1200XL - boots to red screen

[+DrVenkman]

So here's a head-scratcher ... I have two 1200XL's, one in fairly good physical condition, one "ugly duckling" with a horribly yellowed case, broken plasic over the LEDs, one joystick port with a broken pin. Both were working when stored away. A couple years ago, I fitted a U1MB inside the prettier one. Worked great at first but within a day or so started showing instability - settings would reset themselves in the U1MB, occasional garbled display in the U1MB menu, etc. I removed the U1MB, installed that board inside one of my 800XL's and it's worked great since.

 

Today at long last I decided to return that 1200XL to stock configuration - I reinstalled the original MMU, OS ROMS and replaced the jumpers back to their original locations (W7, W8 and W9). I hadn't been using my 9 VAC bricks for awhile so I plugged one in, checked it to be sure it was good (it shows 10.2 VAC without any load) then plugged it the machine and powered up. Red screen. If there's any audio pops I didn't hear them.

 

I then dug out the Ugly Duckling, powered it up and it worked fine - rainbow logo and all. So I started trying to isolate any bad chips. Since it couldn't seem to get running, I decided to see if the CPU was bad first. I removed the CPU from the bad board, plugged it into the Ugly Duckling and gave it a try - now the Ugly Duckling had a red screen too. So to check, I tried the known-good CPU in the first board. Still had a red screen. Huh. So I put the good CPU back into the Ugly Duckling where it came from and now it boots to a red screen too. Hmm. Both machines now boot to a red screen.

 

Anyone have any suggestions?

[toddtmw]

This seems to be a scattergun trouble-shooting guide, but it might be helpful:

 

http://atari.boards.net/thread/974/red-screen-problems-fix-checklist

[Kyle22]

When you reverted that 1200XL back to stock, did you remove the jumper wire(s) that you added for the mod?

[+DrVenkman]

When you reverted that 1200XL back to stock, did you remove the jumper wire(s) that you added for the mod?

 

I did. It was the first thing I did. They were just spot-soldered to vias anyway. I lifted them off the board, reinstalled the 1 ohm jumpers at W7, 8 and 9, then reinstalled the MMU and OS roms. Nada.

 

My real concerns are that now that I've swapped CPUs, and then swapped them back, neither one of the boards is working. Dubya. Tee. Eff. I'm thinking I should touch up the socket pins on both machines just on general principles but that seems long a long shot.

[Kyle22]

Check your voltage regulators on both boards. Something made that CPU go bad.

[+DrVenkman]

Check your voltage regulators on both boards. Something made that CPU go bad.

 

Okay, checked the rectifier - it's putting out 12.04V. The two 7805's are reading 4.98V and 5.01V. I don't see anything wrong there, unfortunately. I'm about done in tonight. I will have to spend more time at it tomorrow, but I am honestly at a bit of a loss.

[Kyle22]

Yeah, that's strange. Voltages are well within spec.

[AtariGeezer]

Are the CPU's getting hotter than before?

[ACML]

 

I lifted them off the board, reinstalled the 1 ohm jumpers at W7, 8 and 9, then reinstalled the MMU and OS roms. Nada.

 

 

You put back W7, 8 and 9, but what about W6? That has to be shorted for the stock 24 pin ROMs as well.

[+DrVenkman]

Are the CPU's getting hotter than before?

 

Nope. None of the chips are particularly hot, only the MMU's are even more than warm.

 

You put back W7, 8 and 9, but what about W6? That has to be shorted for the stock 24 pin ROMs as well.

 

D'oh! I totally forgot about that one! I'll replace it later this morning.

 

That still doesn't explain what might be going on with the SECOND board, which is still in stock configuration, that now won't boot with its original CPU. But I will worry about that one later!

[+DrVenkman]

Okay so after reinstalling W6 I get to the rainbow logo but oddly it hangs when I insert Star Raiders. I'm running the self-test now to see if I can figure out what else might be going on.

[Rybags]

That's a diag mode 8K cart that only uses the first 8K of Ram, so is a game that will work when many others wouldn't.

[+DrVenkman]

That's a diag mode 8K cart that only uses the first 8K of Ram, so is a game that will work when many others wouldn't.

