Old man EE here; if you use one scope channel to trigger on clock rising/falling edge (whichever latches the data), then you can use another channel to inspect the latch output at the moment that the latch event happens. This can help diagnose misbehaving latch or flip-flop ICs. You could even use a third channel to inspect the latch input to see that the latch is functioning 100%.
Pretty good stuff if you've got the gear -- a very valid approach. Adrian might have enough traces somewhere. I kind of like how he just bird-dogged it nose-to-the-ground style and hot on the trail with what he had ready to hand. It shows that the most useful (and the most interesting) piece of test equipment is the human brain.
That's where dual-channel oscilloscope is very useful. I am glad I got Zoyi ZT703S - not the best but does do the job well enough. Following the clock occasionally can help you spot some wonky issues with certain chips.
@@horusfalcon I was pretty sure he had at least 2 traces. He has been doing very advanced diagnostics. Time to graduate from the single-trace school of repair, my friend!
fixing stuff without schematics is already high martial arts, i know so many people (including myself) who would probably rather throw it into the bin... be proud of yourself, you did something only very few people can do! - my deep respect, adrian!
You're the 8-bit Columbo. As someone who runs up against older gear from time to time, I find myself recalling your troubleshooting kung-fu quite often. Thank you for the effort you put into these. It goes without saying it brings you joy. Many thanks for bringing us along.
Good history lesson. If I remember correct, several years later when 3dfx wanted to transform from a chip maker to a graphic card brand they bought STB for their factories/production capabilities.
This was such an awesome video Adrian! Love how you combine a history lesson, technical insights into this obscure adapter and how it works, great troubleshooting and a satisfying outcome. This ones a classic!
Interesting repair. I recall that the Olivetti M24 had a 640x400 pixels mode that is similar. It was imported in USA by AT&T and sold as the AT&T PC 6300
Thanks, you inadvertently taught me how to use my Tektronix 1230 Logic Analyzer, it has mostly been collecting dust, it was meant to be used for the vehicle remote starter I designed and programmed over a decade ago.
Adrian's deductive skills never fail to astonish me. He is like a modern day Sherlock Holmes, using logical reasoning extensively not to solve crimes, but to diagnose faults in vintage computer parts! Bravo, Adrian! 👏
A fault in a vintage computer's circuits is a crime... but now we know whodunit! (The _real crime_ is that they don't make 'em like that any more. At least, not mass-produced...) 😄
@@TheRealRaddicalReggie-o9l "Quick, Watson, the needle!" Yeah, most folks blow right by that. Cocaine was not illegal in Victorian England, though, so he was not a criminal.
@@TheRealRaddicalReggie-o9l But you do apply modern standards to a work of fiction written in the Late Victorian and Edwardian Eras about a fictional detective in the Victorian. That is a bit high and outside, man.
Since the "leaking" clock signal would be trailing the actual clock by a tiny amount, at the following chip, D would always be low as the clock latched the input (hence the always-low Q).
Your willingness to stop, think, interpret your readings, and to go back to test your assumptions got you across the finish line! So very well done. I know this must have led you a merry chase for several hours, maybe even days, but you didn't give up. _That's impressive, man!_
This was a Scanner-Danner level case study! Well done. Great example of how using fundamentals and collecting data can solve problems without needing detailed info like schematics. Applies to any kind of system.
I loved these old cards, made from mostly off the shelf TTL devices. Even without a diagram the various sections of the circuit could be easily worked out.
Excellent troubleshooting process! Btw. These old PALs can be read out. Even some of the locked ones could possibly be read out with a brute force testing inputs / outputs results. It makes sense to make backup of the equations from some rare hardware. These are pretty often stolen by scrappers. /-:
I know I watch a lot of these vids and its just simple repairs but every now and then we get one like this where your knowledge shines through and shows us you really know what's going on in there..... Good job.....
Thanks for the great video! Mid 1980s (West-German) Commodore PC-10s OEM'd ATI's Graphics Solution Rev 3 cards (Commodore calls them AGA, Advanced Graphics Adapter) capable of MDA, CGA, Hercules, Plantronics ColorPlus and special ATI 132 Columns all on a TTL monochrome, RGB or composite monitors.
Another great repair. Sure you took off one IC you didn't have to, but sometimes in discovery that just happens. You still did an amazing job in tracing down the three main ICs that were responsible for the output you were seeing even though it didn't match your schematics. I consider that great work. Working with old tech is often an exploratory adventure these days, especially when you no longer have good documentation available.
Well done, FF were allways a pita on xt and at mb. We used to race repairing boards in a catagory (ram, dma, int, etc). Port80 cards were god back then getting you in the right area qhen there was no video. Thanks for sharing.
