Sergioz82 Posted February 24, 2022 Share Posted February 24, 2022 (edited) Hi. Unfortunately my TI is getting worse every day. It started with some random glitches here and there, these days my screen goes crazy almost every time. In the pictures it's what I see in the splash screen and during a game (Biometeor, in this case). Although it looks like a ram issue I'm not very sure.. is it? Could it be something else? Screen aside, the computer works correctly. I can navigate the final grom menu, or even play the game. Also, if I leave the splash screen on wait, the glitch pattern changes If I check the ram with TI diagnostic I can randomly get 600, 3700, 6400 or 0 errors. The video display, sound and calculation tests always work fine. Edit: while I was writing this post I let the calculation tests running. After some minutes the screen glitched. I couldn't read the text well but it seems I don't get the green square on misc.function and tells me to check manual A reset solves the situation for some minutes or for some hours. What could it be? Edited February 24, 2022 by Sergioz82 Quote Link to comment Share on other sites More sharing options...
WhataKowinkydink Posted February 24, 2022 Share Posted February 24, 2022 Bad RAM would be my first bet here. 1 Quote Link to comment Share on other sites More sharing options...
Stuart Posted February 24, 2022 Share Posted February 24, 2022 It's another case of ... 4 1 Quote Link to comment Share on other sites More sharing options...
apersson850 Posted February 25, 2022 Share Posted February 25, 2022 I had screen garble once. Removed the TMS 9929 and re-seated it in the socket. It's in a socket just to reach high enough to touch the cooling body. Problem disappeared. 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 25, 2022 Author Share Posted February 25, 2022 20 minutes ago, apersson850 said: I had screen garble once. Removed the TMS 9929 and re-seated it in the socket. It's in a socket just to reach high enough to touch the cooling body. Problem disappeared. Thanks, I can try this, I hope I'm lucky enough to solve the problem. Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 25, 2022 Author Share Posted February 25, 2022 Thanks all for the replies. In case I had to replace RAM/CPU or whatever where can I find the original chips or second choice the compatibile ones? I had a quick look on eBay and I found lots of chips from China but what are they? Surplus or just fakes? Quote Link to comment Share on other sites More sharing options...
+arcadeshopper Posted February 25, 2022 Share Posted February 25, 2022 6 hours ago, Sergioz82 said: Thanks all for the replies. In case I had to replace RAM/CPU or whatever where can I find the original chips or second choice the compatibile ones? I had a quick look on eBay and I found lots of chips from China but what are they? Surplus or just fakes? I have 4164's that work better than 4116 as replacements.. https://www.arcadeshopper.com/wp/ti-99-4a-faq-repairs-and-troubleshooting/ 1 Quote Link to comment Share on other sites More sharing options...
+mizapf Posted February 25, 2022 Share Posted February 25, 2022 Do you always get the same output on the master title screen? (last picture) The error symptom is not really easy to interpret. This is the nearest that I could simulate with MAME. It is created by this line when writing to VRAM: if ((m_Addr & 1)==0) data = data ^ 0x08; which would mean to flip the bit 3 (....x...) for every even address. However, it does not really match when you compare the texts. Could you try to create a second shot of the master title screen and see whether it is different? 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 25, 2022 Author Share Posted February 25, 2022 (edited) 2 hours ago, mizapf said: Do you always get the same output on the master title screen? (last picture) The error symptom is not really easy to interpret. This is the nearest that I could simulate with MAME. It is created by this line when writing to VRAM: if ((m_Addr & 1)==0) data = data ^ 0x08; which would mean to flip the bit 3 (....x...) for every even address. However, it does not really match when you compare the texts. Could you try to create a second shot of the master title screen and see whether it is different? The image you recreated it's quite similar. To answer your question no, it's not always the same output. If I don't press any key some characters on the screen change randomly, but not many (let's say 3-4 in the top right part of the screen). At the moment I can't take another picture sorry, but I can add this one I took yesterday in the same occasion. It's the CPU test of the diagnostic cart the moment the screen glitched Edit: here's another picture with a different pattern Edited February 25, 2022 by Sergioz82 Quote Link to comment Share on other sites More sharing options...
+mizapf Posted February 25, 2022 Share Posted February 25, 2022 If it is bit 4 (counted from the left), you should replace the second VRAM chip from the left. See https://www.ninerpedia.org/wiki/Troubleshooting (I said bit 3 in the post above since I counted from the right.) 1 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 25, 2022 Author Share Posted February 25, 2022 5 minutes ago, mizapf said: If it is bit 4 (counted from the left), you should replace the second VRAM chip from the left. See https://www.ninerpedia.org/wiki/Troubleshooting (I said bit 3 in the post above since I counted from the right.) Ok. Thank you very much. Tomorrow I'll try to check the motherboard, just to exclude a problem with solders for example. I doubt I'll be so lucky so next I'll get the ram chip replacement Quote Link to comment Share on other sites More sharing options...
