up to tweak the settings! Also, just an FYI, those POST cards will verify that you've got voltage on the rails, but you're not likely to generate any usable codes on a PC/XT class machine. Still useful to get one though if you're delving into the vintage PC usagi-hole.
When you disconnect one input of an AND gate (like the SN74LS08), it effectively becomes a buffer. by cutting one leg of the AND gate, you’ve transformed it into a buffer.
Does the floating leg always get pulled high if the non-floating leg is pulled high? I thought you had to bridge together the inputs of an AND gate to make it a buffer.
@@rich1051414 In my experience that depends on who made the 74LS08. Midway made a series of arcade game boards where they left some tied high with a resistor and left others open depending on the maker of the chip.
I think you might be thinking about OR, it would be a buffer if you did that. I think cutting the leg means the output is always low, whereas cutting the output would leave a floating line to a possibly more sensitive IC.
@@rich1051414The original 7400 TTL series inputs naturally float high and require substantially more input current to pull them low due to how they are constructed with bipolar transistors. Early on this was an important difference in properties between bipolar and MOS logic even when their logic voltage levels had compatible overlap.
This episode is AMAZING! Not only for for the fact of the Centurion clone, and the history for why it was made, but what impresses me the most comes at the very end where David hooks up the PC as a second data terminal to the Centurion mini and logs in. I ABSOLUTELY LOVE the simplicity and capability of a remote terminal using serial data transfer. There is nothing that beats the fact that you can take a dumb display device and hook it to a smart computing device with three wires (and maybe a little FSK action in between) and do a lot of real work. To this day, I still spend most of my day working in a terminal window. The RS-232 connection has been replaced by SSH, but otherwise things are functionally equivalent. I never think twice about whether the computer that I'm connected to is a VM running on the same laptop as my terminal emulator, or is a computer in another room near me, or is in a data center on the other side of the planet. And if Elon Musk has his way, that computer might someday be located in a colony on Mars, though I suspect that latency might become an issue at that point. 🙂
Thank you so much! I'm fascinated by the RS-232 protocol, it's been around for something like 60 years at this point, and as long as you get the settings right, it just works. It took me less than 30 minutes to the get the EDS PC plugged into the Centurion with a custom cable and functioning. I can totally see why the concept is still being used today!
latency, first step will be a mainframe in space, always in the shadow of Earth/sun to keep it cool, nuclear powered as no sun = no solar power, unless it it sits just on the edge of the shadow and is long/wide enough to have its solar panel in the sunlight.
The cut leg on a TTL chip used to be a fairly common feature - either a retrofit option was added or else it was part of a retroFIX for some error on the motherboard - quite often a RAM configuration or system clock add-in. Yes, the maths coprocessor was an 8087 RGB Monitors Inc. Nice monitors for the time! The horizontal line is for vertical centering of both the display AND convergence on the three electron guns. SCSI/SASI port on the display card - I came across this in a couple of systems - apparently it was for frame capture to an additional device, and for fast frame loading from the same. I *think* I have also seen it used for light pen on a drawing system (that needed the maths co-pro installed). PC/3278 IBM Terminal emulator - I used to see a lot of these when the real terminals became an additional obstruction on the desk and being able to switch over to the mainframe when required was seen as a boon. Hard drive controller - that one should be able to handle MFM drives as well as RLL drives (Jumper selection?) All of my WD controllers back then could handle FM, MFM and RLL drives plus (according the the manuals) several proprietary formats. Bad RAM. One bit is enough to stall POST. How many times have I tried to find the bad chip.
The "flat line" mode in the test switch is likely the same as the "service" switch on a color TV. Used for setting up the CRT bias/screen controls to adjust the greyscale. You adjust the RGB screen controls to get a white/grey line, and the CRT bias to make the line barely visible.
That is so cool. I’ve been watching this series since the beginning and I can’t believe I never discovered that Centurion was headquartered in Richardson. I grew up in Richardson Texas. I went to Richardson high school although it would’ve been purchased by EDS two years before I was born. I look forward to seeing you at UTD and June. Also my à la Mater.
While the 10-1000mb hard drives of yore are largely getting grumpy these days, it's worth giving them a whack to see if they'll behave at least One More Time, with a plan for recovering anything of note. Those noisy whirs, clicks, and clacks are the sound of it thinking.
Do you know the RLL data doubling trick? Buddy of mine did it to an RLL drive back around 1990 or so, but you change the crystal that sets the data clock and you can just about double the capacity of the native drive. One the Seagate's we found that only going to about 60 percent larger was stable. We even threw Super Store at it and doubled the data through compression on top of it, but that really worked the drives. Not going to lie, I hated those 'SeaScrape' drives, always had problems (crash or bearings died) where the Western Digitals and Uniscribe drives seemed to hold up much better and never die. Man does this bring back memories.
@@kleetus92the 238R is just a ST-225, rated to work with an RLL controller. Which basically got you 50% higher recording density on the same drive. I started with a ST-225, on standard MFM and later upgraded to an adaptec RLL controller, and that turned it into a 30 meg drive.
Amazing work! That hard drive - those are the noises it is supposed to make on startup. The PCB doesn't know where the head is so it steps it towards track 0 forcibly even if it's aready there, hence the clicking, then the four softer sounds are sweeps fro track 0 to track 650 and back again, twice. You'll want to boot off an MS-DOS floppy that contains DEBUG. Lauch into debug, type G=C800:5 and hit enter, this wll bring you into the low level format utility on the ROM of the controller card. Once you've low-levelled it you should be able to run FDISK to create a primary partition, then format /s the partition to make it DOS bootable. Good luck!
I know this is a silly thing to notice on a video like this but your appreciation for the person who helped you with the floppy and the troubleshooting, the way you showed genuine appreciation for another man feels rare. Like I don't see men be so genuine often towards other men and it was really cool. Also, great video- I know nothing about old computers like this but it's still my favorite type of entertainment.
I love that you honour the people, places, and efforts of Warrex/Centurion. The pictures really brought it home to me, my late dad had his own small computer business. We all would like to think that we made some difference to the world, but a lot of people’s fine work just slips out from memory - quite quickly! It takes work to keep the memory alive and you’re doing the work. thank you. (Proud patreon member) 🙏
I think one of my favorite things about the Centurion story is that it's really similar to so many other stories out there of small computer businesses. It was the wild west, and people were pouring their heart and soul into this new and exciting industry, and so many of those stories have been lost to the sands of time. But, I think those stories are not only fascinating, but important to remember. They can teach us so much, both from their successes and failures!
I'm still a bit bummed about the ultimate end of Computer Reset. I always planned to visit, and now that I'm finally driving out to Texas, it's gone. It's a regret I'll have to live with.
I managed to get there two times. The second time I went was only about a month or two before they shut down for good, and honestly, by that time, almost everything of note had been picked. The stuff that was left was either fairly common or in such sad shape that it was really repairable. The first time I went, I should have taken a bigger car with me though, they had some really awesome stuff that I was planning on getting on my second trip, but alas, I was too late.
I wonder what the owner was thinking when hoarding, did he knew that the stuff he has is rare, did he sell what he had when he was alive? I am glad that at the end a lot of things got into a hands of people that wanted them
Please save some vacation days and drive out to Edgar Digital in Las Cruces and visit with Roger. Ask him about his Mercury and Apollo collections and enjoy digging through all the vintage electronics, test gear, industrial gear, and computer gear going back decades.
@@Arcadecomponentscom i do wonder when did his hoard get everywhere At first it seams there was some order, but at the end the whole wearhouse and his house was totaly full, (I saw this on lgr video i believe) What happened so that order was lost and all the stairs and all living space was filled I am a hoarder my self but my hoard is in the closet hidden away not to be in the way of my movment My grandfather has it the way usual hoarders has it (stuff everywhwere so crammed u can bearly move)
As someone who worked on the IBM 5150 back in the day, it was so awesome to watch this video today, in 2024. I remember all the chips, and how easy it was to tear apart and customize to any hobbyist desire. Makes me think of WoZ as well, and his protestations when things went non-standard. But I am glad I worked in the PC-Compatible era of tech, it was a wild ride... ALBIET at a much much slowwwweeeer pace than today.
and SASI was SCSI before SCSI was SCSI. Should be able to actually plug a SCSI-1 drive into that display card, if the pin out is known and can be mapped to the drive.
