Hoi Jan. The rainbow you are getting is probably because the composite signal has some severe rippling after the sync pulse, where there should be just a flat front porch. Your oscilloscope images from Part 1 show just that, if you look closely. That rippling is probably being interpreted by the TV as a color burst (see: en.wikipedia.org/wiki/Colorburst ), and that's why you get colors instead of the B&W signal the TRS80 should be producing. Together with the severe rippling in the clock network, that suggests parasitics in the board, which could be caused by e.g. a bad chip hanging from the clock network. I haven't looked at the TRS80's schematics, but it should be possible to back-trace the composite signal to figure out where that rippling/"color burst" is coming from. Once you solve the rainbow issue, what will remain is a screwed-up but stable pattern on the screen, which suggests (video) memory issues. That's probably just some bad chips, which should be easy to diagnose. Don't give up, mate! This is just an 8-bit computer, there aren't many places where the problem could be hiding! (Addendum: the rippling in the clock network, according to your oscilloscope, happens on the low level of the clock. That suggests an undesired parasitic oscillator somewhere other than the CPU (though not 100% certain). That said, Z80s are still made, which are drop-in replacements for the Z80A in there. Here is the part number: Z84C0006PEC. They sell for only EUR 6 at Reichelt.)
@@JanBeta Well Jan this is difficult one huh... actually sounds like a job for Mr. Black ... Mr. Adrian Black (@Adrian's Digital Basement) :-D Are you sure that the power supply delivers and sustains... "at all times" .. all the voltages and amps necessary... That the screen light up .. like normal startup .. and then dim down like that.... It may be because one or more voltages drop at boot .. when the big capacitor load up por example.... and the info in memory gets all screwed up for that momentary lack of voltage .. Check the stability of the voltages with the oscilloscope during the start up When it is on .. have you tried bending the board a little... or tab her... you know.. looking for false contacts in solder joins... or in broken traces...
Jan, as a novice, I find that watching along with your debugging and reading the comments and suggestions posted by your expert viewers I learn a lot that I would never learn watching other computer repair videos. For this reason, I appreciate very much your content and the discussion generated by your viewers. Thanks to both viewers and Jan for sharing your knowledge with us novices. It is very much appreciated. Particularly found the comments by spacedock8 and the byte attic helpful Cheers!
Looking forward to Part 3 Jan! Your channel is a real ‘find’ … as a relative newcomer and newbie I can’t tell you how encouraging + helpful your videos are. Don’t give up on the TRS-80 👍
Easy to sit back and criticize, don't give up. It's 2020, and there has been lots of crazy things happening. Without failure you never learn what you are really capable of doing. I have faith keep at it. Those complaining cannot say they have never failed. I love your innovation and your videos.
No worries that this repair bled over into October. It's an interesting repair. I do have a Model I, was working last time I had it running but that was many years ago. Afraid to use the included power supply as it's so old, perhaps I'll make one like you did. Very much like your channel, greetings from Norther Ohio in the U.S.
Thanks John! More detective work needed for this one. ;) I think you could refurbish the original power supply to an extend by replacing the capacitor and the diodes, should be standard parts as far as I know. I just had to make my own supply because the originals are quite impossible to find over here.
I have one of these and am the original owner. I replaced the Z80 back in Feb 2000. Everything works but the tape recorder circuit (as of 20 Feb 2021). If you want to compare notes, I can make some scope screen shots if that would help. Love the channel!
Interesting video...I have worked on TRS-80 Model I computer back in the 80's doing all sorts of hardware mods. When the unit is first powered up, the ROM BIOS (well it wasn't call that then, but you get the idea) will initialize the video memory and clear the screen. While you do appear to have a video output issue, the display is also not clearing, so the system is not booting up. I would, as some other commenters have suggested, verify power supply ripple, as well as the clock chain as that might be preventing things from starting. A couple of the mods I had made include powering the board from a switching power supply, which improved system stability tremendously (basically fed it +/- 5V and 12V), and took a discrete video output (H/V sync + signal), so that you can feed a standard VGA monitor (I think it was CGA or EGA at the time), by picking off the TTL level signals for H and V sync from Z5, and video output off of Z30 pin 1.
I did some further troubleshooting in the meantime and there's definitely an issue with the clock signal getting messed up somewhere in the chain. I am going to get back to it eventually (the whole process was a bit disheartening to be honest so it takes some time to get back on track with the TRS repair...) Thanks for your input! :)
I feel your pain, brother. I am having similar issues with my CoCo 2 with weird scrambled video output, the video signal is super noisy/ripply. Good luck with this -- I am sure you will prevail.
Hello Jan, It's been a year and I was wondering if you were planning on resurrecting this project for SepTandy this year? I'd love to see it wrapped up. Thank you for all the great videos!
Would love to see a part 3. I do have some suggestions. You had found one of your address or data lines was being held down. This could be the issue. Check the warmth of your chips and see if they are getting warm. I'm pretty sure that video signal in the trash 80 required the processor to be running correctly. Keep at it. I'd like to see you fix it!
Yeah, I am currently at the point where I think the processor is not running correctly, too. I think the problem is the processor clock being too dirty to be interpreted right. I am working on it on and off, there is definitely going to be another video as soon as I have fixed the issue. :)
74HC04 is a recommended replacement of 74C04. HC series are CMOS chips pin compatible with 74 series, but with logic levels compatible with CMOS 4000 series chips (symmetrical) - same as 74C series, but faster. The HCT series are modern CMOS chips, with both pins and logic levels compatible with old 74 series TTL chips (asymmetrical). In most cases CMOS (C/HC) chips will work in TTL logic boards and vice versa, but it is not recommended.