Hi, Adrian. What an amazing historical and technical video, its always a joy to watch you explain, diagnose and repair pc stuff. Although I'm a Laptop repair technician but I watched the video twice to understand what's going on. Noce and wonderful work, looking forward for feature videos. ❤
A very big issue when not having schematics is when you have hard to see broken traces. You can trace traces all you want but if have a broken trace/solder joint that is hidden under a non-socketed chips it can be difficult to find all chips on a particular line. Good thing it was a damaged chip that you had a compatible replacement for.
Adrian: 1) Remember that inside every one of those ICs are transistors and diodes (at least), so if one fails it can do strange things like shorting together two signals, messing up timing so clocking a latch may not happen as expected. Clean failures of an input or output is much easier to pin down, you did fine in your troubleshooting. 2) I was sometimes able to desolder an IC pin and get it isolated from the PCB pad, feed through, and trace. That way you can avoid desoldering the whole chip, or cutting the pin. 3) Since Radio Shack sold the equivalent card, could you try using one of their apps or utilities, like DeskMate for example, that would exercise the higher resolution modes? Good luck, looking forward to the upcoming video!
Hey, there’s a name I haven’t seen in a while. I used to work on the STB assembly line doing QA and burn-in during college. Worked the 3-11 shift along with some other students. Good times, way back when this type of work was available locally in Texas!
i understood what Adrian was describing, the clock should trigger the latch/flipflop on the rising edge, on the falling edge it should do nothing, the fact that both rising and falling clock passed to the output is therefor an error in the ic
You mentioned Clock MUX. I am a trunking and switching tech at a telecommunications company and work on T1s, T3s, STSs etc..... MUX'ing is an everyday term for me and would love to contribute an explanation if you ever needed. You seem VERY knowledgeable tho so I am equally sure you probably already know what MUX'ing is.
@@Katchi_ Firstly, I don't really care what you believe. Secondly, SONET is NOT dead, I have a 22 year career with the telephone company and we still install a lot of T3s, and large capacity STSs and OC circuits for big banks and other telcos that run through our Province. Also, big banks and callcenters who still want to setup 28 phone lines on just a T1. How many calls do you think you can make on a 1.5 meg SIP circuit? 1... maybe...
This looks like an 8-bit version of the 16 bit Olivetti M24/Olivetti M24SP and Olivetti M28 Video Graphics Card GO317/GO318/GO380 Olivetti also produced another 18 bit "add-on" card that then made it possible for 16 color in 640x400 pixel - This was in 1983
Take a look at the AT&T 6300 who resold the Olivetti rebranded as the AT &T. The 6300 had a 16 bit bus and 640x400 resolution on their special monitors. I had the AT&T 6300 in 1986.
Great video Adrian. Fan and small creator in Canada. If I ever come across a broken old computer at a garage sale I will buy it and send it as a collab, but honestly most people here don't try to sell something already broken so I would have to buy it as working and discover it doesn't work.... we'll see what I can find! 😊😊👍
Nice investigatory fault searching, Adrian. You did well on the logic reasoning through a problem as well. Overall good job and interesting to follow :)
I am always surprised how much information I find here in the videos about my previous devices. I had a noname Hercules card in my first XT clone that handled greyscale CGA graphics output without additional drivers. I remember the manual said something about chauffeur compatibility, but we thought that was just a weird translation error at the time. 🥳
In my country we used the NEC PC-9801 from the first mid to late 80. And loved the high res. Meanwhile cga and ega were half that. Even the last and best PC designed and built here, an XT compatible, in 1986, we made its video system to be able to output the same high res, in 16 colors
I used to have an IBM 3270 PC with the GX graphics extension which took up three ISA slots, similar to the IBM Professional Graphics Controller except using a special high-resolution TTL colour monitor. The display controller had an 8080 on board and supported several graphics modes, up to 480 line text or 720x350 graphics at 16 colours (with special software which I never had). Owing to it having a 720 vertical line resolution (actually 800 on the display itself), the 640x200 CGA monochrome mode only used slightly more than half of the display. I had the standard 5272 colour display but it could be used with a standard TTL monochrome monitor or the larger 5279 professional colour display. IBM even offered a TTL gas plasma display with the system to go along with the plasma display terminals which were popular with their mainframes at the time.
Excellent work, I learned a lot. I know it"s a lot of work, but would there be any way for you to reverse engineer the equations in the PALs? They're gonna fail some time...
35:34 Yes, bit-7 of background is the “blinking” attribute; that’s why text modes only have 8 background colors instead of 16 (sacrifices must be made). Underline is (“normally”?) only available in MDA modes and it triggers when the bottom three bits of the foreground color are exactly 001 (either “blue” or “intense-blue”).
