1200XL - boots to red screen

[+bob1200xl]

It looks like a memory problem - could be the delay line. I have seen one or two fail.

 

Look at the 800XL schematic and compare it to the 1200XL drawings. Compare functions, not pin-outs. They are very similar. You can see how the 1200XL does the DRAM clocks...

 

 

Bob

[+DrVenkman]

Thanks, Bob. I’ll do that tonight. Appreciate the tip!

[+DrVenkman]

Still experimenting with the logic probe and learning my way around the 74LSxxx logic on the board ... man, I wish there was a Sam's Computerfacts for the 1200XL! The one for the 800XL seems pretty damn excellent. Anyway, so far as I can tell, everything on the board is generally doing what it's supposed to so and just to be sure, I swapped all the logic one by one again from working donor systems but no change. The only thing I didn't swap is the C060472 "XL Delay" component, as it's soldered to the board. However again, the logic probe shows it's putting out the signals it's supposed to, at least at a coarse, high-level basis. I don't have a 'scope so I can't check pulse frequencies or anything of that nature. To rule it out, unless I solve the problem before then, I will add it to my list of stuff to order from Best in a couple weeks.

 

To address Todd's suggestion of last night, I did try swapping all 8 DRAMs from a donor system but again, no difference. Per Jon's suggestion this morning, I tried flexing the board while it was running but it made no difference. I even tried wiggling the sockets and the chips within the sockets, pressing down on discrete components around the board except for the power components. No change. I will do a bit more testing tomorrow or later in the weeks with the logic probe once I've gotten a better understanding of the schematics - more particularly, I'm going to open up my working machine and try to spot any differences in behavior from one to the other with the probe, specifically around the support logic.

 

I do note a new difference from a few days ago - now, almost invariably, when I boot up the machine, there's a bit of graphical distortion in the display. Specifically, a few missing, dark pixels in the logo. Seems more and more likely to be an ANTIC or GTIA problem but I swapped the ANTIC again tonight with no change and I've previously swapped GTIA as well.

 

The last thing I can think of is some kind of foreign object or debris under a socket between two pins; I've mentioned in the past that this machine was FILTHY when I got it. Unfortunately, several of the sockets are the "solid" style which cover the board underneath. That's good for removing the chips for replacement, but you can't see very well around the edges where the plastic might be slight raised off the board, not completely flush. It's an outside shot but at this point it makes as much sense as almost anything else.

[+kheller2]

Have you replaced all the chips from the donor unit at once? I mean every chip so that basically the flakey unit has every chip replaced from a working unit at the same time? Beside the OS ROMs and maybe delay/mmu everything else should be swappable with an 800xl etc.

[+DrVenkman]

Have you replaced all the chips from the donor unit at once? I mean every chip so that basically the flakey unit has every chip replaced from a working unit at the same time? Beside the OS ROMs and maybe delay/mmu everything else should be swappable with an 800xl etc.

Not yet. My donor systems are frankly looking pretty barren at this point, used to repair others over the years. This particular 1200XL has the CPU from one of my donor XL machines, as it wouldn’t boot at all with its original. I can replace POKEY, ANTIC and the PIA from donor systems (plural) but I’m out of spare GTIA chips - I’d have to pull from an otherwise-complete and working system, so I’m hesitant to do that. I’m frankly not sure it would make a difference. Something is screwy and I don’t know what it is. I’ve been over it front and back until I’m blind.

 

I’ve got a couple things to check tonight and then I’m going to set it aside for a few days and see if a fresh look makes a difference.

 

In the meantime I’m going to dive into the technical documentation out there and try to determine exactly what happens when the HELP key is pressed and what might trigger that crash. I know console switches are read by the GTIA but that’s the extent of my current knowledge.

[flashjazzcat]

The Help key is handled by Pokey like the rest of the standard keys.

[+DrVenkman]

The Help key is handled by Pokey like the rest of the standard keys.

Ah, so it is. I was conflating it along with START, SELECT and OPTION. All the more reason I need to spend time with the documentation to figure out what’s going on here.