I don't mind the 'Octtandy'. I never had one, but it is very interesting to watch. In the Netherlands the Tandy computers weren't that populair like other machines and that makes it even more interesting to watch.
Earlier today, I was thinking about how we didn't get a 3rd video, but forgot to search after work. I' glad to see you kept going on this and used the term OcTandy. Good you can poke fun at yourself. Maybe you need to remove the mods to eliminate them from the equation and do more troubleshooting with the scope or a logic probe.
This one is tricky! I hope you find a solution. The cable looks good and looked like such a lot of work. I just had a random thought, if you could capture and print the 'rainbow' pattern while it is stable, it would almost make a piece of abstract art. All the best.
I was rewatching this video today. What happened to that computer? If you still have it, I think it is about time to have another try in fixing it! As the -5V was bad, I would think the 4116 RAMs are all bad. They die fast whenever one of the 3 voltages is bad. And it would be worth for basic troubleshooting: Voltages, clock, reset behavior, address and data lines and so on...
Personnally i had soldered 2 Bars of pin connectors (MB & Keyboard) and i had use a little old Floppy Cable with 2 34 pin femalle connectors (was easier than solder all wires :-) . Thanks for video on this dinausaurus :-)
I think he would have been better off to put a header on both the keyboard and the motherboard rather than soldering a ribbon cable like that. That's going to be a weak point.
Keep going Jan! I still have my original Model I, (Still working!) It's a fun trip down memory lane. I'm up to the point of messing with the expansion interface, and getting a floppy emulator going so I can load up some disk OSes. :)
Hi Jan - I strongly suspect that the static pattern you are getting means that the video RAM is not being written to. Have you considered replacing the whole modded ROM board with a simple ROM that does nothing but continually write to video RAM? Then you can check that the RAM is continually being written with the value that you expect, and if so, can then try and follow the video signals through. At the moment there is no way to be sure that the reason you are not seeing valid video is that the screen is blank!
I had problems with Q1 and Q2, the video output transistors. Q2 is a PNP (in some issues of the service manual incorrectly marked with MPS3904, a NPN transistor!). Check them and replace if necessary Q2 (MPS3906) with an BC327 (Reichelt, different pinout) and Q1 with a standard NPN transistor. If the sync then comes trough but no video signal (what seems not to be the case on your machine) check Z41.
Hi there Jan, I am suspecting that what you call rainbow is underlying color bars produced by the monitor itself (it seems there is an internal color bar generator in the monitor activated when it has no signal), rather than the TRS-80. Of course I cannot be sure, especially if I am not there, but I would try another monitor if I were you. Preferably an analogue one, in case you are using a composite to HDMI converter.
The pulses on the latter pins of the Z80 look wrong, ie, they should be up and down not a blob at the bottom and top, and I've seen those weird pulses before on Adrian's Digital Basement when he had bad RAM. Compere the early pin pulses to the latter ones, because just seeing activity doesn't mean they are correct.
Ah yes, the Colour Rainbow I've seen so often, especially on LCDs, when repairing MSX and other computers with no colour information. Which of course, the TRS-80 wouldn't. It is purty, though!
You can never have enough rainbows! Although in this case I would prefer a boring "MEMORY?" prompt in black and white... ;) I think it's most likely the upscaler/TV misinterpreting the noisy signal for a color burst.
Hello Jan, did you already check the Electrolytic capacitors? Somtimes a bad video signal is linked to a bad cap. So befor begin, I always change the caps.
@@JanBeta ok, but I would also change the big filter caps. Only to be sure that a bad video signal is not linked with any bad cap. (even you check the cap)
The video seems fine and stable to me. I'm not an expert but to me it looks more like there is memory corruption somewhere and it puts out garbage on the screen. You may want to check the character ROM(s) and memory chips to see if some of those fail. Just a thought. Good luck with the repairs! :)
No, the video isn't stable. It's showing color. The TRS-80 Model 1 is a black-and-white system. It doesn't even have grayscale, so the signal should be very very simple. A front porch, h-sync, back porch, and some high to low transitions for pixels. Somehow his video out is wonky enough that the TV is interpreting noise in the signal as a chroma burst.
Thanks, the video RAM (and character ROM/generator) most likely has issues, the video signal is definitely messed up, too, though. So I started there. ;)
How about some velcro strips to stick it back on? put the Velcro between the chips? use some kind of glue that easily dissolves with a chemical that won't kill the circuit board? (super glue?)
Make some notes put it on the shelf and in September 2021 pull it off the shelf just in time for another round of Septandy. Don't throw it away or it will indeed become a Trash80 which was its nick name back in the day. Sometimes when something frustrates me I shelf it for a while just to clear my mind then when I am in the mood for a challenge I just pull that device off the shelf and work on it again often resolving the issue that once frustrated me. Never give up and your swear bad word vocabulary is bound to get better and when you do resolve the issue its always a joyful event and then you play with it for a while then put it back on the shelf again.
So everything is replaced and it still doesn't work? LUL. I really hope Jan doesn't give up on this, after watching 3 long videos on this I'm really waiting on seeing this work!
Well Jan (@Jan Beta).. this is difficult one huh... actually sounds like a job for Mr. Black ... Mr. Adrian Black (@Adrian's Digital Basement) :-D Are you sure that the power supply delivers and sustains... "at all times" .. all the voltages and amps necessary... That the screen light up .. like normal startup .. and then dim down like that.... It may be because one or more voltages drop at boot .. when the big capacitor load up por example.... and the info in memory gets all screwed up for that momentary lack of voltage .. Check the stability of the voltages with the oscilloscope during the start up When it is on .. have you tried bending the board a little... or tab her... you know.. looking for false contacts in solder joins... or in broken traces...