Not that it would have affected finding the eventual issue in this case, but when you found the output at zero volts, an alternative approach may have been to bust out the thermal camera and carefully inspect the chips and trace connecting them. Not sure if that data line is seeing enough activity to generate a heat difference, but any difference in temperature between it and adjacent pins or chips serving the same function would indicate the output is trying to drive hard against a shorted input, while no apparent difference would imply the output isn't being driven at all. These modern/cheap thermal cameras are surprisingly sensitive to small differences in temperature, especially if you can get their field of view narrow enough that they auto-range down to a degree or two overall across the whole image. Even 0.1C is often quite visible, depending on the exact sensor. Alternatively, using your current-limited power supply to inject a bit of voltage into the suspect data line plus an adjacent one (with the board off firstly, although it's also possible with the board on if you're careful to not overload the output pin) going to the same chip might let you differentiate between them, based on how much current is drawn.
Excellent work Adrian! I‘m wondering why you don’t use the retro chip tester not more often? For example to check the 4 ram chips and the flip flop after unsoldering. Just to make sure your assumptions.
STB Systems, Inc., was an American graphics adapter card manufacturer active from 1981 to 1999. Initially a manufacturer of various expansion cards for the Apple II, the company quickly leaned into the graphics accelerator market for IBM PCs and compatibles, owing to the IBM PC's more open architecture. STB went public in 1995 and was once the second-largest global vendor of multimedia computer products. In 1999, the company was acquired by 3dfx Interactive.
My weird STB is a full-length, dual head PCI video card with a loopback cable. Tseng Labs chip, I think. Basically, it was meant for a video editing rig, and the loopback let you overlay your editor’s video output onto your preview/program windows inside the NLE software (D-Vision in my case). Couldn’t do it over the bus back then. My Creative DXR DVD player did the same thing, using chroma key to display the video.
You just showed the manual where all the calls for special video modes are clearly defined. You could easily write some simple assembly code to switch to that mode and fill memory with some patterns to display at least something
Adrian! I have that same video card you had in the test pc first. what card is it? I cant find info on it and im trying to figure out why its not working right?
If the PALs are not registered they are pretty easy to crack and copy. There' s a method to dump a PAL and re-create the equations then compile to a jedec. i've been doing that for ~15 years and it works well as long as the PAL isn't being using in registered mode. This method should be coming to the RCT soon as I passed on that info to the developer so he can improve it. btw when you find bad logic run it through the RCT as that gives a visual representation of the pins and points out the bad pin which can be shown in the vid.
There were several manufacturers of the 6845 CRTC, and they all had very slightly different behaviours. The Amstrad CPC Wiki gives them "type" numbers...
While it doesn't specifically call out the super graphics / graphics plus 2. An older piece of graphics software that I still have a copy of (complete in box, if you're curious) is the Disney Animation Studio. for IBM PC, XT, AT, PS/1, PS/2 and compatibles including Tandy 1000, 2500, and 3000 series. Worked on VGA, MCGA, EGA, Tandy and CGA modes. The other software I recall using (not sure if I have the discs but I can check) would've been older AutoCad v 2. I recall it having specific uses for Hercules Display.
The video production quality looks very good. Are you color correcting each shot? The sound is clear and understandable. Do you still edit your videos all yourself?
I think I might have had that card in my old XT, or something a lot like it. I remember it was attached to a monochrome monitor, and it supported all the MDA features (high-res, blinking, and underlined text) but CGA graphics modes still worked on it. I know it was an STB, and the name Chauffeur sounds awfully familiar.
I remember having an ATI VGA Wonder XL? card back in the 90s that could drive a MDA monitor and generate 16 shades of grey/green/amber. It could emulate CGA and EGA this way. It also implemented hercules graphics mode as well. Very versatile card.
Well - I've never seen an STB Super Res 400 video card before, but I must say, it seems eerily similar to the video card in my first IBM compatible machine. It was back in '84 or '85. I had moved up from a Heath H-89 CP/M system to a Heath H-200 (AT compatible) which I had outfitted with a Sigma Designs Color 400 video card (driving a Princeton Graphics SR-12 monitor). The card was *Very* similar to the STB - same size and form factor, would do 640x400 high res 16 color graphics over a DB-9 connector to the display. Almost perfectly CGA compatible, but this was still just before the EGA boards hit the scene. [I seem to recall the Sigma Designs board also supported an optical mouse on a metal mouse pad with a grid pattern printed on it, and came with a copy of 'PC Paintbrush', which was probably the ancestor of MS-Paint.] Looking at this video, I now wish I could see my Color 400 side by side with the STB card. I wonder if they shared any DNA?
This card with it's lovely high res text mode would do some interesting things to the hacky cga 160x100 graphics mode that is actually modified 80x50 text mode. Obviously 100% cga compatibility would not even be considered for a card like this. Running cga compatibility tester on this card might be interesting to see how cga it actually is. The manual with programming information could allow someone to program a mode test for this card. I'd suspect it would be real easy to modify code that switches to plantronics modes to work with this card instead
Hey Adrian, are you able to copy or make copies of the custom chips that video card has as a back-up? In case those chips went bad and there are ways to reprogram new ones that are pin compatible.
21:06 is it possible that slot8 on that MIC9283 chipset board is setup like an XT slot 8 and that’s why plugging a card into it causes the system not to boot?