Indeed, most scanners trough the 80's and 90's were either SCSI or Parallel port, so it makes sense for that. Also external enclosures were also most often SCSI or Parallel.
I absolutely love your enthusiasm for repairing and maintaining vintage computer gear! It's a delight watching your joy when something starts working after troubleshooting. Keep up the good work.
When I heard the pc first boot up, the beep, and the floppy drive seeking it literally brought me to tears. I was a SYSOP of The Lair BBS in 1988 and it has been forever since I had last heard the bootup noises. Thank You.
17:19 Ah...you have a Seagate ST-238R in that baby...we're going to get that thing flux imaged for sure! (no matter what!) Forwarding to 24:41 now... Yeah, don't worry...we'll get it! And on to 41:19.... that does NOT sound as bad as you think! We will get it imaged with the Gessweinator! 41:40 It actually finished it's initialize process and is ready! Don't try to get it to work on the PC, we'll get it imaged first at the flux level with the Gessweinator soon!!!
What a great episode! This definitely one of the craziest stories ever! I love it! Between this, the counterfit centurions, and the glue fire, this would make for a great movie!
It physically hurts me whenever I hear about that place, because I didn't have the resources to go there when it was open to the public. And MAN did that look like a good time...
Yeah. I wanted to go really bad before it closed down, but I never found the opportunity. Unfortunate. I hear everything or nearly everything found a home, so thats good though!
RGB Display Corporation has been in the business of producing CRT monitors and LCD panels for gaming and simulation since 1978. The company is a second generation family owned business. ( they also do sheet metalwork and powder coating , so if you wondered where those side panels etc for the Centurion came from.
I worked for EDS for years, and we always had a love/hate relationship with big blue. We really tried to make hitachi mainframes work too but ended up with IBM iron in the end. Fujitsu Sparcs too... yeah, I can 100% believe this was an EDS thing.
I enjoyed my time at Centuion-EDS so much I worked at the EDS Plano campus again two more times before retiring in 2017 from old EDS Plano campus now owned by HP.
This brings back lots of good memories of building and rebuilding literally hundreds of computers in the 1980s (as a job and as hobby). That hard disk sounds perfectly normal by the way. Those Seagate drives all went through this crazy sequence of clicks when you powered them up.
The monitor's construction is really crazy too. Nothing like you'd see from Japan: more like some mil spec high reliability thing. The motherboard on the other hand, stuffed with Japanese ICs a North American manufacturer wouldn't have used at the time, has to be Japanese. I see Mitsubishi, NEC (CPU even!), Toshiba, Hitachi. A couple National Semiconductor UARTs and 8088 support chips but probably because they had no other choice.
You could add a PISCSI, or BlueSCSI II to the video card and connect it to WiFi. Among other things, these devices allow you to load files into a virtual hard drive on the fly and emulate early ethernet connectors.
Awesome video. Reminds me of when I worked at ICL late '80s in the U.K testing/validating 3rd Party data comms products on ICL's first IBM AT compat, the ICL PWS. Did everything from standard dialup on Hayes modems (300baud-19.2Kbaud), IBM LAN (base-band & Token Ring (LU6.2), ICL's own LAN (MicroLan, OSLAN), Terminal Emulators (IBM 3270 SNA, ICL 75xx C03) with dedicated comms cards. When I used to see that mini/mainframe login prompt on the screen after fiddling with myriad IRQ and DMA settings, building comms batch files to log-in etc was such an amazing feeling, so I share your own joy with the EDS PC! Thank you for posting this, it brought back so many happy memories.
Years ago I had an IBM PC/AT with a PC3278 card. I used it to access an IBM System 370 mainframe. The PC was connected to an IBM 3270 display controller via coax cable. The display controller could attach to a local or remote mainframe depending on the model. Later the coax was replaced with TTP. With a series of keystrokes you could switch between PC and main frame mode. When in mainframe mode, the PC was a dump terminal. A Great Leap Forward for the day. I worked for IBM for 37 years.
He picked that up from Adrian's Digital Basement, who picked it up from LGR I think? The app is called Spectroid, it's been on my phones for like a decade.
@@UsagiElectric I did a double take when I saw the colour palette! The monitor label shows "32 colors can be selected from 4096" at TTL levels. It says 3 lines; presumably 8 color RGB, 4 lines; RGBI CGA, and 5 lines; for the 32 color mode. So it's not EGA which uses 6 lines and a fixed 64 colour palette. What video card can drive the 5 line 32 color mode? And does the monitor have built in functions for selecting your 32 colors from 4096 possibilities? I've never seen anything like it.
@@erikj3390 yeah. Mindboggling. Suppose there was something like a programmable palette table in the monitor? In principle you could display Amiga-like graphics on there. But without a clock line, it seems difficult to imaging how you'd avoid really weird fringing effects.
An absobloodylutely fascinating story! So cool to see you bring the PC clone up, even if the HDD is still not there. I don't get the 74LS08 shenanigans either. It's a quad AND gate and cutting one lead off would kill one gate, or even worse, make the output state undefined as the input is not tied to any potential. Maybe someone started doing some dodgy-bodgy workaround but didn't Keri on through the task, haha. It looks like the lead was originally soldered in, which is pretty logical, but has been discombobulated later on.
@@mycosys ah, OK - it has diode input, it'll go low only when one or more of the inputs are pulled down. Leaving them floating will force high on the output.
Absolutely loving this community, what a find. Kinda interesting they made people work on something that was just a bargaining chip, bit of a waste of potential. :) Absolutely lovely work!
That monitor looks a lot like how Ceronix used to make their monitors. They used the same seperate vertical pcbs to modularise the different sections of the monitor.
8X305 !!! Yay! One of the weirdest processor chips ever made, although described by the manufacturers as a microcontroller. They were 8 bit processors capable of 5 MIPs, sometime used in disk controllers, and other devices where high speed (for the time) bit-wise and logical operations were required. It has a *very* limited instruction set, enough to make any seasoned programmer pull their hair out. And I absolutely love the built-in test generator in the monitor. That's a nice find.
It’s always nice to see someone so enthusiastic about retro tech. :) Are you going to make another video showcasing what the clone can do as a standalone PC?
I love your videos because the subject are always interesting and you put your soul on all of them, your joy when something old came alive is wonderful. That warehouse where you find this pc is a paradise for me if I was in USA I would go there to try buying some old PC. I would like to have an Ibm pc again. Congratulatios for the new item for Ceturyon family.
The line test is used to set grayscale. You turn the color pots until the line JUST appears for each color. That sets cutoff for the CRT. What an interesting monitor.... A broadcast monitor has always had those features, but a computer monitor? That's pretty unique.
@@Runco990I've never seen a monitor chassis with externally accessible analogue controls for the RGB drive either. I always thought that was exclusively adjusted via pots on the neckboard until digital OSDs came along.
Hey...thanks for the walk down memory lane. Been a LONG time since I played with systems like that. And Oh...the number of dead systems like that which I heaped into the trash bins over time. I have been working computers since I started my business in '83. Am 73...and still doing work. In fact...just did a CMOS flash on a 11900 system this morning...smiles. Anyway...nice show.
I suspect you are right. Not much else makes sense. From memory, 74LS08 is a quad two input AND gate. Pin 2 is lifted which is input 1a. Pin 3 is input 1b and pin 4 is the gate's output.
Great Video! RLL vs MFM is the type of encoding. Back in the day, it was more common to run into MFM drives (the ST225 being the most famous). I've never gotten a clear answer on why, but I suppose it had to do with price points and availability/production costs.
RLL drives needed a better quality of magnetic platters to be reliable. I had both MFM and RLL drives back in the day. I had software that tested hard drives at a low level (spinrite and disk technician) and reportedly you could take an MFM drive, connect it an RLL controller and get 50% more storage after testing and validation to detect bad blocks
🌠The suspenseful bootup is a _feature_ 🌠 Great video. I especially enjoy how jazzed you get when things work! Congrats on your success with this _very cool_ PC. I love the job you gave it, too. 👍
Hi dave!, thank you for all these Wonderfull times you spend with us, The 3270 board comes with a special software loaded as "resident program" with a special key combination it swaps between pc session and mainframe session. and there is also a special api called via a software interruption allowing tou to interract with the mainframe session. i've used it in 1987 in our micro computer section to read accountant data from the mainframe faking a user session in the back of the proud "computer specialists"..