[mellis]

Still experimenting with the logic probe and learning my way around the 74LSxxx logic on the board ... man, I wish there was a Sam's Computerfacts for the 1200XL! The one for the 800XL seems pretty damn excellent. Anyway, so far as I can tell, everything on the board is generally doing what it's supposed to so and just to be sure, I swapped all the logic one by one again from working donor systems but no change. The only thing I didn't swap is the C060472 "XL Delay" component, as it's soldered to the board. However again, the logic probe shows it's putting out the signals it's supposed to, at least at a coarse, high-level basis. I don't have a 'scope so I can't check pulse frequencies or anything of that nature. To rule it out, unless I solve the problem before then, I will add it to my list of stuff to order from Best in a couple weeks.

 

To address Todd's suggestion of last night, I did try swapping all 8 DRAMs from a donor system but again, no difference. Per Jon's suggestion this morning, I tried flexing the board while it was running but it made no difference. I even tried wiggling the sockets and the chips within the sockets, pressing down on discrete components around the board except for the power components. No change. I will do a bit more testing tomorrow or later in the weeks with the logic probe once I've gotten a better understanding of the schematics - more particularly, I'm going to open up my working machine and try to spot any differences in behavior from one to the other with the probe, specifically around the support logic.

 

I do note a new difference from a few days ago - now, almost invariably, when I boot up the machine, there's a bit of graphical distortion in the display. Specifically, a few missing, dark pixels in the logo. Seems more and more likely to be an ANTIC or GTIA problem but I swapped the ANTIC again tonight with no change and I've previously swapped GTIA as well.

 

The last thing I can think of is some kind of foreign object or debris under a socket between two pins; I've mentioned in the past that this machine was FILTHY when I got it. Unfortunately, several of the sockets are the "solid" style which cover the board underneath. That's good for removing the chips for replacement, but you can't see very well around the edges where the plastic might be slight raised off the board, not completely flush. It's an outside shot but at this point it makes as much sense as almost anything else.

 

 

You have done a thorough job troubleshooting so far.

 

This reminds me of a troubleshooting session I had with a malfunctioning Roomba (yes, the robotic vacuum). After extensive troubleshooting, I decided that there had to be some undetectable conductive material messing with the logic. Figuring I had nothing to lose as the board was going to be trash anyway, I ran it through a dishwasher cycle. Here's exactly what I did:

1. Emptied the dishwasher

2. Loaded just the circuit board into the top rack (glassware rack)

3. Ran a no-heat-dry cycle with normal detergent

 

After the cycle, I bathed the board in isopropyl alcohol to displace any water. I then let the board air dry for two days.

 

After all of that, I have the Roomba working flawlessly.

 

I am not promising this will work for you, but being as the board was filthy, it might be the only way to eliminate the hidden conductive material.

 

Good luck!

Edited June 28, 2017 by mellis

[+DrVenkman]

In lieu of running the maching through the dishwasher and incurring the wrath of MrsVenkman ( ) I flushed under and around each socket copiously with isopropyl alcohol using a syringe bottle. It flushsed some little bits of dust and fluff out but little else; I sprayed eveything out with canned air, let it dry and then tried to boot the machine again. No change. *sigh* Just for giggles I tried swapping first ANTIC, then GTIA and finally POKEY, given Jon's helpful reminder that POKEY handles the HELP key. No change until after swapping POKEY ... then the machine wouldn't boot at all again, even after swapping back. Grrr. I put the board aside and found more productive ways to spend my time (like hitting my head with a hammer, counting carpet fibers, watching the yard grow ...)

Tonight I came back to it. I started removing chips and examining sockets themselves very closely. First thing I noted is that the lower OS socket grips the chips relatively loosely, at least compared to the upper socket and the rest of the main IC sockets. The wipes were all nice and shiny from the cleaning I gave them a couple weekends back. I examined each socket closely in turn and the rest of them all looked fine. And then I got to POKEY ...