• 21:38 - Is that solder from eBay? I have the same spool. How's your experience with it? I find that cheap Chinese solder is pretty crappy and makes soldering very difficult. I've recently started using an old spool of very thick brand-name "electrical" solder I got from a reputable store a long time ago and so far, it's been better than I remember. I guess it's the old thing about "you don't know how good you have it until you experience something worse to compare it to". 🤷 • 24:10 - Ugh! I made an IDE-to-edge-connector adapter a long time ago to make it easy to swap hard-drives in an out and this was such a pain. So many wires. 😕 • 26:10 - ♪ _Oops, there goes a billion kilowatt __-dam-__ RAM_ ♪
@@JanBeta I've only seen that happen three times, 2 was a flaky Z80, the other was due to the trim pots. I can't remember if they were broken, or needed cleaning. That was in the 80's when I was manager of the Tandy Computer Repair Center in Alexandria.
Hi Jan, I repair old 8-bit computers as a hobby. I've worked on several model 1s. I think you need to take a step back and start with what you know is wrong. What you know is wrong is that the video output signal is somehow messed up. The fact that you're getting a color signal means that the output is so messed up that your television thinks that it's seeing a "color burst" in the composite line. The TRS-80 doesn't _send_ a color burst signal because it's black and white. Here's a picture of a black and white NTSC composite video wave from a working model 1. imgur.com/gallery/OJwQkS4 We need to see what yours looks like, and from that we can start making decisions about where to look next. Is the computer sending proper sync signals? We don't know. Is the waveform distorted in particular parts of the signal? (Front porch messed up, back porch messed up, specific parts of the luminance part messed up?) We don't know. We don't have anything to go off of right now, and we're just sort of replacing chips in a shotgun approach. We need to narrow the scope by looking at the symptom closer.
I scoped the video output at various points in the previous video. It looked like a composite signal but with a lot of noise added, as you said. The signal you linked definitely looks way cleaner. The sync signals actually looked okay (also shown in the previous video) but the vsync was flickering for some reason (hence I why I replaced the chips in this video). I'm going to have to do more detective work and start at the beginning of the video divider chain. Thanks!
@@JanBeta Okay. Well, the video circuitry for this thing is pretty basic. It's got a sync side and a pixel side. Pin 1 of Z30 is the pixel data and pin 8 of Z5 is the combined sync signal. Here is a photo of the scope on pin 1 of Z30 (which is pixel data). imgur.com/gallery/YGyIznx Here is a photo of the scope on pin 8 of Z5 (sync data). It also shows a vertical sync fly by every once in a while (shows up as a longer high signal), but that is not depicted in the photo. imgur.com/gallery/uOskzEr If yours looks identical to these, then the problem is on the right side (the analog side) of the composite circuitry, which is essentially two transistors, half of a peripheral driver chip 75452, and some voltage divider resistors. If yours looks different than these, then the problem is further up the video chain. So I would suggest we divide the problem in half by comparing these images to yours.
By the way, just so you're aware, one of those transistors is mislabeled in some versions of the service manual schematic. Q2 is suppose to be a 2N3906. Some versions of the schematic show it as a 2N3904.
@Jan Beta I just rewatched your previous video. You need to go to time 39:05 in the first video and look at your composite signal again. Either your scope is set to the wrong time and voltage divisions, or your composite signal is very messed up. What I see is that your scope is set to 1V per division, and it stays floating at 2V all the time, and then occasionally drops down to 0V for horizontal sync. That makes no sense, because it means that you are sending white pixel data continuously, and there is no front or back porch after the horizontal sync. You need to look at pin 1 of Z30. If you have pixel data there, then one of the discrete components after Z30 (Z41, R23, R27, R28, R30, R22, C7, C2, Q1, or R30) are bad. Since it _appears_ that Q1 is always switched on, we could imagine that pin 4 of Z41 might not be grounded (which would allow back-flow through the base collector junction of Z41), or Q2 is somehow leaking current, or Q1 itself is just bad and is stuck on.
You're going to have to get a good look at the composite video signal to perhaps get some clues about what's wrong. In the last video you didn't accomplish that from what I saw.
Hi Jan, I've just been diagnosing my TRS80.. Your vertical and horizontal sync seem to be stable.. What concerns me is the graphics blobs you see on the screen.. I found out that if you have graphics blobs, corruption or snow on the screen, the fastest way to eliminate the Video Ram is to short the output of Video Ram bit D7 to ground. If you check the tech ref manual it tells you that if D7 Video ram bit is on the rest of the video ram data bits i.e. D0 - D6 will be interpreted as graphics characters in the upper ascii character set. On mine, the D7 bit was wonky and corrupted the screen.. See here th-cam.com/video/FCQNedxqE7A/w-d-xo.html
Kind of frustrating to watch Jan. Maybe we are missing all the methodical bits you are doing as you examine the schematic and logically starting at the clock, divider chain, putting the cro into a rate/sync mode for video so you can actually see the whole signal and not some random edge.
I did a lot of scoping in the previous video, including the video sync and clock signals. The vsync looked strange so that's why I replaced the ICs I suspected. Good point about putting the scope in the correct mode, obviously. Didn't think of that.