That was truly an epic adventure and I admire your dedication and drive to do what you did to make it go! I worked with real 5150s in high school (yes, as a student!) and I agree - the design was insanely accessible.
You are fortunate to have working floppy drives. The TEC drives have history with capacitors leaking and corroding traces on the floppy drive's PCB. There is also one capacitor on the motor PCB which is a hassle to get to for replacement.
those sideways cards are a simple way to make single-layer (double sided) pcb function like multi-layer pcb. and you can fix those easily because they are single layer pcb. its not progress if maintenance becomes harder.
The RGB Display Corporation of California are still round making signage! The style looks really familiar, i did some research but cant find who they OEMed, did find them in the Apple //c expansion manual though
At first I thought of RGB Spectrum (that make video wall equipment) but it sounds more likely that it might be that RGB Display company in Grass Valley ☺
74LS08 Is the low-power Schottky version of a quad 2-input AND gate. Pin 2 is the second input of the first AND gate. Since it's cut, pin 3 (the first gate output) will never fire. You have to be careful with the LS variants of TTL chips because they are extremely sensitive to static discharge, much more so than standard TTL logic.
I have no idea how many hours I've spent staring at Procomm Plus connected to a BBS but it's a big number. That old Seagate ain't happy at all, the ringing sound makes me think there's a little sticky spot (waterspot from condensation probably) on a platter somewhere. Of course since it's dead you've got nothing to lose by trying to clean it up, a little DIRO water might just take care of it. But it does spin and it does seek so chances are if you can get the crud off the platter it might just be savable.
Same here. I was a little late to the BBS world, as I didn't get a modem until the family business bought a second PC in 1987... and in 1991 all my Internet access was through my college's VAX, so Procomm Plus was constantly running... I even used it under OS/2 in order to test the dial in lines for the ISP I helped start. Never liked the competitor software... and I liked the ability to extend Procomm... like when you had to manually setup bleeding edge software like Zmodem before it became commonplace.
Thank you Marco , LGR , and everyone who got into Computer Reset . I have over 1000 Systems and Documentation , Books , Software etc. . and I hope it and finds a good Home :) QC
I started my career working for IBM (1980-1991) working on mainframes (3090, 4381, AS 400, etc). Later on I worked for a company called Attachmate. They made 3278 cards for the IBM type PC's that would attach PC's to those mainframes. The architectures were so different that they used a protocol called "screen scraping". Basically the card would scrape the data of of what was being displayed on the mainframe monitor and just do some conversion for the PC. It was slow and inefficient and kind of an early stepping stone in the PC to Mainframe talk to each other days. Those cards were expensive at the time but were very successful for Attachmate in the early years. It was a real garage startup (just like Apple earlier).
74LS08 is a quad, 2 input AND gate the cut leg is the B input for the first gate. Also great video, bit of history and a look over the machine plus a crazy backstory? What more could one ask for. Also while you might not hear the highpitched whine of the CRT you'r muc captures well enough for me to hear (or at least i think i do).
That startup clicking on the hard drive is how ST-251/ST-277R drives sound :) Nice drives which have an early auto-park mechanism. Fingers crossed you can get the data off, but it's quite common for drives of that era to have "lost" their low level format and need to have a fresh factory format done (usually a firmware routine on the controller itself). Recently I used the MFM Emulator project (over at pdp8online) to backup a Quantum 40MB out of a Wang PC - might be worth trying it on that drive before attempting a LLF.
The line test mode on the monitor is to support aligning the 3 colors. There are adjustments on the yoke to trim the alignment of each of the color beams as far as pulling up/down on the left and up/down on the right. The goal is to get a flat horizontal white line with no divergence of the color guns from edge to edge. I think there are other adjustments there as well but it's been since the early 80's since I've worked on color CRTs and memory has gotten more than a little fuzzy.
Lots of different date codes on those 74xx series ics. I saw years 82, 84, 85, 86, 88 and 89. It might have been built as a prototype but seems to have been well used. Great vid !
I feel so old watching this... The ST-238R was a way to get 30MB storage out of an ST225's 20MB for not much more $$$. I remember being told (unsure if true) that the 238 was simply a 225 that was re-labelled if it passed a specific test. My first clone build used a 238R and I don't recall having any issues with it - I was very pleased with it. Yes, they are NOISY. The STB 3278 card is indeed an IBM 3270 emulator. If you were an IBM shop, these things (and their clones) were plentiful as we couldn't wait to get rid of the humoungus 3270s. IIRC (again, it's been years), they connected via the BNC connector on the back thru RG-59(?) coax to the mainframe via whatever IBM called routers/switches back in the day ( we just called them controllers). Eventually, the coax was replaced by twisted pair with an RJ connector on it, and then thru a balun (RJ in/BNC out) to the emulator/terminal. That Power Supply itself is something else! Who made it?
That's exactly true about the ST-225/ST-238. Although a large number of ST-225s with few or zero defects didn't get the ST-238 label simply because there weren't that many people that could justify the more expensive RLL controller and there were some suspicions that the higher capacity was somehow more failure prone.
RG62 if I remember correctly. This was one of the reasons why some businesses migrated to Arcnet, they already had the cabling for the IBM terminal network.
Actually I think I heard this story before, fits perfectly with who Ross Perot was. But this channel brings back sooo many memories of when I used to work on more hardware than I do now (nowadays it's all about the cloud - boy I miss the old days...)
If the HDD could not be made working - please investigate the option to apply some magnetic paint to reveal the data optically - and then photograph the platters. With 32MB in 4 heads, thats like 8 MB / 32 Mbit per side - quite likely reconstructable (if you make a polar panorama scanner).
I've spent the last 5 months restoring an IBM 5150 and it's just about done. With regards to the hard disk though, it had an ST513 which would only boot if it was the correct way "up", otherwise it would make the exact same knocking sound as yours. I'm guessing you've probably already tried it in various orientations but just in case, it's worth a try.
The line is a CRT OK test, it bypasses most of the circuits to fire a line to check if the CRT is running and roughly aligned. Tube and early solid state TVs often have a switch that does that marked "test" or "service".
@@UsagiElectric Well hard to say, but you might get in, and still see remnants of the plant there, though the office space likely still closely resembles the original, aside from having had a dozen remodels over time. Go to the power inlet and you probably will find Centurion branded equipment panel plates there though, or at least on the city engineering plans. Those drawings and plans are never updated name wise, because they get referred to by so many other plans, so keeping them in sync is near impossible. By me there are still substations where the plans refer to the original builder, even though the building has changed hands over time, the substation and power rooms, plus telecoms rooms, all still have layers of old names around. Worked in a building which was called Olivetti House, but they were long gone, and the substation in it had feeds from House of Delegates (now dept of education), and off to Forsdicks Motors (now a part of VAG dealerships, as Forsdicks moved out a half century before), and another feed marked as Receiver of revenue, which is still there, though they now occupy 4 buildings instead of 1 as they rebuilt over time from a tiny extra to being the main office. Another town the substation says Roberts Heights, with later feed additions saying Verwoerdburg ext 2, and latest now saying Centurion ext 4, and they still are the only block of substations that supply power at a nominal 250VAC, as opposed to the later standard of 240VAC, and the new harmonised standard of 230VAC. Incandescent lamps, stoves and heaters have a very short and very bright life there still, peak voltage of 263VAC and minimum of 226VAC, barely within spec, and almost always running on the high side around 245VAC. But hard to change because of the LV and MV interconnects you would need to do all at the same time to fix it, and change all the transformer taps same power interruption, so they simply set new transformers to the same voltage settings. Only new developments, with a single 11kV 3 phase supply, got 230VAC mains fed to them, the old layout with parallel transformers in a ring stay the same.
The question is ... and perhaps Mr Usagi can do a first here. Eat some Pilgrims' Pride chicken nuggets in front of a Centurion terminal. I mean, the most complete computer restoration action would be to remove food particles made in the same factory as the computer. How many times in history has that been done? How many times has it knowingly been done? HOW MANY TIMES WAS IT RECORDED, AND POSTED ON TH-cam?
