Maxlean

Yeah, I reinstalled the removed resistor. 3k ohms, I believe. Hadn’t thought about cold solder joints on the sockets, etc. I’ll check all that next. With as many times as I’ve removed and installed most of the chips, and cleaning with deoxit, I’d think they’re sparkly clean at this point. Lol.

Ok guys, I'm stumped. I've got my original 800XL from 1985, and dug it out of storage a few months ago to start having some modern Atari fun. Fired it up and was greeted by a black screen. This machine has a Newell Industries 256k upgrade, btw. I deoxed all the sockets and re-seated the chips, and she came back to life, but was unstable. Ran the memory test, and it would produce errors, sometimes video corruption, and sometimes lock up. I removed the Newell 256k, but left the ram chips since I don't have the original 64k chips, but I've been told that this is fine, no change. Bought new 256k ram chips, (like 32 of them), and swapped them. Thought there was some improvement, but it'll still randomly fail the ram test, and sometimes has the video corruption, and still sometimes locks. Sometimes, it'll run for 10 minutes on the ram test with no faults. So I dug out my new in box 1200XL, which tests perfectly fine, and decided to swap chips, one at a time. Swapped the ram, ANTIC, GTIA, POKEY, CPU, and PIA, all one at time. Nothing changed. I haven't swapped the smaller ICs yet. I'm kinda leaning toward the MMU, but unfortunately, the 1200XL uses a different one, so I can't swap that, and of course can't swap the OS roms. I've also measured the power supply while running a failing mem test, and voltage is stable at 5.09v. Any suggestions? Thanks, Scott

Every chip on the board got the deoxit and reseat before any serious diagnosing was started.

I like that! Not surprising that I can't remember the specifics (or anything about this mod) after 36 years. Heck, I'm grateful that I can still remember how to turn the darn thing on! And on a related note, I de-modded the board. Removed the 256kxl, reinstalled the 74ls258 that the 256kxl replaces, reinstalled resistor r32, and removed the mystery mod. Powered it up, and it's working pretty much as it was before demodding, as in it's randomly failing the memory test. I had ordered 24 replacement 256k memory chips (cheaper in bulk) a couple of weeks ago and picked 8 to replace the ones that I had on the board from the 256kxl mod. As expected, they seem to work fine on a stock board, at 64k of course. However, I still was getting random memory test failures. Tried reinstalling the original 256k ram chips, and was greeted with a black screen. Installed 8 of the new chips, and it functions but still has the random errors. Eventually found a combination of chips that didn't produce memory errors, and then methodically swapped one chip at a time until errors returned. Thought I had settled on two bad ones. One from the original set, and one from the new set. Success, right? Not quite. While I'm still getting no errors, I can no longer replicate a faulty chip no matter what combination of suspected bad chips I substituted. It almost seems that the longer I let the machine run, the more and more stable it became. Now before anyone says it was a poor contact on one of the sockets, I had already used Deoxit on ALL the chips (fully socketed board). Removed each chip, sprayed the socket and chip, and then reinserted multiple times. All this on each and every chip. So I don't think this latest round of memory swapping and success were a result of a poor connection, but you never know. Now as luck would have it, I have a keyboard fault. The Q key seems to be electrically stuck. I've never had an issue with this keyboard, and all of a sudden, the Q key is stuck. WTF? Scott

Thanks for the replies everyone. Like reifsnyderb noticed, it appears that I tapped into the system clock on the GTIA and connected it to the 74ls258 that was piggybacked on chip cd4050 which is u20. And like it was mentioned, u20 handles video processing. I still have no memory of doing this, but since no one seems to know what it is either, I should probably remove it too. Maybe it was something that I read in Antic magazine. At some point I'd like to install the ultimate 1mb. I assume this would mean I'd have to remove the 256kxl anyway? (I'm guessing they don't coexist) Scott

Hi there everyone! Got my first 800xl new in '85, and I'm trying to revive it. Seems that it has a memory related problem, but I'll save that for another thread. I installed a Newell 256kxl way back in '86, and in the interest of diagnosis, I'm going to remove it, but I found that I had done another mod that I don't remember doing and I don't know what it's for. I don't think its related to the 256kxl install as I've re-read the install instructions, and there's no mention of this. It concerns chip u20. It seems that I took the 74ls258 that is removed from u27 for the 256kxl, and piggy backed it on u20, utilizing only a few of the pins, and attached a bodge wire to the GTIA. any idea what this is for? I'm attaching pics of this. Also, on a different subject, can I just remove the 256kxl, reinstall a 74ls258 in u27, reinstall the 3k resistor in r32, and keep the 256k ram chips (I don't have the 64k chips anymore), and they'll just function like the 64k chips? Thanks in advance! Scott