That might very well be! I think the main fault is the clock signal that is messed up somewhere but there's probably more faults down the line, possibly including the video ram. I am still working on it on and off but got a bit fed up with the machine frankly. I'll get there eventually, I guess... :)
Difficult to watch this one Jan. Semi-random component replacement in the hope that things will get better rather than structured troubleshooting and fault finding. Using a screwdriver like a chisel to lift a desoldered IC - almost guaranteed to rip traces and lift pads. Made me wince. CPU clock signal looks horrible and this runs the whole system including the video timing. Start there. What's the output from the crystal like? Trace the signal to the CPU checking it at the entry and exit of each device. Where does it get dirty? Without a good quality system clock signal nothing will work properly. Try removing the Z80 and see what the clock looks like without it - the CPU may be bad and corrupting the clock.
It really seems to me like you're just replacing random parts and not doing any real troubleshooting. If you suspect that a capacitor is bad then pull one lead out of the circuit and measure it -- it takes about three minutes (less time than replacing it) and costs nothing. Yes, parts are cheap and you have a nice desoldering gun but you're wasting time and money doing it this way.
Some computers can't be fixed. I know, that isn't right but you can't replace all parts. In the US the TRS-80 was called Trash Eighty and that must be for a reason. I remember Tandy / Radio Shack, they sold electronic parts on blisters. If you needed one resister you had to buy five, not spending fifteen or twenty cents on a resister but something like ƒ1,95 for the five and then. And then you had four of them in storage till you needed them again. All parts sold at Tandy / Radio Shack had questionable origins, like Hong Kong or even worst The Peoples Republic of China, and we all know what comes from China! The PRC copied Hong Kong and the Colony copied it from Japan. In retrospect I can say that the complete computer industry spoiled not millions of Dollars but thousands of billions of Dollars. Almost all computers ever produced could not be used for more then a few years, outdated because of shortcomings in the design. And the electronics and computer industry did the same with mobile telephones, they did their upmost best to produce equipment that is totally unrepairable. There is no need to repair if the design is old-fashioned when it leaves the factory, or when more money can be made selling new and newer stuff to the public. Therefor I can say that repairing a TRS-80 is against the will (and testimony) of Tandy (may God rest his soles). These computers were only produced to be sold to the illiterate public and then to be moved to the attic gathering dust. A year later another model came out, and again and again. It does not surprise me that Tandy is no longer in business. The pure hate for the customers was to obvious.
Enjoying the channel. Looks like you will do what I usually do for electronic repairs. Buy everything it could be and start swapping. It will work eventually. I recommend posting it on Noels discord. Maybe somebody there has an idea.
Hoi Jan. The rainbow you are getting is probably because the composite signal has some severe rippling after the sync pulse, where there should be just a flat front porch. Your oscilloscope images from Part 1 show just that, if you look closely. That rippling is probably being interpreted by the TV as a color burst (see: en.wikipedia.org/wiki/Colorburst ), and that's why you get colors instead of the B&W signal the TRS80 should be producing. Together with the severe rippling in the clock network, that suggests parasitics in the board, which could be caused by e.g. a bad chip hanging from the clock network. I haven't looked at the TRS80's schematics, but it should be possible to back-trace the composite signal to figure out where that rippling/"color burst" is coming from. Once you solve the rainbow issue, what will remain is a screwed-up but stable pattern on the screen, which suggests (video) memory issues. That's probably just some bad chips, which should be easy to diagnose. Don't give up, mate! This is just an 8-bit computer, there aren't many places where the problem could be hiding! (Addendum: the rippling in the clock network, according to your oscilloscope, happens on the low level of the clock. That suggests an undesired parasitic oscillator somewhere other than the CPU (though not 100% certain). That said, Z80s are still made, which are drop-in replacements for the Z80A in there. Here is the part number: Z84C0006PEC. They sell for only EUR 6 at Reichelt.)
Yes, that's the most reasonable explanation I've read so far. Thanks for that. I'm going to do some more detective work. :)
@@JanBeta Well Jan this is difficult one huh... actually sounds like a job for Mr. Black ... Mr. Adrian Black (@Adrian's Digital Basement) :-D
Are you sure that the power supply delivers and sustains... "at all times" .. all the voltages and amps necessary...
That the screen light up .. like normal startup .. and then dim down like that.... It may be because one or more voltages drop at boot .. when the big capacitor load up por example.... and the info in memory gets all screwed up for that momentary lack of voltage ..
Check the stability of the voltages with the oscilloscope during the start up
When it is on .. have you tried bending the board a little... or tab her... you know.. looking for false contacts in solder joins... or in broken traces...
Please work on this some more! The TRS-80 was my first computer, and I have very fond memories. Good luck!
Jan, as a novice, I find that watching along with your debugging and reading the comments and suggestions posted by your expert viewers I learn a lot that I would never learn watching other computer repair videos. For this reason, I appreciate very much your content and the discussion generated by your viewers. Thanks to both viewers and Jan for sharing your knowledge with us novices. It is very much appreciated. Particularly found the comments by spacedock8 and the byte attic helpful Cheers!
Hey, thanks! Glad that you like the learning by doing kind of approach. I learn a lot from the comments/discussion every time. :)
Looking forward to Part 3 Jan! Your channel is a real ‘find’ … as a relative newcomer and newbie I can’t tell you how encouraging + helpful your videos are. Don’t give up on the TRS-80 👍
Easy to sit back and criticize, don't give up. It's 2020, and there has been lots of crazy things happening. Without failure you never learn what you are really capable of doing. I have faith keep at it. Those complaining cannot say they have never failed. I love your innovation and your videos.