Centurion had 1780 Jay Ell built-out for our company from the shell building to our spec in 1978-1979. After Centurion then EDS-Centurion the Centurion line was sold to ZTRON and we moved the temo site in Richardson before moving to a new building in Allen, TX on West Bethany Dr. After Centurion move out of Jay Ell the building became EDS Tech Products Div ( building stuff for GM ). Later Jay Ell became a Data Center for Ross Perot's new company "Perot Systems". After years of being a data center for Perot System the data center was sold to Pilgrims Pride as a Data Center ( maybe to count chicken parts ).
As a 31-year EDS veteran, this was a great trip down Memory Lane. Actually used one of these for quite a while at EDS. They were actually pretty reliable. Thanks for the deep dive on both the history and the machine itself.
Fascinating. I never worked for them but I knew people who did and I was familiar with their prominence in the DC area govcon scene. They had an excellent reputation and the most professional project manager I ever encountered was trained there in the mid 1990s. It seems hilarious to me, though, that they made desktop hardware for a while. It would be the modern day equivalent of someone like Booz Allen introducing a smartphone. Or imagine if Palantir came out with a smartphone - that would be even more laughable, for obvious reasons.
I was with EDS for more than 20 years - I never saw one. I supported GM and all I saw ( in pc market ) was IBM - and 1 compaq. I saw one of these at the Toledo Ham Fest ( once a year flea market - SUPERSIZED - blocks in size ). I wondered if it was my EDS or someone else who used that logo. Now I know.
David, words cannot describe how pleased i am that my uncle chose you to become the caretaker and story teller of Centurion Computer. Your excitement and enthusiasm for both the hardware and the people behind the scenes is truely amazing. Thank you for being the man that you are, and continuing share with the world, the time of Centurion Computer.
Looking forward to meeting you someday. 😮
@johnnywarren6617. You may not recall I meet you helping your dad work on your dad's boat at Lake Texaoma in 1975-1976. Your dad [John] & mom [Zoe] along with Bud, where good people.
such a funny flex to build an entire PC to just get a better price
Right? It shows how easy it was to make a PC clone.
It's the most EDS thing that EDS could have done, lol.
Honestly, I'm surprised they didn't just go that way anyways. The IBM PC was almost entirely built of off the shelf components anyways (78xx logic and a few PLAs), and the clone market was just starting to spring up. They could have made a killing, and maybe stayed around a bit longer.
@@BlackEpyon It does look like the Compaq PC before Compaq built the desktop one.
@@BlackEpyon Well IBM actually used the original Intel reference design, and built it exactly like that, so cloning it was easy, it was an Intel computer system, and provided you bought the Intel support set with the processor, you had an automatic copyright to build the reference design. The most difficult thing was the IBM BIOS, but IBM published the BIOS code entry and exit specs, so anybody could build a clone provided they had a BIOS that followed that code entry and exit strategy, and not get a lawyer dog on them.
Though many simply went and copied the IBM BIOS, and simply went and did a search and replace of IBM with their own 3 letters, and changed a few copyright strings to either spaces or null characters, and those figured they would sell a few thousand before big Blue found out, and they could close shop, move to another address, and open up as a new clone seller again. Plenty of clones like that were around, basically using the same exact board layout and components, just the silkscreen changed every few months to reflect a new company name instead. Same artwork for the copper boards all the time.
My father was a tv repair man. I remember he used to press a service button inside the set to reduce the picture to a single horizontal line. He would then adjust the color controls to get a pure white color balance.
That makes sense! I never thought about the single line being used for adjusting the white balance!
I worked in my father's shop. The way to adjust is to go into service mode for that horizontal line. Turn all colors down until the screen is dark. Turn one color up until you can see it, then slowly turn it back down until the line disappears. Lather, rinse, repeat for the other two colors. That's all it takes to make a good grey-scale.
Specifically, that single line mode disables the vertical deflection (so, don't keep it on in that mode for any length of time; as the entire screen's worth of electron beam(s) is hitting that single line of phosphors).
I'm not an expert, this is just from watching old TV repair videos on TH-cam.
Another possibility is that the single white line is used to set the convergence of the R G and B guns. When all three guns are aligned perfectly you would get a crisp white line.
@@nzbsgfan That doesn't work for the vertical deflection; in all the videos I've watched the convergence is setup using the grid test pattern. Again, though, my knowledge is only from watching TH-cam, so.... 🤔
I worked for EDS back in 1987-1997. At my first position, they had several EDS PCs. I never knew they were more than just a cheap clone.
Whoa, interesting!
Were they the same style PC as this, or were they a newer type of PC? I'm sure EDS took more than one crack at making a PC clone in its history.
@@UsagiElectric It's been over 30 years, and I don't remember ever looking at the back of the PCs. But the ones that I saw looked exactly like yours in the front. My location at EDS in 1988 was sort of a "backwater", so no doubt that we got cast-offs from other locations.
I started working for EDS in the Netherlands in 1999 and remember seeing one of those exact PCs in the EDS "museum" display case at the Spijkenisse datacenter.
12:19 Those modules are called thick film hybrid modules. The resistors are the black strips printed onto the ceramic substrate.
That's it! Those things are so weird looking. Wild that they would print the resistors onto the ceramic substrate and then solder on IC components. What a wild way to build stuff!
@@UsagiElectric Not really weird, this is still common in high reliability applications, and where you want to have precision resistors in an application, and do all the trimming and calibration during manufacture, and have zero setup needed on the line, or during repair, as each module is laser adjusted during production to be identical. Your multimeter uses this for the thick or thin film resistor array that provides the voltage selection, and also it is common in things like terminators, where you can get a precise production of a complex part with simple silk screen printing.
Very common in military equipment, and also the method used to build the LNB that your satellite receiver uses, as a reproducible part that is both stable and precise, and which lasts well under extreme temperature swings. You also will find it used in pretty much every sensor in your cars, and in the voltage regulator modules on the alternator, as a way to get a robust reliable package for the electronics.
@@SeanBZA You also find them in a lot of retro kit, the Amiga 500's video output hybrid jumps immediately to mind (basically an R-2R ladder DAC with some driver transistors).
@@SomeMorganSomewhere Yes Burr Brown used this to make ultra precise units, where you would have a 24digit DAC that would not skip codes and be monotonic over the full -55 to +125C military temperature range, as the resistors all track each other precisely. Once factory calibrated, and sealed in a metal case with another thin film to do the reference voltage, they would never drift.
Usagi upload: This is a good day
Usagi uploads a video about an IBM clone: HOLY SHIT WHAT I NEED TO WATCH THAT
Not just any IBM clone, a Centurion IBM clone!
@@UsagiElectric how can you copy infor from the clone copies to a centurion , for execution, or storage, using the comms programm ?
@@UsagiElectric I heard that in the posh English TV advert voice for MnS. Not just any food, this is MnS Food.
@@highpath4776 So, we can't yet. The Centurion was never really intended to interface with... well, anything. The PC has a pretty good serial transfer program, so we might be able to write some custom software for the Centurion's OpSys that can receive, but we have some big hurdles to overcome to do that. Most notably, the PC will send things in 512 byte chunks, but the Centurion uses 400 bytes per sector, so moving data around gets real confusing.
But, that's still something that I want to try to tackle someday!
This will be the fifth video on PC clones I've watched this week.
10 years at EDS. My first job at the company was working for the Tech Products Division in Warren, Michigan. I repaired these as well as IBM PC Machines. There were quite a few of them in the organization. The failure rate on these machines was high and our client, General Motors, hated them. They were happy to replace them with IBM machines. Glad to see that you could get one working. No such thing as a ram tester in those days. You would just populate one bank with 64K of known good chips and add banks until one failed, and then swap chips in the bank one at a time to find the culprit. One PC could take all day to fix, but they did give us overtime!
RAM errors are a PITA to diagnose on these early PCs. One of my XT machines was giving RAM errors, but the RAM chips themselves were perfectly fine. Turned out the issue was a single bad NAND gate that wasn't properly switching the active banks, but I had to reverse-engineer a quarter of the motherboard in order to find where the common link was. Replaced that 74x00, and it's worked perfectly ever since.