+mizapf Posted February 25, 2022 Share Posted February 25, 2022 As you can see, the errors seem to be distributed in a non-deterministic way. At the marked positions we have this bit flip. Changing the RAM circuit may help (since it is always that bit that is flipped), but it is not guaranteed. I am a bit concerned about this flip kind; why should a 1-bit flip to 0, and a 0-bit flip to 1? This could also mean a problem with the data bus input of the video processor, and then you'd need a new video processor. 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 26, 2022 Author Share Posted February 26, 2022 (edited) 13 hours ago, mizapf said: As you can see, the errors seem to be distributed in a non-deterministic way. At the marked positions we have this bit flip. Changing the RAM circuit may help (since it is always that bit that is flipped), but it is not guaranteed. I am a bit concerned about this flip kind; why should a 1-bit flip to 0, and a 0-bit flip to 1? This could also mean a problem with the data bus input of the video processor, and then you'd need a new video processor. Today I checked the motherboard. At a visual inspection everything looks fine. However I identified a couple of intervention: 1) replace the thermal paste 2) Replace some capacitors (I don't like the bulge on one side.. they're not leaking but that's not the normal look This aside, after I reassembled the computer the problem disappeared and I'm puzzled. Considering that last day it glitched in a matter of minutes and resets only temporary solved the situation and now I used it for an hour and it worked fine it's weird. I just did these things: 1) push the Groms on their sockets (I heard a click in a couple of them) 2) spread a little the thermal paste (I haven't got any at home at the moment) 3) Clean the cartridge connector with isopropyl alcohol Edited February 26, 2022 by Sergioz82 3 Quote Link to comment Share on other sites More sharing options...
Tursi Posted February 26, 2022 Share Posted February 26, 2022 The VDP will overheat rather quickly if it's not making good contact with the heat sink -- on some systems. We have some people who've run without a heat sink for hours, I've had systems that start to glitch within minutes. So the thermal paste may well have helped. 1 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 27, 2022 Author Share Posted February 27, 2022 (edited) 12 hours ago, Tursi said: The VDP will overheat rather quickly if it's not making good contact with the heat sink -- on some systems. We have some people who've run without a heat sink for hours, I've had systems that start to glitch within minutes. So the thermal paste may well have helped. Thanks. Unfortunately the problem reappeared some time later, but this time I managed to catch the change in the glitch patterns. It happens about 30 seconds after the screen gets corrupted. It can happen at cold start or 10-15 minutes after it's turned on. In the first two pictures, the cyan dirt near the game over text looks like a sprite, it keeps scrolling downard. If I understood correctly, a RAM problem gives a fixed pattern (correct me if I'm wrong) so could it be a VDP-related problem? Anyway, I'll surely replace the thermal paste which is almost totally dry. I hope it will be enough. Edited February 27, 2022 by Sergioz82 Quote Link to comment Share on other sites More sharing options...
+mizapf Posted February 27, 2022 Share Posted February 27, 2022 Do you have an infrared thermometer? Try to measure the temperature of the VDP chip or of other chips. If it does not get really warm, it's probably not an overheat problem. 1 Quote Link to comment Share on other sites More sharing options...
Ed in SoDak Posted February 27, 2022 Share Posted February 27, 2022 Or a chill spray. Those spray cans that blow out your keyboard, if inverted, unleash a frigid blast. Or there's freeze sprays specifically meant for chilling circuit parts. Hit the VDP or other questionable chip when it becomes flaky, if it clears up, good chance you have found your offender. I run a small DC confuser fan over my console's ventilation slots behind the cart port. Power it off a 3-volt DC wallwart so it's nice and quiet but still moves a decent amount of air. The ol' "coffee warmer" stays nice and cool so I know it's helping and lockups became a thing of the past. 1 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 28, 2022 Author Share Posted February 28, 2022 12 hours ago, mizapf said: Do you have an infrared thermometer? Try to measure the temperature of the VDP chip or of other chips. If it does not get really warm, it's probably not an overheat problem. Unfortunately I don't. I'm going to do the following steps: 1) Test with chill spray as @Ed in SoDak suggested 2) If heat dissipation is the problem then replace the thermal paste 3) replace faulty RAM chip(s) as you suggested. Thank you all for the suggestions. I'll let you know how it ends. Quote Link to comment Share on other sites More sharing options...
Stuart Posted February 28, 2022 Share Posted February 28, 2022 I don't think you've said whether you've popped it out of the socket and put it back in again, just in case ...? 1 Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted February 28, 2022 Author Share Posted February 28, 2022 18 minutes ago, Stuart said: I don't think you've said whether you've popped it out of the socket and put it back in again, just in case ...? I didn't. This is another thing to try. Thanks. Quote Link to comment Share on other sites More sharing options...
Sergioz82 Posted March 6, 2022 Author Share Posted March 6, 2022 A small update. Last Wednesday I disconnected and reconnected the VDP from its socket, then I replaced the thermal paste. The first hour it worked worse than usual: instant glitches, flickering screen (never happened before), random sounds and so on. I thought the damage was worsening. At that point I ordered VDP and RAM replacement. Now the weird part: the next day I used the TI for a while to see what happened next. Well.. nothing. I've been using the computer since then and, I don't say it too loud, for the moment no glitches even after some hours of functioning. I think thermal paste helped, what I can't explain is why it took a while to work Quote Link to comment Share on other sites More sharing options...
GDMike Posted March 6, 2022 Share Posted March 6, 2022 (edited) I've got 3 ea 9918s and none work after removing, but the machines work with my 9918A card installed. 1 of the 9918s gives a black flameout loop flipping kinda thing, but that's some kind of life. I'm not sure what happened either, just popped em out with a small flat blade on a non static table, I didn't have shoes on, I don't even think it mattered if my tounge was grounded, no signs of bent pins, I don't see any crap in the sockets either. But whatever..I'll replace all with F18As anyway. But doesn't that piss you off. Edited March 6, 2022 by GDMike 3 Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.