No worries that this repair bled over into October. It's an interesting repair. I do have a Model I, was working last time I had it running but that was many years ago. Afraid to use the included power supply as it's so old, perhaps I'll make one like you did. Very much like your channel, greetings from Norther Ohio in the U.S.
Thanks John! More detective work needed for this one. ;) I think you could refurbish the original power supply to an extend by replacing the capacitor and the diodes, should be standard parts as far as I know. I just had to make my own supply because the originals are quite impossible to find over here.
@@JanBeta I'll do that. I'm not much of a repair guy, but I can handle replacing the capacitor and the diodes. Thanks !
I have one of these and am the original owner. I replaced the Z80 back in Feb 2000. Everything works but the tape recorder circuit (as of 20 Feb 2021). If you want to compare notes, I can make some scope screen shots if that would help. Love the channel!
This series is nail-biting, I am at the edge of my seat. Can't wait for part 3!
Interesting video...I have worked on TRS-80 Model I computer back in the 80's doing all sorts of hardware mods. When the unit is first powered up, the ROM BIOS (well it wasn't call that then, but you get the idea) will initialize the video memory and clear the screen. While you do appear to have a video output issue, the display is also not clearing, so the system is not booting up. I would, as some other commenters have suggested, verify power supply ripple, as well as the clock chain as that might be preventing things from starting. A couple of the mods I had made include powering the board from a switching power supply, which improved system stability tremendously (basically fed it +/- 5V and 12V), and took a discrete video output (H/V sync + signal), so that you can feed a standard VGA monitor (I think it was CGA or EGA at the time), by picking off the TTL level signals for H and V sync from Z5, and video output off of Z30 pin 1.
I did some further troubleshooting in the meantime and there's definitely an issue with the clock signal getting messed up somewhere in the chain. I am going to get back to it eventually (the whole process was a bit disheartening to be honest so it takes some time to get back on track with the TRS repair...) Thanks for your input! :)
Just getting an early start on Septandy 2021, don't apologise.
Haha, good point!
Good morning Jan, I will look forward to watching this tonight . All the best .
Thanks! Hope you liked it. ;)
Keep trying Jan. I feel your frustration .
This was my first PC. I sold it to buy a C-64. Both are rubbish now. Great video.
You weren't the only one, SepTandy videos kind of ramped up more at the end of the month, time flies by with other things to do. ;)
There's still some OcTandy videos coming from other participants, too, afaik. ;)
I feel your pain, brother. I am having similar issues with my CoCo 2 with weird scrambled video output, the video signal is super noisy/ripply. Good luck with this -- I am sure you will prevail.
I'm not giving up yet, many more things to try. ;) Hope your CoCo issues can be fixed, too! And if it takes a whole Tandyear, so what! ;)
@@JanBeta Agreed. Hopefully, I'll get an #OctTandy video out and then #TandyYear is not out of the question.
Well done Jan. Keep at it. Doesn't have to be fixed right away, but you can consign the TRS 80 to the scrap pile.
Hello Jan, It's been a year and I was wondering if you were planning on resurrecting this project for SepTandy this year? I'd love to see it wrapped up. Thank you for all the great videos!
Would love to see a part 3. I do have some suggestions. You had found one of your address or data lines was being held down. This could be the issue. Check the warmth of your chips and see if they are getting warm. I'm pretty sure that video signal in the trash 80 required the processor to be running correctly. Keep at it. I'd like to see you fix it!
Yeah, I am currently at the point where I think the processor is not running correctly, too. I think the problem is the processor clock being too dirty to be interpreted right. I am working on it on and off, there is definitely going to be another video as soon as I have fixed the issue. :)
74HC04 is a recommended replacement of 74C04. HC series are CMOS chips pin compatible with 74 series, but with logic levels compatible with CMOS 4000 series chips (symmetrical) - same as 74C series, but faster. The HCT series are modern CMOS chips, with both pins and logic levels compatible with old 74 series TTL chips (asymmetrical). In most cases CMOS (C/HC) chips will work in TTL logic boards and vice versa, but it is not recommended.
I don't mind the 'Octtandy'. I never had one, but it is very interesting to watch. In the Netherlands the Tandy computers weren't that populair like other machines and that makes it even more interesting to watch.
Septandy is approaching again, will you make another video on this one?
Check the resistor values if you haven't done it yet. Carbon resistors can drift quite a bit over time.
I checked the ones relevant for the vsync timings, they were okay. I'll check more, thanks for the hint.
Earlier today, I was thinking about how we didn't get a 3rd video, but forgot to search after work. I' glad to see you kept going on this and used the term OcTandy. Good you can poke fun at yourself. Maybe you need to remove the mods to eliminate them from the equation and do more troubleshooting with the scope or a logic probe.
rager1969 3rd Video, noTandy?
Definitely not giving up on this one yet. Although it might take some serious detective work still. ;)
This one is tricky! I hope you find a solution. The cable looks good and looked like such a lot of work. I just had a random thought, if you could capture and print the 'rainbow' pattern while it is stable, it would almost make a piece of abstract art. All the best.
Thanks! Maybe I should get a screenshot of the rainbow indeed. :D
I use a lot of ribbon cable for bulk LCD displays
I was rewatching this video today. What happened to that computer? If you still have it, I think it is about time to have another try in fixing it! As the -5V was bad, I would think the 4116 RAMs are all bad. They die fast whenever one of the 3 voltages is bad. And it would be worth for basic troubleshooting: Voltages, clock, reset behavior, address and data lines and so on...