74ls20 ?
@@rallyscoot No, a 74S00. The '00 is a quad 2-input NAND, and the '20 is a dual 4-input NAND.
We used to have two banks of ram and eprom's as their access times were not fast enough for the 8080 (i think that was the chip but could be wrong). So one set of ram / eprom on even address and the second on the odd address, this effectively halved the speed of the bus. Added an extra step in programming the eprom's as as you had to strip the data for two eproms before you could program a software change in to them. this was done on a micro vax system. :-)
The horizontal line test is reminiscent of the "beam find" function on analog oscilloscopes. That's a freaking amazing monitor! Just for the convenience of not having to open it up to tweak the settings!
Also, just an FYI, those POST cards will verify that you've got voltage on the rails, but you're not likely to generate any usable codes on a PC/XT class machine. Still useful to get one though if you're delving into the vintage PC usagi-hole.
I find your channel very nostalgic (I am 75) and I love your enthusiasm! Keep going please!
When you disconnect one input of an AND gate (like the SN74LS08), it effectively becomes a buffer. by cutting one leg of the AND gate, you’ve transformed it into a buffer.
Does the floating leg always get pulled high if the non-floating leg is pulled high? I thought you had to bridge together the inputs of an AND gate to make it a buffer.
@@rich1051414 In my experience that depends on who made the 74LS08. Midway made a series of arcade game boards where they left some tied high with a resistor and left others open depending on the maker of the chip.
I think you might be thinking about OR, it would be a buffer if you did that. I think cutting the leg means the output is always low, whereas cutting the output would leave a floating line to a possibly more sensitive IC.
@@rich1051414The original 7400 TTL series inputs naturally float high and require substantially more input current to pull them low due to how they are constructed with bipolar transistors. Early on this was an important difference in properties between bipolar and MOS logic even when their logic voltage levels had compatible overlap.
@@teknifix That depends on whether or not the cut leg floats high. If not, then you're wrong. If it does, then you're right.
The SASI adapter is on the graphics card to allow you to connect a SCSI/SASI Graphics Digitiser tablet ;)
Weren’t there scsi scanners?
@@ppokorny99 yeah, they were a thing too, either would make sense for access to the vmem but i would guess its intended for a digitiser
@@mycosys early HP Scanjets used a scsi interface
Ah nice I came here to say this too! People forget that SCSI was designed as a generic device bus, not just for disks.
At this point, scsi portion of the graphics card was used for scanners but eventually supported first generation CD-ROM readers.
This episode is AMAZING! Not only for for the fact of the Centurion clone, and the history for why it was made, but what impresses me the most comes at the very end where David hooks up the PC as a second data terminal to the Centurion mini and logs in. I ABSOLUTELY LOVE the simplicity and capability of a remote terminal using serial data transfer. There is nothing that beats the fact that you can take a dumb display device and hook it to a smart computing device with three wires (and maybe a little FSK action in between) and do a lot of real work. To this day, I still spend most of my day working in a terminal window. The RS-232 connection has been replaced by SSH, but otherwise things are functionally equivalent. I never think twice about whether the computer that I'm connected to is a VM running on the same laptop as my terminal emulator, or is a computer in another room near me, or is in a data center on the other side of the planet. And if Elon Musk has his way, that computer might someday be located in a colony on Mars, though I suspect that latency might become an issue at that point. 🙂
Thank you so much!
I'm fascinated by the RS-232 protocol, it's been around for something like 60 years at this point, and as long as you get the settings right, it just works. It took me less than 30 minutes to the get the EDS PC plugged into the Centurion with a custom cable and functioning. I can totally see why the concept is still being used today!
If you go to mars, you’ll have to bring your data with you….an AI database! Get a daily digest from earth, send a daily digest back.
latency, first step will be a mainframe in space, always in the shadow of Earth/sun to keep it cool, nuclear powered as no sun = no solar power,
unless it it sits just on the edge of the shadow and is long/wide enough to have its solar panel in the sunlight.
The cut leg on a TTL chip used to be a fairly common feature - either a retrofit option was added or else it was part of a retroFIX for some error on the motherboard - quite often a RAM configuration or system clock add-in. Yes, the maths coprocessor was an 8087
RGB Monitors Inc. Nice monitors for the time! The horizontal line is for vertical centering of both the display AND convergence on the three electron guns.
SCSI/SASI port on the display card - I came across this in a couple of systems - apparently it was for frame capture to an additional device, and for fast frame loading from the same. I *think* I have also seen it used for light pen on a drawing system (that needed the maths co-pro installed).
PC/3278 IBM Terminal emulator - I used to see a lot of these when the real terminals became an additional obstruction on the desk and being able to switch over to the mainframe when required was seen as a boon.
Hard drive controller - that one should be able to handle MFM drives as well as RLL drives (Jumper selection?) All of my WD controllers back then could handle FM, MFM and RLL drives plus (according the the manuals) several proprietary formats.
Bad RAM. One bit is enough to stall POST. How many times have I tried to find the bad chip.
The "flat line" mode in the test switch is likely the same as the "service" switch on a color TV. Used for setting up the CRT bias/screen controls to adjust the greyscale. You adjust the RGB screen controls to get a white/grey line, and the CRT bias to make the line barely visible.
That is so cool. I’ve been watching this series since the beginning and I can’t believe I never discovered that Centurion was headquartered in Richardson. I grew up in Richardson Texas. I went to Richardson high school although it would’ve been purchased by EDS two years before I was born. I look forward to seeing you at UTD and June. Also my à la Mater.
That sound of the HDD is very healthy, it's what ST-238R is supposed to sound like ;)
Or the 225. I ran a BBS on four 225s for a few years, man they were loud.
While the 10-1000mb hard drives of yore are largely getting grumpy these days, it's worth giving them a whack to see if they'll behave at least One More Time, with a plan for recovering anything of note. Those noisy whirs, clicks, and clacks are the sound of it thinking.
Do you know the RLL data doubling trick? Buddy of mine did it to an RLL drive back around 1990 or so, but you change the crystal that sets the data clock and you can just about double the capacity of the native drive. One the Seagate's we found that only going to about 60 percent larger was stable. We even threw Super Store at it and doubled the data through compression on top of it, but that really worked the drives.
Not going to lie, I hated those 'SeaScrape' drives, always had problems (crash or bearings died) where the Western Digitals and Uniscribe drives seemed to hold up much better and never die. Man does this bring back memories.
@@kleetus92the 238R is just a ST-225, rated to work with an RLL controller. Which basically got you 50% higher recording density on the same drive. I started with a ST-225, on standard MFM and later upgraded to an adaptec RLL controller, and that turned it into a 30 meg drive.
Amazing work! That hard drive - those are the noises it is supposed to make on startup. The PCB doesn't know where the head is so it steps it towards track 0 forcibly even if it's aready there, hence the clicking, then the four softer sounds are sweeps fro track 0 to track 650 and back again, twice. You'll want to boot off an MS-DOS floppy that contains DEBUG. Lauch into debug, type G=C800:5 and hit enter, this wll bring you into the low level format utility on the ROM of the controller card. Once you've low-levelled it you should be able to run FDISK to create a primary partition, then format /s the partition to make it DOS bootable. Good luck!
tend to agree - as at the end you heard it do a full seek without hitting the mechanical stop.
It should be backed up before wiped over so carelessly.
I know this is a silly thing to notice on a video like this but your appreciation for the person who helped you with the floppy and the troubleshooting, the way you showed genuine appreciation for another man feels rare. Like I don't see men be so genuine often towards other men and it was really cool. Also, great video- I know nothing about old computers like this but it's still my favorite type of entertainment.
I love that you honour the people, places, and efforts of Warrex/Centurion. The pictures really brought it home to me, my late dad had his own small computer business.
We all would like to think that we made some difference to the world, but a lot of people’s fine work just slips out from memory - quite quickly! It takes work to keep the memory alive and you’re doing the work. thank you. (Proud patreon member) 🙏
I think one of my favorite things about the Centurion story is that it's really similar to so many other stories out there of small computer businesses. It was the wild west, and people were pouring their heart and soul into this new and exciting industry, and so many of those stories have been lost to the sands of time. But, I think those stories are not only fascinating, but important to remember. They can teach us so much, both from their successes and failures!