The graphics adapter Compaq included in the Compaq Portable functions as a dual CGA/MDA board. I think it can be configured to act solely as a CGA board? I know it's possible to force it into CGA mode for display on a external display using CTRL-ALT-< and CTRL-ALT->.
My lil boy wants retro hardware. If you could send anything to California Eli would be beyond the moon. I have to go to the computer museum to discuss my baby "liveworks".
The video card in original Compaq Portable can also act both as MDA and CGA and switch on the fly, it's normal 8bit ISA card with DB9 and composite outputs altough I've never tried it outside the Compaq.
@@snap_oversteer the Compaq card does that with help if the Compaq BIOS so in a normal PC I think it will act just like a CGA card (but I also haven't tried it)
I was watching this and noticed the huge difference in the use of sockets. Game consoles would be so much easier to repair if they had used some sockets. Even if it was just the CPU and GPU.
Suggestion - since you can not 'replace' those chips, you should take them out, and make copies of them, that way if you ever find/need to be able to replace one, you might be able to reprogram and fill in one of those chips. (The MUX chips) and label the MUX programs as what card they came from.
Growing up, we had a Compaq DeskPro 286 with amber monitor that did cga in 16 shades of amber. Pretty sure it's the same card they used in their luggable that did this.
You’d need to pull the part from the circuit board to use that, and he’s trying to minimise the number of parts desoldered. If they were all socketed then it’s an option.
@@stevetodd7383 , the suspect RAM was socketed, the unsoldered latch was found working, etc. When you unsolder a suspect chip, it's good practice to test it to confirm if it's faulty or not. Same with testing a used part you're going to put back in the circuit. It's just good practice, otherwise you set yourself up for unsuspected errors.
@@No-mq5lw , I've been using a RCT for many years now. Never seen that it did not test a TTL chip correctly. There are/were some issues with some DRAM, but that's been sorted.
@@jannievanzyl4073 The RAM wasn’t suspect, he was removing it to minimise the possible areas to check. You’ll note that he was trying to reduce the RAM to a minimum for the subsequent logic tests. The logic ICs, while in circuit, weren’t testable. Once removed granted he could tested the suspect latch on the chip tester, but settled for lifting the suspect pin and seeing if that solved the issue. Since this was part of a video card then it was running at speeds beyond that of the chip tester (roughly 24MHz from memory, so that approach better matched the test requirements).
![How to build a PC - Step by Step guide for Beginners [EVERYONE CAN DO IT!]](http://i.ytimg.com/vi/4XwGAliimSU/mqdefault.jpg)
![[LIVE] : ONE ลุมพินี 88 | คู่เอก "ป้อมเพชร vs อัสลามจอน"](http://i.ytimg.com/vi/1ZqTVVbMgbo/mqdefault.jpg)
Old man EE here; if you use one scope channel to trigger on clock rising/falling edge (whichever latches the data), then you can use another channel to inspect the latch output at the moment that the latch event happens. This can help diagnose misbehaving latch or flip-flop ICs. You could even use a third channel to inspect the latch input to see that the latch is functioning 100%.
@@thirstyCactus excellent tip, thanks!
Wow...
That's a geek DEEP CUT right there.
Pretty good stuff if you've got the gear -- a very valid approach. Adrian might have enough traces somewhere. I kind of like how he just bird-dogged it nose-to-the-ground style and hot on the trail with what he had ready to hand. It shows that the most useful (and the most interesting) piece of test equipment is the human brain.
That's where dual-channel oscilloscope is very useful. I am glad I got Zoyi ZT703S - not the best but does do the job well enough. Following the clock occasionally can help you spot some wonky issues with certain chips.
@@horusfalcon I was pretty sure he had at least 2 traces. He has been doing very advanced diagnostics. Time to graduate from the single-trace school of repair, my friend!
fixing stuff without schematics is already high martial arts, i know so many people (including myself) who would probably rather throw it into the bin... be proud of yourself, you did something only very few people can do! - my deep respect, adrian!
Schematics should be online somewhere.
You're the 8-bit Columbo.
As someone who runs up against older gear from time to time, I find myself recalling your troubleshooting kung-fu quite often.
Thank you for the effort you put into these. It goes without saying it brings you joy. Many thanks for bringing us along.
I'm a retired technician with over 50 years of electronics T/S&R experience, and I still learn stuff from Adrian. Dude's got _talent._
Good history lesson.
If I remember correct, several years later when 3dfx wanted to transform from a chip maker to a graphic card brand they bought STB for their factories/production capabilities.
Correct.
This was such an awesome video Adrian! Love how you combine a history lesson, technical insights into this obscure adapter and how it works, great troubleshooting and a satisfying outcome. This ones a classic!
Thank you really appreciate it!