Personnally i had soldered 2 Bars of pin connectors (MB & Keyboard) and i had use a little old Floppy Cable with 2 34 pin femalle connectors (was easier than solder all wires :-) . Thanks for video on this dinausaurus :-)
I think he would have been better off to put a header on both the keyboard and the motherboard rather than soldering a ribbon cable like that. That's going to be a weak point.
Hi @Jan Beta , I suspect that one or more Ram Chips are bad, also check or replace the Roms with the Eproms with the TRS-80 Kernal
The capacitors were likely mounted above the board like that in an attempt to isolate them from heat.
Keep going Jan! I still have my original Model I, (Still working!) It's a fun trip down memory lane. I'm up to the point of messing with the expansion interface, and getting a floppy emulator going so I can load up some disk OSes. :)
I will get back to it, I promise!
Hi Jan - I strongly suspect that the static pattern you are getting means that the video RAM is not being written to. Have you considered replacing the whole modded ROM board with a simple ROM that does nothing but continually write to video RAM? Then you can check that the RAM is continually being written with the value that you expect, and if so, can then try and follow the video signals through. At the moment there is no way to be sure that the reason you are not seeing valid video is that the screen is blank!
I had problems with Q1 and Q2, the video output transistors. Q2 is a PNP (in some issues of the service manual incorrectly marked with MPS3904, a NPN transistor!). Check them and replace if necessary Q2 (MPS3906) with an BC327 (Reichelt, different pinout) and Q1 with a standard NPN transistor. If the sync then comes trough but no video signal (what seems not to be the case on your machine) check Z41.
Thanks! I checked the transistors and they seem to be fine from what I can see. The signal gets messed up somewhere before, it seems.
Hi there Jan, I am suspecting that what you call rainbow is underlying color bars produced by the monitor itself (it seems there is an internal color bar generator in the monitor activated when it has no signal), rather than the TRS-80. Of course I cannot be sure, especially if I am not there, but I would try another monitor if I were you. Preferably an analogue one, in case you are using a composite to HDMI converter.
cooool video .. you didn't cross the fingers during tests ...so it fails : - )
please keep up this interesting work
Michael
Thanks Michael! I'll try to remember crossing my fingers for the next tests! ;D
Interesting as always, thanks for the great effort you put into this video
Thanks! :)
The pulses on the latter pins of the Z80 look wrong, ie, they should be up and down not a blob at the bottom and top, and I've seen those weird pulses before on Adrian's Digital Basement when he had bad RAM. Compere the early pin pulses to the latter ones, because just seeing activity doesn't mean they are correct.
You get the dang thing fixed then have all year to collect more goodies for it and next 2021. Time for trash-80 again. :)
Haha, good idea. :D
CuriousMarc showed that crystals can go bad. He repaired HP clock boards this way.
The crystal looks fine on the scope. I'm currently suspecting one of the 74 ICs in the so called video divider chain. We'll see.
Ah yes, the Colour Rainbow I've seen so often, especially on LCDs, when repairing MSX and other computers with no colour information. Which of course, the TRS-80 wouldn't. It is purty, though!
You can never have enough rainbows! Although in this case I would prefer a boring "MEMORY?" prompt in black and white... ;) I think it's most likely the upscaler/TV misinterpreting the noisy signal for a color burst.
@@JanBeta Bonus Amiga demo effects are usually appreciated, just not helpful.
If it keeps on being tricky, maybe get a second (working) machine to swap parts & check component values?
2020 is the year of the Tandy takeover
Hello Jan, did you already check the Electrolytic capacitors? Somtimes a bad video signal is linked to a bad cap. So befor begin, I always change the caps.
Yes, replaced the electrolytics in the previous video (not the big filter caps, but they tested fine so I'm not too worried about those).
@@JanBeta ok, but I would also change the big filter caps. Only to be sure that a bad video signal is not linked with any bad cap. (even you check the cap)
The video seems fine and stable to me. I'm not an expert but to me it looks more like there is memory corruption somewhere and it puts out garbage on the screen. You may want to check the character ROM(s) and memory chips to see if some of those fail. Just a thought. Good luck with the repairs! :)
No, the video isn't stable. It's showing color. The TRS-80 Model 1 is a black-and-white system. It doesn't even have grayscale, so the signal should be very very simple. A front porch, h-sync, back porch, and some high to low transitions for pixels. Somehow his video out is wonky enough that the TV is interpreting noise in the signal as a chroma burst.
Thanks, the video RAM (and character ROM/generator) most likely has issues, the video signal is definitely messed up, too, though. So I started there. ;)
How about some velcro strips to stick it back on? put the Velcro between the chips? use some kind of glue that easily dissolves with a chemical that won't kill the circuit board? (super glue?)
enjoy, you have the only TRS-80 which can make million of colours!
Haha, good point. It's a miracle machine!
Like the "BETA" on the wrist strap.
Thanks! :)
I would have just started pulling out the socketed chips one by one to see if that made any difference to the picture initially.. :)
Oh, I did pull the RAM in the previous video. Maybe I should try removing the whole ROM assembly, too.
@@JanBeta it wouldn't hurt and you would be able to see if there's any corrosion in any of the sockets or on the chips too..
Any change with the removal of ROMs or CPU?
This biggest capacitor I have ever seen ... hundredth of a Farad.
You should see my 1 faradcap. It's bigger than a tallboy! :D
Now into OcTandy
Might be some more months even... ;D
Did this get launched into the bin or is it on the back burner?
I would disconnect the expansion board. That's one thing less that could case an issue.