I'm still a bit bummed about the ultimate end of Computer Reset. I always planned to visit, and now that I'm finally driving out to Texas, it's gone. It's a regret I'll have to live with.
I managed to get there two times. The second time I went was only about a month or two before they shut down for good, and honestly, by that time, almost everything of note had been picked. The stuff that was left was either fairly common or in such sad shape that it was really repairable.
The first time I went, I should have taken a bigger car with me though, they had some really awesome stuff that I was planning on getting on my second trip, but alas, I was too late.
I wonder what the owner was thinking when hoarding, did he knew that the stuff he has is rare, did he sell what he had when he was alive?
I am glad that at the end a lot of things got into a hands of people that wanted them
@@Veso266 He knew what he had and loved to chat about the PCjr.
Please save some vacation days and drive out to Edgar Digital in Las Cruces and visit with Roger. Ask him about his Mercury and Apollo collections and enjoy digging through all the vintage electronics, test gear, industrial gear, and computer gear going back decades.
@@Arcadecomponentscom i do wonder when did his hoard get everywhere
At first it seams there was some order, but at the end the whole wearhouse and his house was totaly full, (I saw this on lgr video i believe)
What happened so that order was lost and all the stairs and all living space was filled
I am a hoarder my self but my hoard is in the closet hidden away not to be in the way of my movment
My grandfather has it the way usual hoarders has it (stuff everywhwere so crammed u can bearly move)
As someone who worked on the IBM 5150 back in the day, it was so awesome to watch this video today, in 2024. I remember all the chips, and how easy it was to tear apart and customize to any hobbyist desire. Makes me think of WoZ as well, and his protestations when things went non-standard. But I am glad I worked in the PC-Compatible era of tech, it was a wild ride... ALBIET at a much much slowwwweeeer pace than today.
22:40 SCSI was long the interface of choice for scanners.
and SASI was SCSI before SCSI was SCSI. Should be able to actually plug a SCSI-1 drive into that display card, if the pin out is known and can be mapped to the drive.
Indeed, most scanners trough the 80's and 90's were either SCSI or Parallel port, so it makes sense for that.
Also external enclosures were also most often SCSI or Parallel.
Yes, and "SCSI" was originally supposed to be pronounced "sexy" but someone said "scuzzy" and it was funny, so it stuck :-)
I still use scsi scanners every day.
I absolutely love your enthusiasm for repairing and maintaining vintage computer gear! It's a delight watching your joy when something starts working after troubleshooting. Keep up the good work.
Didn’t feel long at all. Congrats and thanks for sharing
Thank you so much, glad I managed to get the pacing right!
I love to see your sucess stories of bringing old computers back to life…..
Thank you so much!
When I heard the pc first boot up, the beep, and the floppy drive seeking it literally brought me to tears.
I was a SYSOP of The Lair BBS in 1988 and it has been forever since I had last heard the bootup noises.
Thank You.
You owned one of these EDS machines??
17:19 Ah...you have a Seagate ST-238R in that baby...we're going to get that thing flux imaged for sure! (no matter what!) Forwarding to 24:41 now... Yeah, don't worry...we'll get it! And on to 41:19.... that does NOT sound as bad as you think! We will get it imaged with the Gessweinator! 41:40 It actually finished it's initialize process and is ready! Don't try to get it to work on the PC, we'll get it imaged first at the flux level with the Gessweinator soon!!!
Please, only one comment instead of 10 million
Excellent!
Looking forward to maybe getting some data off the old girl!
I love the sound of those old Seagate drives!
I had the ST-238 in my clone, a thin WYSE PC that I wish I never gave away...
@@Cmdrbzrd that was only one comment, sequentially on the HDD, nice and easy to follow
@@volvo09 I think a WYSE ? terminal passed through my hands once in the UK too
What a great episode! This definitely one of the craziest stories ever! I love it! Between this, the counterfit centurions, and the glue fire, this would make for a great movie!
15:00 - That horizontal line is for setting the Red Green and Blue gain controls, achieving a white line.
RIP Computer Reset
It physically hurts me whenever I hear about that place, because I didn't have the resources to go there when it was open to the public. And MAN did that look like a good time...
It was a fun 3-year project to empty it and get all that gear out to new homes. Volunteers get together from time to time when we can. :)
@@sterlingphoenix Same.
@@sterlingphoenixSame here.
Yeah. I wanted to go really bad before it closed down, but I never found the opportunity. Unfortunate. I hear everything or nearly everything found a home, so thats good though!
RGB Display Corporation has been in the business of producing CRT monitors and LCD panels for gaming and simulation since 1978. The company is a second generation family owned business. ( they also do sheet metalwork and powder coating , so if you wondered where those side panels etc for the Centurion came from.
Does that mean David might be able to get a technical description of how to create the bona fide Centurion blue color out of RGB's dusty old files?
Thanks, I did wonder where that brushed stainless plug panel came from! Didn't look cheap, or necessary by any means!
I worked for EDS for years, and we always had a love/hate relationship with big blue. We really tried to make hitachi mainframes work too but ended up with IBM iron in the end. Fujitsu Sparcs too... yeah, I can 100% believe this was an EDS thing.
I enjoyed my time at Centuion-EDS so much I worked at the EDS Plano campus again two more times before retiring in 2017 from old EDS Plano campus now owned by HP.
Ah, the Fujitsu SPARCs.
This brings back lots of good memories of building and rebuilding literally hundreds of computers in the 1980s (as a job and as hobby).
That hard disk sounds perfectly normal by the way. Those Seagate drives all went through this crazy sequence of clicks when you powered them up.
At 12:04 it does say "made by hitachi japan" on the sticker, but I guess that's just the CRT tube?
that logo looked like RCA, which had Japanese partners.
The monitor's construction is really crazy too. Nothing like you'd see from Japan: more like some mil spec high reliability thing. The motherboard on the other hand, stuffed with Japanese ICs a North American manufacturer wouldn't have used at the time, has to be Japanese. I see Mitsubishi, NEC (CPU even!), Toshiba, Hitachi. A couple National Semiconductor UARTs and 8088 support chips but probably because they had no other choice.
You could add a PISCSI, or BlueSCSI II to the video card and connect it to WiFi. Among other things, these devices allow you to load files into a virtual hard drive on the fly and emulate early ethernet connectors.
Awesome video. Reminds me of when I worked at ICL late '80s in the U.K testing/validating 3rd Party data comms products on ICL's first IBM AT compat, the ICL PWS. Did everything from standard dialup on Hayes modems (300baud-19.2Kbaud), IBM LAN (base-band & Token Ring (LU6.2), ICL's own LAN (MicroLan, OSLAN), Terminal Emulators (IBM 3270 SNA, ICL 75xx C03) with dedicated comms cards. When I used to see that mini/mainframe login prompt on the screen after fiddling with myriad IRQ and DMA settings, building comms batch files to log-in etc was such an amazing feeling, so I share your own joy with the EDS PC! Thank you for posting this, it brought back so many happy memories.
Years ago I had an IBM PC/AT with a PC3278 card. I used it to access an IBM System 370 mainframe. The PC was connected to an IBM 3270 display controller via coax cable. The display controller could attach to a local or remote mainframe depending on the model. Later the coax was replaced with TTP. With a series of keystrokes you could switch between PC and main frame mode. When in mainframe mode, the PC was a dump terminal. A Great Leap Forward for the day. I worked for IBM for 37 years.
that spectrum analyzer on the phone was genious, I wish more people did this
He picked that up from Adrian's Digital Basement, who picked it up from LGR I think? The app is called Spectroid, it's been on my phones for like a decade.
man, that built-in test pattern. wild!
It was so cool to see that in person, I had no clue it was there!
@@UsagiElectric I did a double take when I saw the colour palette! The monitor label shows "32 colors can be selected from 4096" at TTL levels. It says 3 lines; presumably 8 color RGB, 4 lines; RGBI CGA, and 5 lines; for the 32 color mode. So it's not EGA which uses 6 lines and a fixed 64 colour palette. What video card can drive the 5 line 32 color mode? And does the monitor have built in functions for selecting your 32 colors from 4096 possibilities? I've never seen anything like it.
