telmnstr

Hello, I hit the issue you are talking about last night. I took the SIPP module from the top of the 4MB Best upgrade board, desoldered it, then soldered it on the other side. I think the original board probably had two SIMM sockets where the chips were at 45 degrees facing the expansion slot underneath? I could be wrong. I assume Best soldered on the SIPP modules and maybe isn't aware that it makes it a tough fit. I shot them an email about it. Now my board has the two SIPPs facing the downward expansion slot and the cover fits on fine. But for some reason with the cover on, the system isn't coming up right. But it definitely worked on the bench before re-assembly. I did a ASCI2SD using the blue pill thing on mine, and 3d printed a mount so I could remove the SD card from outside. I should have taken the guts of a gotek and added that as a 2nd floppy, but I have a gotek mounted in an external Atari ST 3rd party drive box, that should be good enough. I was thinking about trimming down my cover a little around the LCD cable though, to make it easier to put back together.

Working on a 1MB Atari STacy project. Would like to bump it to at least 2MB or 4MB of RAM.

Replaced all 40 pin dip sockets, no change. Sockets on the way for the 16 pin (RAM, others) ICs. Will order the common logic chips on my digikey order this weekend.

Actually that is what I did, I programmed the RAM tester into both the top and bottom, but the top side was what was needed to make it work. When it crashes and the RAM tester is running, it points to random ICs. Sometimes it will point to 4 of them, or any of them. But the RAM was all tested in a working 800XL and seemed fine. When it's crashing it's going pretty left. It usually crashes in 20 seconds or so, but doesn't always run on boot -- takes a number of tries. Centipede cart will usually start to get graphics noise before it stops executing. The fact that it works cold then always stops is strange. Maybe tonight I will run it, get it in a crashed state and oscope the data and address lines to see if any are stuck? I looked around before and nothing stood out but I didn't have anything specific I was looking for.

Okay, I didn't have a 27C128 EPROM but I had a 27C256. I read on a corvette forum they are pin compatible. I wrote the diag utility to the lower half, it didn't work. Wrote it to the upper half (starting at 4000) and it booted right up. It crashes as it always does. I can get it to run for a good while, it never really reports errors until things blow up. I replaced the DIP socket under the OS ROM with a SIP socket (they are strips of machined pin holes that are breakable. Microcenter had them, but I don't know if they have any more. If they did I would buy more.) I did replace the 74LS08 with a chip I had on hand, no difference. I don't have all of the other logic chips. By pressing on chips it seems like I can get it to run longer but not always. It's so strange. It's like there is a mechanical break somewhere but for the life of me I can't see where. Or it could be something else messing with my mind. Next plan of attack might be to replace the 5 40 pin sockets? I just feel like with the number of times I've pulled and placed chips plus the deoxit I don't think the socket contact would be an issue. And I reflowed all of them. I'm worried it's one of the traces and the foil is cracked right at a solder pad. I had that on a synthesizer and you couldn't tell from looking at it that it was broken. Much thanks for all the tips! The 800XL PCB on the bench... still: https://imgur.com/a/L8VAo8t Maybe I will warm it up, see if that changes anything....

Okay, so last night I took my good candidate 800XL and de-soldered the MMU and added a socket (had one on hand.) Dropped the good 800XL MMU in the problem 800XL. No difference. On a whim, in the spirit of the cold spray I put an ice cube in a plastic baggie and sit it on the OS ROM. OS ROM is one of the chips I couldn't swap since candidate isn't socketed. The machine ran longer than I ever have seen, Centipede made it to the bottom of the screen and back up several times! Thinking I had a win, I de-soldered the OS ROM from the working 800XL and dropped it in the socket of the bad 800XL... and it didn't fix it. Past that point the ice cube didn't seem to work anymore. I don't have sockets that fit the OS ROM, and neither does the local Microcenter. I wonder if it's a cracked solder joint on the top of the PCB under a socket or something like that. That RAM tester program looks good! I think I have some 27c128s in the stash, if so I can write it out. Awesome how it shows which chip is failing, if only the built in RAM test did something like that.

Usually when it works it runs for about 30 seconds max, and then crashes. If I power cycle it, the crashes usually happen much quicker. I have swapped all the RAM between a working system and the failed system and no changes. So it doesn't seem to be the RAM. Tonight I socketed the MMU on my working system and tried swapping those. Nothing different. I used a multimeter and checked the address lines through the RAM ICs, they seem fine. Tested from a pin on say chip 1 up through all others. Found broken traces on a Robotron machine doing this once. Also toned the data bus from the BASIC ROM back to the CPU, it was okay. Played with OScope on the address lines a little, it looked strange to me on the non-working system but seemed to match the working. Haven't gone fully crazy. Somewhere there is a list if pins should be high or low on a system sitting in BASIC I think, I need to find that. I haven't looked at the delay line stuff yet. I saw it in the schematics. I had it running with a fan on it to cool it down, that didn't seem to make a difference. I have used a heat gun on electrolytics to change the ESR for troubleshooting those in the past. Last night I couldn't get it into BASIC and it was black screened, but tonight it is working. It is either a marginal component or a trace/socket issue? One chip I haven't swapped is the OS PROM. It's socketed on my problem child but not socketed on the working one. Would a cart (Centipede and Star Raiders are my test carts) even use this OS PROM? It didn't seem to work with it pulled.