Good point, I'm going to try that.
septandy..... now octandy?????? next notandy lolololol :D
Next video is probably going to be NoTandy. I need a break from this... ;D
Make some notes put it on the shelf and in September 2021 pull it off the shelf just in time for another round of Septandy. Don't throw it away or it will indeed become a Trash80 which was its nick name back in the day. Sometimes when something frustrates me I shelf it for a while just to clear my mind then when I am in the mood for a challenge I just pull that device off the shelf and work on it again often resolving the issue that once frustrated me. Never give up and your swear bad word vocabulary is bound to get better and when you do resolve the issue its always a joyful event and then you play with it for a while then put it back on the shelf again.
I'm definitely not giving up on it yet. Still quite a few things I need to check. :)
This is going to become the Trash of Theseus, isn't it?
So everything is replaced and it still doesn't work? LUL. I really hope Jan doesn't give up on this, after watching 3 long videos on this I'm really waiting on seeing this work!
Well Jan (@Jan Beta).. this is difficult one huh... actually sounds like a job for Mr. Black ... Mr. Adrian Black (@Adrian's Digital Basement) :-D
Are you sure that the power supply delivers and sustains... "at all times" .. all the voltages and amps necessary...
That the screen light up .. like normal startup .. and then dim down like that.... It may be because one or more voltages drop at boot .. when the big capacitor load up por example.... and the info in memory gets all screwed up for that momentary lack of voltage ..
Check the stability of the voltages with the oscilloscope during the start up
When it is on .. have you tried bending the board a little... or tab her... you know.. looking for false contacts in solder joins... or in broken traces...
You mentioned the timing be a little off , perhaps you might want to investigate that.
Is SepTandy going to last as long as DecAmber?
I sure hope not! But then again, there's going to be another SepTandy next year... ;)
Given the weird pattern, would that indicate some kind of corruption in the video RAM, ie could it be faulty RAM?
Yes, most likely the video RAM is corrupted. The system RAM has been replaced.
• 21:38 - Is that solder from eBay? I have the same spool. How's your experience with it? I find that cheap Chinese solder is pretty crappy and makes soldering very difficult. I've recently started using an old spool of very thick brand-name "electrical" solder I got from a reputable store a long time ago and so far, it's been better than I remember. I guess it's the old thing about "you don't know how good you have it until you experience something worse to compare it to". 🤷
• 24:10 - Ugh! I made an IDE-to-edge-connector adapter a long time ago to make it easy to swap hard-drives in an out and this was such a pain. So many wires. 😕
• 26:10 - ♪ _Oops, there goes a billion kilowatt __-dam-__ RAM_ ♪
If I remember correct a Model I doesn't boot right when the keyboard is not connected.
It doesn't, but it should at least show the "Memory?" prompt. It basically keeps "typing" nonsense at that prompt when the keyboard isn't connected.
@@Calphool222 yes! Now I remember. You are correct.
Yes, that's what I figured. It won't boot fully but should at least not display garbage. ;)
#octandy now :D
Yay!
When you buy a new soldering iron
@Ken Mason pace ads 200 it is a very good soldering iron and it is cheap for her power and quality
I keep screaming into my monitor, 'Replace the Z80", but you don't seem to be hearing me, for some reason....
I tried a known good one, same result. :)
@@JanBeta I've only seen that happen three times, 2 was a flaky Z80, the other was due to the trim pots. I can't remember if they were broken, or needed cleaning. That was in the 80's when I was manager of the Tandy Computer Repair Center in Alexandria.
I’ll inspect the potentiometers again, thanks for the tips!
check the ram
I replaced the RAM in the previous video, also tested without any RAM.
Hi Jan, I repair old 8-bit computers as a hobby. I've worked on several model 1s. I think you need to take a step back and start with what you know is wrong. What you know is wrong is that the video output signal is somehow messed up. The fact that you're getting a color signal means that the output is so messed up that your television thinks that it's seeing a "color burst" in the composite line. The TRS-80 doesn't _send_ a color burst signal because it's black and white.
Here's a picture of a black and white NTSC composite video wave from a working model 1.
imgur.com/gallery/OJwQkS4
We need to see what yours looks like, and from that we can start making decisions about where to look next. Is the computer sending proper sync signals? We don't know. Is the waveform distorted in particular parts of the signal? (Front porch messed up, back porch messed up, specific parts of the luminance part messed up?) We don't know. We don't have anything to go off of right now, and we're just sort of replacing chips in a shotgun approach. We need to narrow the scope by looking at the symptom closer.
I scoped the video output at various points in the previous video. It looked like a composite signal but with a lot of noise added, as you said. The signal you linked definitely looks way cleaner. The sync signals actually looked okay (also shown in the previous video) but the vsync was flickering for some reason (hence I why I replaced the chips in this video). I'm going to have to do more detective work and start at the beginning of the video divider chain. Thanks!
@@JanBeta Okay. Well, the video circuitry for this thing is pretty basic. It's got a sync side and a pixel side. Pin 1 of Z30 is the pixel data and pin 8 of Z5 is the combined sync signal.
Here is a photo of the scope on pin 1 of Z30 (which is pixel data).
imgur.com/gallery/YGyIznx
Here is a photo of the scope on pin 8 of Z5 (sync data). It also shows a vertical sync fly by every once in a while (shows up as a longer high signal), but that is not depicted in the photo.
imgur.com/gallery/uOskzEr
If yours looks identical to these, then the problem is on the right side (the analog side) of the composite circuitry, which is essentially two transistors, half of a peripheral driver chip 75452, and some voltage divider resistors. If yours looks different than these, then the problem is further up the video chain. So I would suggest we divide the problem in half by comparing these images to yours.