@@erikj3390 yeah. Mindboggling. Suppose there was something like a programmable palette table in the monitor? In principle you could display Amiga-like graphics on there. But without a clock line, it seems difficult to imaging how you'd avoid really weird fringing effects.
@@erikj3390it's basically the Amiga palette. 12 bit RGB... even the 32 out of 4096 matches with the Amiga. :)
An absobloodylutely fascinating story! So cool to see you bring the PC clone up, even if the HDD is still not there.
I don't get the 74LS08 shenanigans either. It's a quad AND gate and cutting one lead off would kill one gate, or even worse, make the output state undefined as the input is not tied to any potential. Maybe someone started doing some dodgy-bodgy workaround but didn't Keri on through the task, haha. It looks like the lead was originally soldered in, which is pretty logical, but has been discombobulated later on.
Looking at the SN74LS08 schematic, that input will float high turning it into a buffer im pretty sure. Its a ground triggered schottky input
@@mycosys ah, OK - it has diode input, it'll go low only when one or more of the inputs are pulled down. Leaving them floating will force high on the output.
@@KeritechElectronics Thats what it looked like, yeah.
7:39 "there's a link somewhere around here"....standing by for the YT InfoCard on that...
Updated the description with the link!
(Thanks for reminding me, I totally forgot, lol)
So cool! Enjoyed watching that. Your excitement for the old technology is contagious!
My dad was an early EDS manager and he remembers this computer. As well as the fact that Ross eventually bought everyone IBM PCs.
Absolutely loving this community, what a find. Kinda interesting they made people work on something that was just a bargaining chip, bit of a waste of potential. :) Absolutely lovely work!
34:21 Ah, the shirt!
Gotta represent!
Fantastic work getting it up and running at all!
great video and what a cool back story on the particular unit
That monitor looks a lot like how Ceronix used to make their monitors. They used the same seperate vertical pcbs to modularise the different sections of the monitor.
8X305 !!! Yay! One of the weirdest processor chips ever made, although described by the manufacturers as a microcontroller. They were 8 bit processors capable of 5 MIPs, sometime used in disk controllers, and other devices where high speed (for the time) bit-wise and logical operations were required. It has a *very* limited instruction set, enough to make any seasoned programmer pull their hair out.
And I absolutely love the built-in test generator in the monitor. That's a nice find.
Amazing work there, Dave!
It’s always nice to see someone so enthusiastic about retro tech. :) Are you going to make another video showcasing what the clone can do as a standalone PC?
I expect it to be the same as any other XT clone, would be nice to run Checkit, speedtest etc. It should run any XT program/games etc.
Excellent! I love this sort of video, in depth troubleshooting and a great result!
I love your videos because the subject are always interesting and you put your soul on all of them, your joy when something old came alive is wonderful. That warehouse where you find this pc is a paradise for me if I was in USA I would go there to try buying some old PC. I would like to have an Ibm pc again. Congratulatios for the new item for Ceturyon family.
Awesome setup. I am completely envious. Great job!
Awesome job! Congratulations on the effort to rescue the history and bring it back to life. Sensational!
13:50 I wonder if the line test shows the signals coming into the monitor
The line test is used to set grayscale. You turn the color pots until the line JUST appears for each color. That sets cutoff for the CRT. What an interesting monitor.... A broadcast monitor has always had those features, but a computer monitor? That's pretty unique.
@@Runco990 Huh, thanks for the info!
@@jakesteampson7043 You're welcome!
@@Runco990I've never seen a monitor chassis with externally accessible analogue controls for the RGB drive either. I always thought that was exclusively adjusted via pots on the neckboard until digital OSDs came along.
Fascinating story especially about the relationship of the eds and ibm. Thanks for sharing 😊
Hey...thanks for the walk down memory lane. Been a LONG time since I played with systems like that. And Oh...the number of dead systems like that which I heaped into the trash bins over time. I have been working computers since I started my business in '83. Am 73...and still doing work. In fact...just did a CMOS flash on a 11900 system this morning...smiles. Anyway...nice show.
Cool! And the connection to Reset is wonderful.
Watching a rare PC compatible come back to life is always fun, but I think the best part of the video was the kitties at the end. XD
I would guess there was a error in the board design and they lifted the leg to fix it.
Ah, that would make sense! I didn't even think about it being an official bodge, I was thinking someone along the line somewhere had modified it.
It might be like a dog lifting its leg to mark its territory .... ?!
I suspect you are right. Not much else makes sense.
From memory, 74LS08 is a quad two input AND gate.
Pin 2 is lifted which is input 1a. Pin 3 is input 1b and pin 4 is the gate's output.
Great Video! RLL vs MFM is the type of encoding. Back in the day, it was more common to run into MFM drives (the ST225 being the most famous). I've never gotten a clear answer on why, but I suppose it had to do with price points and availability/production costs.
RLL drives needed a better quality of magnetic platters to be reliable. I had both MFM and RLL drives back in the day.
I had software that tested hard drives at a low level (spinrite and disk technician) and reportedly you could take an MFM drive, connect it an RLL controller and get 50% more storage after testing and validation to detect bad blocks
@ppokorny99 Yes. Also RLL generally required better performance in the data separator circuit on the disc controller board.
🌠The suspenseful bootup is a _feature_ 🌠
Great video. I especially enjoy how jazzed you get when things work! Congrats on your success with this _very cool_ PC. I love the job you gave it, too. 👍
Your enthusiasm is always really a joy 🤗👍
"absolutely filthy inside"
Shows just a little bit of dust
man thats a cool setup, congrats on getting it going!
I love everything about the video! Great work!
About HDD, you have to provide drive geometry to the controller to make it work.
I can't remember how it was done, probably dip-switches.
Hi dave!, thank you for all these Wonderfull times you spend with us,
The 3270 board comes with a special software loaded as "resident program"
with a special key combination it swaps between pc session and mainframe session.
and there is also a special api called via a software interruption allowing tou to interract with the mainframe session. i've used it in 1987 in our micro computer section to read accountant data from the mainframe faking a user session in the back of the proud "computer specialists"..
That was truly an epic adventure and I admire your dedication and drive to do what you did to make it go! I worked with real 5150s in high school (yes, as a student!) and I agree - the design was insanely accessible.
You are fortunate to have working floppy drives. The TEC drives have history with capacitors leaking and corroding traces on the floppy drive's PCB. There is also one capacitor on the motor PCB which is a hassle to get to for replacement.
That brings so many enjoyable memories from the past. Thank you!
those sideways cards are a simple way to make single-layer (double sided) pcb function like multi-layer pcb. and you can fix those easily because they are single layer pcb. its not progress if maintenance becomes harder.
Makes it easy to reverse-engineer too, which is important if you can't find a schematic anywhere.
The RGB Display Corporation of California are still round making signage!
The style looks really familiar, i did some research but cant find who they OEMed, did find them in the Apple //c expansion manual though
At first I thought of RGB Spectrum (that make video wall equipment) but it sounds more likely that it might be that RGB Display company in Grass Valley ☺
74LS08 Is the low-power Schottky version of a quad 2-input AND gate. Pin 2 is the second input of the first AND gate. Since it's cut, pin 3 (the first gate output) will never fire.
You have to be careful with the LS variants of TTL chips because they are extremely sensitive to static discharge, much more so than standard TTL logic.
Won't the input "float" high, turning the gate into a buffer? Or do the LS series gates either float low or become indeterminate if left floating?
@@melkiorwiseman5234Unconnected LS inputs float high.
I have no idea how many hours I've spent staring at Procomm Plus connected to a BBS but it's a big number.
That old Seagate ain't happy at all, the ringing sound makes me think there's a little sticky spot (waterspot from condensation probably) on a platter somewhere. Of course since it's dead you've got nothing to lose by trying to clean it up, a little DIRO water might just take care of it. But it does spin and it does seek so chances are if you can get the crud off the platter it might just be savable.
Same here. I was a little late to the BBS world, as I didn't get a modem until the family business bought a second PC in 1987... and in 1991 all my Internet access was through my college's VAX, so Procomm Plus was constantly running... I even used it under OS/2 in order to test the dial in lines for the ISP I helped start.