Interesting repair. I recall that the Olivetti M24 had a 640x400 pixels mode that is similar. It was imported in USA by AT&T and sold as the AT&T PC 6300
you know it's a great video when you spend over an hour watching troubleshooting on hardware you have absolutely zero interest in. 👍👌
Thanks, you inadvertently taught me how to use my Tektronix 1230 Logic Analyzer, it has mostly been collecting dust, it was meant to be used for the vehicle remote starter I designed and programmed over a decade ago.
Adrian's deductive skills never fail to astonish me. He is like a modern day Sherlock Holmes, using logical reasoning extensively not to solve crimes, but to diagnose faults in vintage computer parts! Bravo, Adrian! 👏
A fault in a vintage computer's circuits is a crime... but now we know whodunit! (The _real crime_ is that they don't make 'em like that any more. At least, not mass-produced...) 😄
Shelrock holmes was addicted to cocaine, you sure you want to compare him to Adrian?
@@TheRealRaddicalReggie-o9l "Quick, Watson, the needle!" Yeah, most folks blow right by that. Cocaine was not illegal in Victorian England, though, so he was not a criminal.
@@horusfalcon sure it wasn’t illegal, but I don’t go around comparing my friends drug addicts
@@TheRealRaddicalReggie-o9l But you do apply modern standards to a work of fiction written in the Late Victorian and Edwardian Eras about a fictional detective in the Victorian. That is a bit high and outside, man.
Since the "leaking" clock signal would be trailing the actual clock by a tiny amount, at the following chip, D would always be low as the clock latched the input (hence the always-low Q).
Your willingness to stop, think, interpret your readings, and to go back to test your assumptions got you across the finish line! So very well done. I know this must have led you a merry chase for several hours, maybe even days, but you didn't give up. _That's impressive, man!_
you got it fixed without the shematics ..Respect .....
This was a Scanner-Danner level case study! Well done. Great example of how using fundamentals and collecting data can solve problems without needing detailed info like schematics. Applies to any kind of system.
I loved these old cards, made from mostly off the shelf TTL devices. Even without a diagram the various sections of the circuit could be easily worked out.
Excellent troubleshooting process! Btw. These old PALs can be read out. Even some of the locked ones could possibly be read out with a brute force testing inputs / outputs results. It makes sense to make backup of the equations from some rare hardware. These are pretty often stolen by scrappers. /-:
Yup, hkz, the person who created the CGA Redux clone, which Adrian used the schematics of, also created the DuPAL for exactly this purpose.
I know I watch a lot of these vids and its just simple repairs but every now and then we get one like this where your knowledge shines through and shows us you really know what's going on in there..... Good job.....
Thanks for the great video! Mid 1980s (West-German) Commodore PC-10s OEM'd ATI's Graphics Solution Rev 3 cards (Commodore calls them AGA, Advanced Graphics Adapter) capable of MDA, CGA, Hercules, Plantronics ColorPlus and special ATI 132 Columns all on a TTL monochrome, RGB or composite monitors.
Loved that brush probe. Never seen one before. Useful.
This is pure magic. I struggeled a bit following your think process. But it all makes sense now.
Another great repair. Sure you took off one IC you didn't have to, but sometimes in discovery that just happens. You still did an amazing job in tracing down the three main ICs that were responsible for the output you were seeing even though it didn't match your schematics. I consider that great work.
Working with old tech is often an exploratory adventure these days, especially when you no longer have good documentation available.
Well done, FF were allways a pita on xt and at mb.
We used to race repairing boards in a catagory (ram, dma, int, etc).
Port80 cards were god back then getting you in the right area qhen there was no video. Thanks for sharing.
Awesome fix
It's like shooting in the dark and getting a coin at 100 yards, you're a wizard ❤
Your Lion King audio cassette is the MVP
Hi, Adrian. What an amazing historical and technical video, its always a joy to watch you explain, diagnose and repair pc stuff. Although I'm a Laptop repair technician but I watched the video twice to understand what's going on. Noce and wonderful work, looking forward for feature videos. ❤
Using the 'period correct' Lion King cassette to prop up the card was a nice touch. Always entertaining to watch your troubleshooting videos.
It's crazy how videos like this keep me on the edge of my seat. Really got interesting around 59:05
Absolutely love these type of videos.
Another job well done. 💪
A very big issue when not having schematics is when you have hard to see broken traces. You can trace traces all you want but if have a broken trace/solder joint that is hidden under a non-socketed chips it can be difficult to find all chips on a particular line. Good thing it was a damaged chip that you had a compatible replacement for.
Adrian:
1) Remember that inside every one of those ICs are transistors and diodes (at least), so if one fails it can do strange things like shorting together two signals, messing up timing so clocking a latch may not happen as expected. Clean failures of an input or output is much easier to pin down, you did fine in your troubleshooting.
2) I was sometimes able to desolder an IC pin and get it isolated from the PCB pad, feed through, and trace. That way you can avoid desoldering the whole chip, or cutting the pin.