By the way, just so you're aware, one of those transistors is mislabeled in some versions of the service manual schematic. Q2 is suppose to be a 2N3906. Some versions of the schematic show it as a 2N3904.
@Jan Beta I just rewatched your previous video. You need to go to time 39:05 in the first video and look at your composite signal again. Either your scope is set to the wrong time and voltage divisions, or your composite signal is very messed up. What I see is that your scope is set to 1V per division, and it stays floating at 2V all the time, and then occasionally drops down to 0V for horizontal sync. That makes no sense, because it means that you are sending white pixel data continuously, and there is no front or back porch after the horizontal sync. You need to look at pin 1 of Z30. If you have pixel data there, then one of the discrete components after Z30 (Z41, R23, R27, R28, R30, R22, C7, C2, Q1, or R30) are bad. Since it _appears_ that Q1 is always switched on, we could imagine that pin 4 of Z41 might not be grounded (which would allow back-flow through the base collector junction of Z41), or Q2 is somehow leaking current, or Q1 itself is just bad and is stuck on.
Still not working? Why does God hate you Jan? What did you do?
You're going to have to get a good look at the composite video signal to perhaps get some clues about what's wrong. In the last video you didn't accomplish that from what I saw.
Hi Jan, I've just been diagnosing my TRS80.. Your vertical and horizontal sync seem to be stable.. What concerns me is the graphics blobs you see on the screen.. I found out that if you have graphics blobs, corruption or snow on the screen, the fastest way to eliminate the Video Ram is to short the output of Video Ram bit D7 to ground. If you check the tech ref manual it tells you that if D7 Video ram bit is on the rest of the video ram data bits i.e. D0 - D6 will be interpreted as graphics characters in the upper ascii character set. On mine, the D7 bit was wonky and corrupted the screen.. See here th-cam.com/video/FCQNedxqE7A/w-d-xo.html
Kind of frustrating to watch Jan. Maybe we are missing all the methodical bits you are doing as you examine the schematic and logically starting at the clock, divider chain, putting the cro into a rate/sync mode for video so you can actually see the whole signal and not some random edge.
I did a lot of scoping in the previous video, including the video sync and clock signals. The vsync looked strange so that's why I replaced the ICs I suspected. Good point about putting the scope in the correct mode, obviously. Didn't think of that.
Hi jan the video you are looking at is a bad video ram chip 2102
That might very well be! I think the main fault is the clock signal that is messed up somewhere but there's probably more faults down the line, possibly including the video ram. I am still working on it on and off but got a bit fed up with the machine frankly. I'll get there eventually, I guess... :)
Looks like bad RAM to me.
The video RAM is most likely bad, the system RAM has been replaced, also tested without it.
Difficult to watch this one Jan. Semi-random component replacement in the hope that things will get better rather than structured troubleshooting and fault finding. Using a screwdriver like a chisel to lift a desoldered IC - almost guaranteed to rip traces and lift pads. Made me wince. CPU clock signal looks horrible and this runs the whole system including the video timing. Start there. What's the output from the crystal like? Trace the signal to the CPU checking it at the entry and exit of each device. Where does it get dirty? Without a good quality system clock signal nothing will work properly. Try removing the Z80 and see what the clock looks like without it - the CPU may be bad and corrupting the clock.
Mwa.. could have been much much much worse: th-cam.com/video/Wh2OCBZpzZ8/w-d-xo.html
Bas van Haren yeah, he even disabled the comments!
@@andrewlittleboy8532 Yeah I probably removed the whole video if I were him.
@@basvanharen2904 Yes, that was extremely painful to watch!
Video is static ram. You can get schematics off net and in its. Service manual.
It really seems to me like you're just replacing random parts and not doing any real troubleshooting. If you suspect that a capacitor is bad then pull one lead out of the circuit and measure it -- it takes about three minutes (less time than replacing it) and costs nothing. Yes, parts are cheap and you have a nice desoldering gun but you're wasting time and money doing it this way.
Some computers can't be fixed. I know, that isn't right but you can't replace all parts. In the US the TRS-80 was called Trash Eighty and that must be for a reason. I remember Tandy / Radio Shack, they sold electronic parts on blisters. If you needed one resister you had to buy five, not spending fifteen or twenty cents on a resister but something like ƒ1,95 for the five and then. And then you had four of them in storage till you needed them again. All parts sold at Tandy / Radio Shack had questionable origins, like Hong Kong or even worst The Peoples Republic of China, and we all know what comes from China! The PRC copied Hong Kong and the Colony copied it from Japan.
In retrospect I can say that the complete computer industry spoiled not millions of Dollars but thousands of billions of Dollars. Almost all computers ever produced could not be used for more then a few years, outdated because of shortcomings in the design. And the electronics and computer industry did the same with mobile telephones, they did their upmost best to produce equipment that is totally unrepairable. There is no need to repair if the design is old-fashioned when it leaves the factory, or when more money can be made selling new and newer stuff to the public.
Therefor I can say that repairing a TRS-80 is against the will (and testimony) of Tandy (may God rest his soles). These computers were only produced to be sold to the illiterate public and then to be moved to the attic gathering dust. A year later another model came out, and again and again. It does not surprise me that Tandy is no longer in business. The pure hate for the customers was to obvious.
Enjoying the channel. Looks like you will do what I usually do for electronic repairs. Buy everything it could be and start swapping. It will work eventually. I recommend posting it on Noels discord. Maybe somebody there has an idea.
Hi..The kukotronix rules forever 🫡🫡🫡⚡️⚡️⚡️