Never liked the competitor software... and I liked the ability to extend Procomm... like when you had to manually setup bleeding edge software like Zmodem before it became commonplace.
Well done Marco! Brilliant move man. Absolutely brilliant.
Oh, god! This video looks like a best detective/drama movie. Thanks a lot!
Thank you Marco , LGR , and everyone who got into Computer Reset . I have over 1000 Systems and Documentation , Books , Software etc. . and I hope it and finds a good Home :) QC
What an incredible and bizarre story behind this thing. Developing a whole PC basically as a flex! Wow. Glad this one survived!
I started my career working for IBM (1980-1991) working on mainframes (3090, 4381, AS 400, etc). Later on I worked for a company called Attachmate. They made 3278 cards for the IBM type PC's that would attach PC's to those mainframes. The architectures were so different that they used a protocol called "screen scraping". Basically the card would scrape the data of of what was being displayed on the mainframe monitor and just do some conversion for the PC. It was slow and inefficient and kind of an early stepping stone in the PC to Mainframe talk to each other days. Those cards were expensive at the time but were very successful for Attachmate in the early years. It was a real garage startup (just like Apple earlier).
74LS08 is a quad, 2 input AND gate the cut leg is the B input for the first gate.
Also great video, bit of history and a look over the machine plus a crazy backstory? What more could one ask for. Also while you might not hear the highpitched whine of the CRT you'r muc captures well enough for me to hear (or at least i think i do).
That startup clicking on the hard drive is how ST-251/ST-277R drives sound :) Nice drives which have an early auto-park mechanism. Fingers crossed you can get the data off, but it's quite common for drives of that era to have "lost" their low level format and need to have a fresh factory format done (usually a firmware routine on the controller itself). Recently I used the MFM Emulator project (over at pdp8online) to backup a Quantum 40MB out of a Wang PC - might be worth trying it on that drive before attempting a LLF.
I worked for EDS! I'm glad I stuck with this to reach the point of realizing that this is an EDS mini computer! Clearly a prototype!
Don't mind me, just spinning up the ol' Seagate...
_jackhammer noises intensify_
The line test mode on the monitor is to support aligning the 3 colors. There are adjustments on the yoke to trim the alignment of each of the color beams as far as pulling up/down on the left and up/down on the right. The goal is to get a flat horizontal white line with no divergence of the color guns from edge to edge. I think there are other adjustments there as well but it's been since the early 80's since I've worked on color CRTs and memory has gotten more than a little fuzzy.
I looked at the SN74LS08 schematic - i think its inputs will float high, so cutting that one leg of the and gate will turn it into a buffer.
Lots of different date codes on those 74xx series ics. I saw years 82, 84, 85, 86, 88 and 89. It might have been built as a prototype but seems to have been well used. Great vid !
14:30, wow i can hear the hum change volume when you fiddle with those dials
I feel so old watching this...
The ST-238R was a way to get 30MB storage out of an ST225's 20MB for not much more $$$. I remember being told (unsure if true) that the 238 was simply a 225 that was re-labelled if it passed a specific test. My first clone build used a 238R and I don't recall having any issues with it - I was very pleased with it. Yes, they are NOISY.
The STB 3278 card is indeed an IBM 3270 emulator. If you were an IBM shop, these things (and their clones) were plentiful as we couldn't wait to get rid of the humoungus 3270s. IIRC (again, it's been years), they connected via the BNC connector on the back thru RG-59(?) coax to the mainframe via whatever IBM called routers/switches back in the day ( we just called them controllers). Eventually, the coax was replaced by twisted pair with an RJ connector on it, and then thru a balun (RJ in/BNC out) to the emulator/terminal.
That Power Supply itself is something else! Who made it?
I think they used RG58 like 10base2 ethernet, because of the 50 ohm impedance vs 75 of tv coax...
@@freeculture you may be right, because I don’t recall the coax being as thick as RG 59 was
That's exactly true about the ST-225/ST-238. Although a large number of ST-225s with few or zero defects didn't get the ST-238 label simply because there weren't that many people that could justify the more expensive RLL controller and there were some suspicions that the higher capacity was somehow more failure prone.
RG62 if I remember correctly. This was one of the reasons why some businesses migrated to Arcnet, they already had the cabling for the IBM terminal network.
Actually I think I heard this story before, fits perfectly with who Ross Perot was. But this channel brings back sooo many memories of when I used to work on more hardware than I do now (nowadays it's all about the cloud - boy I miss the old days...)
If the HDD could not be made working - please investigate the option to apply some magnetic paint to reveal the data optically - and then photograph the platters. With 32MB in 4 heads, thats like 8 MB / 32 Mbit per side - quite likely reconstructable (if you make a polar panorama scanner).
I've spent the last 5 months restoring an IBM 5150 and it's just about done. With regards to the hard disk though, it had an ST513 which would only boot if it was the correct way "up", otherwise it would make the exact same knocking sound as yours. I'm guessing you've probably already tried it in various orientations but just in case, it's worth a try.
The line is a CRT OK test, it bypasses most of the circuits to fire a line to check if the CRT is running and roughly aligned. Tube and early solid state TVs often have a switch that does that marked "test" or "service".
SCSI CDROM if you can find the drivers, or a tape backup. Like the Media Vision sound boards have. But that might be too old.
As ever, your joy at something working again after decades of not working has me smiling and remembering exactly why we do this bonkers hobby.
Putting Keyboard keycaps in a Tied at the end pillow case and then running through clothe washer/Drier has worked wonderfully for me.
Lol 1780 Jay Ell is now a Pilgrims Pride chicken nugget plant. That says something about American production.
I wonder if they'll let me in to wander around, lol.
Haha well chicken nuggets are important too
@@UsagiElectric Well hard to say, but you might get in, and still see remnants of the plant there, though the office space likely still closely resembles the original, aside from having had a dozen remodels over time. Go to the power inlet and you probably will find Centurion branded equipment panel plates there though, or at least on the city engineering plans. Those drawings and plans are never updated name wise, because they get referred to by so many other plans, so keeping them in sync is near impossible.
By me there are still substations where the plans refer to the original builder, even though the building has changed hands over time, the substation and power rooms, plus telecoms rooms, all still have layers of old names around. Worked in a building which was called Olivetti House, but they were long gone, and the substation in it had feeds from House of Delegates (now dept of education), and off to Forsdicks Motors (now a part of VAG dealerships, as Forsdicks moved out a half century before), and another feed marked as Receiver of revenue, which is still there, though they now occupy 4 buildings instead of 1 as they rebuilt over time from a tiny extra to being the main office.
Another town the substation says Roberts Heights, with later feed additions saying Verwoerdburg ext 2, and latest now saying Centurion ext 4, and they still are the only block of substations that supply power at a nominal 250VAC, as opposed to the later standard of 240VAC, and the new harmonised standard of 230VAC. Incandescent lamps, stoves and heaters have a very short and very bright life there still, peak voltage of 263VAC and minimum of 226VAC, barely within spec, and almost always running on the high side around 245VAC. But hard to change because of the LV and MV interconnects you would need to do all at the same time to fix it, and change all the transformer taps same power interruption, so they simply set new transformers to the same voltage settings. Only new developments, with a single 11kV 3 phase supply, got 230VAC mains fed to them, the old layout with parallel transformers in a ring stay the same.
The question is ... and perhaps Mr Usagi can do a first here. Eat some Pilgrims' Pride chicken nuggets in front of a Centurion terminal. I mean, the most complete computer restoration action would be to remove food particles made in the same factory as the computer. How many times in history has that been done? How many times has it knowingly been done? HOW MANY TIMES WAS IT RECORDED, AND POSTED ON TH-cam?
Centurion had 1780 Jay Ell built-out for our company from the shell building to our spec in 1978-1979. After Centurion then EDS-Centurion the Centurion line was sold to ZTRON and we moved the temo site in Richardson before moving to a new building in Allen, TX on West Bethany Dr. After Centurion move out of Jay Ell the building became EDS Tech Products Div ( building stuff for GM ). Later Jay Ell became a Data Center for Ross Perot's new company "Perot Systems". After years of being a data center for Perot System the data center was sold to Pilgrims Pride as a Data Center ( maybe to count chicken parts ).