3) Since Radio Shack sold the equivalent card, could you try using one of their apps or utilities, like DeskMate for example, that would exercise the higher resolution modes?
Good luck, looking forward to the upcoming video!
Lessons learned and Wisdom delivered! Another great video. Thanks, Adrian!
Serious head scratching diagnosis there. Well done Adrian! Yes - was screaming at the screen when the clock was visible on the output...
Hey, there’s a name I haven’t seen in a while. I used to work on the STB assembly line doing QA and burn-in during college. Worked the 3-11 shift along with some other students. Good times, way back when this type of work was available locally in Texas!
One of your most interesting posts. Well done.
Thanks for the video! Amazing work, kudos!
i understood what Adrian was describing, the clock should trigger the latch/flipflop on the rising edge, on the falling edge it should do nothing, the fact that both rising and falling clock passed to the output is therefor an error in the ic
You mentioned Clock MUX. I am a trunking and switching tech at a telecommunications company and work on T1s, T3s, STSs etc..... MUX'ing is an everyday term for me and would love to contribute an explanation if you ever needed. You seem VERY knowledgeable tho so I am equally sure you probably already know what MUX'ing is.
Find it difficult to believe you are employed in any meaningful capacity. SONET is dead.
@@Katchi_ Firstly, I don't really care what you believe. Secondly, SONET is NOT dead, I have a 22 year career with the telephone company and we still install a lot of T3s, and large capacity STSs and OC circuits for big banks and other telcos that run through our Province. Also, big banks and callcenters who still want to setup 28 phone lines on just a T1. How many calls do you think you can make on a 1.5 meg SIP circuit? 1... maybe...
Excellent job as always Sir!
I used this card to do Autocad back in the late 80' early 90'
Good morning Adrian!
Watching this and the sunset while the pup sleeps next to me.
I think a coffee is in order! This will be a good one!
This looks like an 8-bit version of the 16 bit Olivetti M24/Olivetti M24SP and Olivetti M28 Video Graphics Card GO317/GO318/GO380
Olivetti also produced another 18 bit "add-on" card that then made it possible for 16 color in 640x400 pixel - This was in 1983
Take a look at the AT&T 6300 who resold the Olivetti rebranded as the AT &T. The 6300 had a 16 bit bus and 640x400 resolution on their special monitors. I had the AT&T 6300 in 1986.
Thanks for sharing your knowledge
Great video Adrian. Fan and small creator in Canada. If I ever come across a broken old computer at a garage sale I will buy it and send it as a collab, but honestly most people here don't try to sell something already broken so I would have to buy it as working and discover it doesn't work.... we'll see what I can find! 😊😊👍
You must have been Sherlock Hemlock in another life! Excellent deductive skills.
Nice investigatory fault searching, Adrian. You did well on the logic reasoning through a problem as well. Overall good job and interesting to follow :)
STB= simply the best🥂. I have one, but mine is a VLB card (mine is not super res), and has vga out connector.
I am always surprised how much information I find here in the videos about my previous devices.
I had a noname Hercules card in my first XT clone that handled greyscale CGA graphics output without additional drivers. I remember the manual said something about chauffeur compatibility, but we thought that was just a weird translation error at the time. 🥳
In my country we used the NEC PC-9801 from the first mid to late 80. And loved the high res. Meanwhile cga and ega were half that.
Even the last and best PC designed and built here, an XT compatible, in 1986, we made its video system to be able to output the same high res, in 16 colors
The Olivetti M24 has a similar video subsystem and the technical reference contains the contents of the crtc register translation prom.
interesting video especially as you didnt have schematics. great that you fixed it in one video!
Adrian, I take my hat off to you! That is one extremely neat repair! Permission to engage smug mode (look up Red Dwarf)
Suggest you connect the probe ground to the board you are measuring on,
you wont pickup so much garbage signals.
I used to have an IBM 3270 PC with the GX graphics extension which took up three ISA slots, similar to the IBM Professional Graphics Controller except using a special high-resolution TTL colour monitor. The display controller had an 8080 on board and supported several graphics modes, up to 480 line text or 720x350 graphics at 16 colours (with special software which I never had). Owing to it having a 720 vertical line resolution (actually 800 on the display itself), the 640x200 CGA monochrome mode only used slightly more than half of the display. I had the standard 5272 colour display but it could be used with a standard TTL monochrome monitor or the larger 5279 professional colour display. IBM even offered a TTL gas plasma display with the system to go along with the plasma display terminals which were popular with their mainframes at the time.
Excellent work, I learned a lot. I know it"s a lot of work, but would there be any way for you to reverse engineer the equations in the PALs? They're gonna fail some time...
Wow! Great work!!!
35:34 Yes, bit-7 of background is the “blinking” attribute; that’s why text modes only have 8 background colors instead of 16 (sacrifices must be made). Underline is (“normally”?) only available in MDA modes and it triggers when the bottom three bits of the foreground color are exactly 001 (either “blue” or “intense-blue”).