 

Yep, I know. That's why I was trying it. So I connected a keyboard and ran the self-test. Everything seemed to pass. So I tried Star Raiders again - this time, with the keyboard attached, it works! I tried a slew of other carts and they all booted up normally. I didn't want to jinx anything but since I had the board out on the bench, and I had a second rev D UAV board from Bryan that I've been holding onto for a couple weeks, I went ahead and installed that too. Wow! Looks great except for some RF interference, probably because I haven't yet disconnected and removed the RF modulator. I'll take care of the disconnecting part with my snips after lunch!

 

So I'm going to declare victory and retire from the field for now. Still can't figure out WTF could possibly have happened to my second Ugly Duckling 1200XL. It's sitting there in bone-stock original configuration and now boots to a red screen too. I must have some cold solder joints on one of those ICs I pulled and reseated yesterday. Anyway, for now, my main 1200XL is back in operation! I'm going to buy another U1MB in a few weeks and install it inside this one, as I did a couple years ago, but this time I'll try little capacitor mod to the U1MB and/or swap the 74SL08 chip with one from a donor machine.

 

Anyway, thanks for everyone's suggestions, and if anyone else has any wild ideas why my Ugly Duckling is currently comatose, I'd appreciate it.

[ACML]

If it continues to hang on Star Raiders, its most likely RAM timing. It will pass all tests, but hangs on game. The solution is to swap out RAM chips one by one and test with games that locks it up. Usually its the first RAM chip on the left that is the most picky timing wise.

Edited May 30, 2017 by ACML

[ACML]

on your other board, try reseating your two ROM OS chips.

[+DrVenkman]

on your other board, try reseating your two ROM OS chips.

 

So to update my progress on the Ugly Duckling ...

 

Tonight I had 30 minutes so I removed and reseated all the main socketed ICs (I didn't reseat the RAM or the miscellaneous logic chips) but that had no effect. Next I checked the outputs of power system - the 12V rectifier was putting out 11.98V and the two 7805's were putting out 4.98V and 4.97V, respectively.

 

So next I left the system running for about 20 minutes and came back. The ANTIC chip and 6502 were a bit warmer than ambient but not much. The rest of the chips were fairly cool including the RAM. But the MMU was darn near scorching. That doesn't seem normal to me. I guess it's time to track down a 1200XL MMU and see how it goes.

[+bob1200xl]

Track one down? I have a ton of the 09 chips. PM me.

 

They do run rather warm, however.

 

Bob

[+DrVenkman]

Track one down? I have a ton of the 09 chips. PM me.

 

They do run rather warm, however.

 

Bob

 

Thanks, Bob - I messaged you.

[+DrVenkman]

Alright - update time.

 

Machine #2: Ugly Duckling 1200XL:

 

I do *NOT* need a new MMU for the bad machine. I had my "good" unit open to clean up some wiring on the UAV video mod I put in last week and while I was at it, I swapped the MMU from the bad machine into the good one, and it booted right up. Huh. So then I swapped all the 74-series logic chips and several RAM chips selected at random into the good machine and they all worked fine there. I had already swapped the 6502 into another machine and it worked like a charm there. Those are the main chips associated with booting up, at least without anything on the SIO bus, so I don't know what else to try without a logic probe or scope. Anyone have suggestions of components along the boot-up/power-on sequence to test or replace?

 

 

Machine #1: "Pretty" 1200XL:

 

So after a few days where I thought all was fine, the "good" machine is showing signs of system noise and instability. Through both Composite and to a lesser extent through S-video, typing on the screen produces faint horizontal system noise lines in the video display. The system Bell sound makes several faint broad horizontal noise "bars" while it chimes with a BASIC error message.

 

However, if I turn off the keyclick sounds (the 1200XL OS allows this with one of the function key combos), typing produces no screen noise. Also interesting, disk activity sounds from the SIO bus do not produce system noise on the screen either - only keyclicks. I haven't yet tried any other sounds - I might right a quick BASIC tone generator or find a music program to see if it's ALL system sounds, or just the keyclicks.

 

The keyclick system noise is enough, in fact, to get the system to reset and/or lockup if I type very quickly and start filling the screen. I thought that might indicate a bad POKEY, so I've swapped a presumable good POKEY into the system. The typing nose is still there but - so far - no resets or lockups. Might this be a PIA problem instead?

 

Any and all suggestions welcomed!

 

EDITED TO ADD: Update again - just allowing the system to sit idling for 15 - 20 minutes and now it's locked up. So my "good" machine has issues too, it seems.