Reflowed every RAM solder joint, all 40 pin chips and some of the others. Found one poly cap that barely had any solder on it from the factory. Back to black screen. Tonight if I get a chance I'm going to start ringing out with a multimeter the data and address lines and make sure they go through all of the RAM chips properly Sometimes it will boot now with Star Raiders, haven't seen BASIC in a while though. Tough one.

The XL styled power supply does work as well, I just haven't been using it and haven't checked it out. I went for the "last known working" which is the one I used when I repaired a 130XE that had failed RAM ICs. I could replace all the sockets with machine pin sockets. I'll try deoxit and reseat tonight on the RAM ICs and maybe reflow. I have a Hakko 808 desoldering tool so that would help speed up replacement of all the sockets but I would need to order the sockets.

Yea, I've been using the one that I think is from the 130XE for both. I saw that the "ingot" supply is considered bad. This machine came with the power supply that has the "XL" styling which is awesome, but I haven't been using it. I plan to, but IIRC they usually need recap or overhaul first. My original 800XL had one of the XL styles power supplies but died when I was young.

Okay swapped the 74LS08 last night (I just can't stop working on projects once I start!) Black screen back again! Which is kind of good. I pulled PCB out and started working around on the bottom with something plastic trying to find any broken solder joints. Seems like in the RAM area it started working for a bit. So going to go full reflow tonight and deoxit the RAM IC sockets. Will report back.

Hello! I double checked some of them but will redo it. I actually just pulled all four of the 40 pin chips from the failing machine, dropped them in my working 800XL (not the pokey) and it ran fine for 40 minutes. Everything on the non-working machine is stock. Working machine is stock + dual pokey. The caps I swapped were all the electrolytics except the 3 bipolars near the cart slot. I just double checked, polarity looks right. Didn't touch the 4 poly caps or anything. There are components on the board that look like diodes that have Cxx part numbers silkscreened, didn't touch those. Tomorrow I will reflow all RAM ICs, and all of the 40 pin dip sockets (only did the CPU so far.) There *was* what looked like reflow or something around the 08 and other ICs in that area, or flux wasn't cleaned off the board all the way. Looking for an 74LS08, I know I had them somewhere. Thanks for the pointers!

Hello! The Atari 800XL was my first computer. I still have my original from my childhood, which I modified to dual POKEY in the 2000s or so. It works fine. I recently came into possession of a non working 800XL. And this weekend started to take a crack at it. Prior to this I recapped the electrolytic caps in it except the bipolars but that made no change, set it on a shelf back then. Would like to get it running to bring it out to public events when they start again. Symptoms were black screen. All chips socketed on the dead unit, RAM And most customs socketed on my childhood one but not the LS logic, ROMs and MMU. Reseated chips, no change. Swapped RAM and CPU and wala! It started working! Progress! So I thought. Star Raiders cart as I've heard that is a good place to start. But after X seconds the screen turns to garbage and it dies. Checked DC voltage at the POKEY and RAM, solid 5.05VDC through the crash. It crashes in game (Star Raiders, Centipede cart) or BASIC. I have tried swapping GTIA, ANTIC, SALLY (CPU), PIA with my working childhood machine. It seems to lock up still. After frustration, I unsodered a few wires and swapped the dual POKEY stack into the failing one and no change. The CPU gets hot, but I checked my working machine and it feels similar and is stable. I reflowed the socket pins to the mainboard on the SALLY thinking that maybe heat.... When it's working I can "percussive test" it and that doesn't seem to kill it. I shot deoxit into the sockets of the main chips (not RAM) and no difference. It seems to run longer on a cold boot than a warm reboot. System diags (built in not SALT cart) crash. I did have one LS logic chip that I swapped but that made no difference. My working machine MMU and OS ROM are soldered in.on the donor (as they would be inside 130XEs.) I did pull the BASIC ROM just incase that was throwing off data/address bus somehow. Anyone encounter such an issue or know of anything on these machines that lead to this? I haven't scoped it yet... but not sure what I would really look for other than maybe a clock oscillator issue or noise on the DC rail?

Are they actually shaped different on the return key or is it the same? Would the one from the Spacebar fit?

Hello! I was recently working on restoring a 1040ST for a music and gaming festival I am involved with. The spacebar has a balancing rod under it that keeps the space bar even regardless of where it's hit. These are common. There are these two plastic parts that insert into the spacebar keycap that capture the metal rod. One of them was broken from age, the other one was broken by me removing the spacebar to clean under it. I managed to model out the part and was able to print satisfactory replacements using a pretty run of the mill sub $250 3d printer (Ender 3.) I posted the STL file on Thingiverse in-case others run into this issue. Good chance this keyboard part is used by many of the ST series. Here is the link to the Thingiverse download: https://www.thingiverse.com/thing:3887995 I also attached the same file to this post. Enjoy. Atari_1040ST_SpaceBar_Balancer_Thing.stl