Fantastic video Adrian
Not that it would have affected finding the eventual issue in this case, but when you found the output at zero volts, an alternative approach may have been to bust out the thermal camera and carefully inspect the chips and trace connecting them. Not sure if that data line is seeing enough activity to generate a heat difference, but any difference in temperature between it and adjacent pins or chips serving the same function would indicate the output is trying to drive hard against a shorted input, while no apparent difference would imply the output isn't being driven at all.
These modern/cheap thermal cameras are surprisingly sensitive to small differences in temperature, especially if you can get their field of view narrow enough that they auto-range down to a degree or two overall across the whole image. Even 0.1C is often quite visible, depending on the exact sensor.
Alternatively, using your current-limited power supply to inject a bit of voltage into the suspect data line plus an adjacent one (with the board off firstly, although it's also possible with the board on if you're careful to not overload the output pin) going to the same chip might let you differentiate between them, based on how much current is drawn.
Excellent work Adrian!
I‘m wondering why you don’t use the retro chip tester not more often? For example to check the 4 ram chips and the flip flop after unsoldering. Just to make sure your assumptions.
ATI Graphics solution 3 can do grey scale on b/w monitors too. (STB=3dfx?)
katakana doesn't really need high resolution, but hiragana would be more readable with it, and it's certainly necessary for kanji.
STB Systems, Inc., was an American graphics adapter card manufacturer active from 1981 to 1999. Initially a manufacturer of various expansion cards for the Apple II, the company quickly leaned into the graphics accelerator market for IBM PCs and compatibles, owing to the IBM PC's more open architecture. STB went public in 1995 and was once the second-largest global vendor of multimedia computer products. In 1999, the company was acquired by 3dfx Interactive.
Awesome, no, really awesome debugging on a vid bd with NO schematic! Simply awesome!
Well done buddy!
To repair boards, it is better to make a riser (adapter on the loop), and you can safely make repairs or settings in a convenient way.
My weird STB is a full-length, dual head PCI video card with a loopback cable. Tseng Labs chip, I think. Basically, it was meant for a video editing rig, and the loopback let you overlay your editor’s video output onto your preview/program windows inside the NLE software (D-Vision in my case). Couldn’t do it over the bus back then. My Creative DXR DVD player did the same thing, using chroma key to display the video.
Another great video, tks Adrian.
LS input logic floats just over 2v which will likely be a high. I found opens that way. CMOS floats at 0v which made opens harder to find on boards.
Nice! never seen a failed chip which leaks one signal to another pin like that (clock).
You just showed the manual where all the calls for special video modes are clearly defined. You could easily write some simple assembly code to switch to that mode and fill memory with some patterns to display at least something
Adrian! I have that same video card you had in the test pc first. what card is it? I cant find info on it and im trying to figure out why its not working right?
I like to work with two scope probes. I put one on the chip enable so the other probe triggers only when it is valid.
If the PALs are not registered they are pretty easy to crack and copy. There' s a method to dump a PAL and re-create the equations then compile to a jedec. i've been doing that for ~15 years and it works well as long as the PAL isn't being using in registered mode. This method should be coming to the RCT soon as I passed on that info to the developer so he can improve it. btw when you find bad logic run it through the RCT as that gives a visual representation of the pins and points out the bad pin which can be shown in the vid.
It’s been impossible to find a single documentary going through the evolution of computer graphics mode, and how the standards can to be.
There were several manufacturers of the 6845 CRTC, and they all had very slightly different behaviours. The Amstrad CPC Wiki gives them "type" numbers...
While it doesn't specifically call out the super graphics / graphics plus 2. An older piece of graphics software that I still have a copy of (complete in box, if you're curious) is the Disney Animation Studio. for IBM PC, XT, AT, PS/1, PS/2 and compatibles including Tandy 1000, 2500, and 3000 series. Worked on VGA, MCGA, EGA, Tandy and CGA modes.
The other software I recall using (not sure if I have the discs but I can check) would've been older AutoCad v 2. I recall it having specific uses for Hercules Display.
Awesome work man!!
The video production quality looks very good. Are you color correcting each shot?
The sound is clear and understandable.
Do you still edit your videos all yourself?
Oh man. Amazing. A shame I am to far to send you a special machine for endless fun.
I think I might have had that card in my old XT, or something a lot like it. I remember it was attached to a monochrome monitor, and it supported all the MDA features (high-res, blinking, and underlined text) but CGA graphics modes still worked on it. I know it was an STB, and the name Chauffeur sounds awfully familiar.
Excellent use of a Lion King tape!
30:19 "Lets see if this card works with these chips removed"
*Screen wipe*
"No."
*Screen wipe*
This had me laughing my ass off.
Very good info thanks
I remember having an ATI VGA Wonder XL? card back in the 90s that could drive a MDA monitor and generate 16 shades of grey/green/amber. It could emulate CGA and EGA this way. It also implemented hercules graphics mode as well. Very versatile card.