[_The Doctor__]

but you just told us you moved chips from bad ugly to pretty good machine... maybe you swap them back and it's normal again...?

[+DrVenkman]

but you just told us you moved chips from bad ugly to pretty good machine... maybe you swap them back and it's normal again...?

 

I did that to see if the Ugly Duckling had any obviously-failed chips preventing it from booting. After swapping them one by one, I moved them back.

 

Where I stand now is the Ugly Duckling won't boot - still goes to a red screen on power up. The good machine boots up and appears to run fine for about 20 minutes. Then it reboots, sometimes just while idling, and eventually locks up with no video output. Power cycling doesn't work; I can then uplug it, power cycle it while unplugged (drain the big 10,000 uF cap, I guess) and then it will power up but still be glitchy. On this machine I'm leaning towards maybe bad power regulators that are failing as they heat up (which I will test later today) or an IC that is failing as it heats up.

 

Does that seem logical to anyone else? Any other suggestions?

[+bob1200xl]

C7 is the RESET timing capacitor. They go bad sometimes, giving you boot problems and quirky failures while the system is just sitting there. Measure the voltage on the '+' side - it should slowly climb (5 seconds) to 5v and stay there.

 

C7 is about 1 inch below pin 1 of the 6520 PIA.

 

Bob

[+DrVenkman]

Thanks Bob! The field service manual says that's a 47 uF electrolytic. I assume that's correct - my two machines have caps like that in several locations around the board that don't have a capacitance rating, only a part number and max temperature. I'll check the behavior of that cap later tonight or tomorrow and if that's the issue, I'll be ordering new caps next weekend. I probably need to anyway, those 47 uF and 22 uF electrolytics are all over these boards. I probably should have spares on hand.

 

As for my "pretty" machine, the one I got up and running last weekend again, I think I solved the problem. First, some backstory ...

 

All week it seemed to work great when I used it in the evenings, but I never had it powered on longer than 20 - 30 minutes, until Thursday night, when I got called away and it was running for a couple hours. When I returned to the machine, there was no video output and it was unresponsive. I was able to power cycle it and it booted up and worked normally. Then Friday night I was fooling around with several systems (I have an A8 machine, a 2600 and a 7800 all connected to the same CRT) - when I returned to the 1200XL it was dark again. This time a power cycle didn't reboot it, but I could pull the plug, toggle the power switch to drain the onboard caps quickly, and then connect the power and boot it. However, it reset it self while I was typing. So I started running tests which carried over into yesterday much of the afternoon. I ended up being able to type in a simple BASIC infinite loop program and induce a crash within anywhere from 5 - 15 minutes or so with just:

 

10 ? "Testing 1, 2, 3 ..."

20 GOTO 10

 

and then just let it run. Eventually the screen would glitch and ultimately crash, or it would cold boot and eventually hang with no video. Similarly, I found I could induce a crash by having BASIC inserted and letting a key auto-repeat long enough to trigger the BASIC line input error bell a few times.

 

So today I sat down with a donor 800XL I have, and started swapping chips. In order of likeliness, I tried first the CPU, then ANTIC, then POKEY (because I could induce a crash via the keyboard), but in each case the donor machine worked fine with the suspect chip and the 1200XL would still crash after a few minutes. So in desperation I tried PIA (what the hell, it's sitting right there in the middle of the 1200XL board) and finally GTIA, which in retrospect I should have started with. Yes, GTIA was the culprit. The system ran the infinite loop test for 40 minutes or so with no glitches or errors, accepted multiple keyboard buffer error bells in BASIC without a problem. Even more tellingly, in the 800XL, that GTIA induced the same sort of crash/no video after about 5 minutes of run-time of my infinite loop BASIC test.

 

So, tl; dr: if you're having weird screen glitches and reboots/hangs, replace the GTIA!

[DavidMil]

Just curious... Was the GTIA chip getting hot to the touch?

 

DavidMil

[+DrVenkman]

Just curious... Was the GTIA chip getting hot to the touch?

 

DavidMil

 

Nope, not really. The 6502 was pretty warm, the POKEY and ANTIC somewhat warm. The MMU was pretty scorching but as Bob says, I guess they seem to run hot. The GTIA was about as cool as the PIA, RAM or any of the random 74xxx logic chips.

 

And now after replacing the GTIA, the machine has been running over 6 hours straight without a problem.