There was a cga and a ega wonder card available before that with similar features
There was a cga and a ega wonder card available before that with similar features
Well - I've never seen an STB Super Res 400 video card before, but I must say, it seems eerily similar to the video card in my first IBM compatible machine. It was back in '84 or '85. I had moved up from a Heath H-89 CP/M system to a Heath H-200 (AT compatible) which I had outfitted with a Sigma Designs Color 400 video card (driving a Princeton Graphics SR-12 monitor).
The card was *Very* similar to the STB - same size and form factor, would do 640x400 high res 16 color graphics over a DB-9 connector to the display. Almost perfectly CGA compatible, but this was still just before the EGA boards hit the scene. [I seem to recall the Sigma Designs board also supported an optical mouse on a metal mouse pad with a grid pattern printed on it, and came with a copy of 'PC Paintbrush', which was probably the ancestor of MS-Paint.]
Looking at this video, I now wish I could see my Color 400 side by side with the STB card. I wonder if they shared any DNA?
At my old workplace, which was literally a double garage. We had boxes of those cards (faulty) of course.
This card with it's lovely high res text mode would do some interesting things to the hacky cga 160x100 graphics mode that is actually modified 80x50 text mode. Obviously 100% cga compatibility would not even be considered for a card like this. Running cga compatibility tester on this card might be interesting to see how cga it actually is. The manual with programming information could allow someone to program a mode test for this card. I'd suspect it would be real easy to modify code that switches to plantronics modes to work with this card instead
Another great fix.
Awesome job tacking that one down. I'd call only two chips a win, too. (and Hip waders activated 🙂 )
Hey Adrian, are you able to copy or make copies of the custom chips that video card has as a back-up? In case those chips went bad and there are ways to reprogram new ones that are pin compatible.
21:06 is it possible that slot8 on that MIC9283 chipset board is setup like an XT slot 8 and that’s why plugging a card into it causes the system not to boot?
MS Works for DOS supported a high-res text editor mode. I used it on a Tandy 1000TX which supported the same extended video mode.
The graphics adapter Compaq included in the Compaq Portable functions as a dual CGA/MDA board. I think it can be configured to act solely as a CGA board? I know it's possible to force it into CGA mode for display on a external display using CTRL-ALT-< and CTRL-ALT->.
My lil boy wants retro hardware. If you could send anything to California Eli would be beyond the moon. I have to go to the computer museum to discuss my baby "liveworks".
I remember having one of those STB full length boards back in the mid 1980's. Ours was working properly as long as we had it! 🤣 Bravo, Adrian! 🧐😎🥳
The video card in original Compaq Portable can also act both as MDA and CGA and switch on the fly, it's normal 8bit ISA card with DB9 and composite outputs altough I've never tried it outside the Compaq.
@@snap_oversteer the Compaq card does that with help if the Compaq BIOS so in a normal PC I think it will act just like a CGA card (but I also haven't tried it)
I was watching this and noticed the huge difference in the use of sockets. Game consoles would be so much easier to repair if they had used some sockets. Even if it was just the CPU and GPU.
This "brush probe" is cool! What is it officially called since I can't find it anywhere?
Suggestion - since you can not 'replace' those chips, you should take them out, and make copies of them, that way if you ever find/need to be able to replace one, you might be able to reprogram and fill in one of those chips. (The MUX chips) and label the MUX programs as what card they came from.
Growing up, we had a Compaq DeskPro 286 with amber monitor that did cga in 16 shades of amber. Pretty sure it's the same card they used in their luggable that did this.
@Adrian, why don't you use the Retrochip Tester anymore to test potentially faulty ICs?
You’d need to pull the part from the circuit board to use that, and he’s trying to minimise the number of parts desoldered. If they were all socketed then it’s an option.
IIRC the retro chip tester pro tends to let marginal parts pass at times
@@stevetodd7383 , the suspect RAM was socketed, the unsoldered latch was found working, etc. When you unsolder a suspect chip, it's good practice to test it to confirm if it's faulty or not. Same with testing a used part you're going to put back in the circuit. It's just good practice, otherwise you set yourself up for unsuspected errors.
@@No-mq5lw , I've been using a RCT for many years now. Never seen that it did not test a TTL chip correctly. There are/were some issues with some DRAM, but that's been sorted.
@@jannievanzyl4073 The RAM wasn’t suspect, he was removing it to minimise the possible areas to check. You’ll note that he was trying to reduce the RAM to a minimum for the subsequent logic tests. The logic ICs, while in circuit, weren’t testable. Once removed granted he could tested the suspect latch on the chip tester, but settled for lifting the suspect pin and seeing if that solved the issue. Since this was part of a video card then it was running at speeds beyond that of the chip tester (roughly 24MHz from memory, so that approach better matched the test requirements).