Frankly, in the past few weeks I've paid POKEY very little attention. When the machine first wouldn't boot, it was the 6502 that was the culprit and then that socket had a broken wipe from all the manipulations during troubleshooting, so I had to replace it. Then the machine would boot but crash, so I focused on memory, support logic, ANTIC ... hardly a thought to POKEY again.

So tonight, well, I'll let the video tell the tale.

 

[RodLightning]

Nice. Now you can decide how far to go in making things right. Putting a new precision socket in for Pokey would make the problem go away for good. I would be torn between that or just trying to bodge the existing faulty pin back into position.

 

Thanks for making the video. Very cool.

[+DrVenkman]

Nice. Now you can decide how far to go in making things right. Putting a new precision socket in for Pokey would make the problem go away for good. I would be torn between that or just trying to bodge the existing faulty pin back into position.

 

Thanks for making the video. Very cool.

 

Oh, no question ... I'm going to replace the socket. Absolutely no question about that.

 

Tonight, as an experiment I inserted a tinned wire into the socket hole, gently bent the leg of POKEY Pin 7 out enough to clear the socket and reinserted the chip. Then I made a sharp crease in the other end of the wire and lightly crimped it around the exposed leg of the chip. The machine boots again now but the problem of crashing on attempting to enter Self Test persists. Tomorrow afternoon I'll replace the socket and then we'll see how things go from there. Things are still awry and I need to have the machine plugged in and running while I run more checks with the logic probe.

[+DrVenkman]

I replaced the POKEY socket on my troublesome machine today and while the machine still freezes rather than going to Self Test, at least it boots up. I then put in a Joust cartridge since it has an attract mode and there will be lots of data moving around the system. I probed the main ICs with a logic probe and here's what I measured:

 

POKEY

1. Vss - Low
2. D3 - High/Low pulse (slow)
3. D4 - High/Low pulse (slow)
4. D5 - High/Low pulse (slow)
5. D6 - High/Low pulse (slow)
6. D7 - High/Low pulse (slow)
7. Phi2 - High/Low pulse (slow)
8. P6 - Low
9. P7 - Low
10. P4 - Low
11. P5 - Low
12. P2 - Low
13. P3 - Low
14. P0 - Low
15. P1 - Low
16. KR2 - High
17. Vcc - High
18. K5 - High/Low pulse (fast)
19. K4 - High/Low pulse (fast)
20. K3 - High/Low pulse (slow)
21. K2 - High/Low pulse (slow)
22. K1 - High/Low pulse (slow)
23. K0 - High/Low pulse (slow)
24. SID - High
25. KR1 - High
26. BCLK - High
27. ACLK - High
28. SOD - High
29. IRQ - High
30. CS0 - High pulse (fast)
31. CS1 - High
32. R/W - High pulse (fast)
33. A3 - High/Low pulse (slow)
34. A2 - High/Low pulse (slow)
35. A1 - High/Low pulse (slow)
36. A0 - High/Low pulse (slow)
37. AUD - High
38. D0 - High/Low pulse (slow)
39. D1 - High/Low pulse (slow)
40. D2 - High/Low pulse (slow)

6502C

1. Vss - Low
2. RDY - High
3. Phi1 - Low
4. IRQ - High
5. N/C
6. NMI - High pulse
7. SYNC - High/Low pulse (slow)
8. Vcc - High
9. A0 - High/Low pulse (fast)
10. A1 - High/Low pulse (slow)
11. A2 - High/Low pulse (slow)
12. A3 - High/Low pulse (slow)
13. A4 - High/Low pulse (slow)
14. A5 - High/Low pulse (slow)
15. A6 - High/Low pulse (slow)
16. A7 - High/Low pulse (slow)
17. A8 - High/Low pulse (slow)
18. A9 - High/Low pulse (slow)
19. A10 - High/Low pulse (slow)
20. A11 - High/Low pulse (slow)
21. Vss - Low
22. A12 - High/Low pulse (slow)
23. A13 - High/Low pulse (slow)
24. A14 - High/Low pulse (slow)
25. A15 - High/Low pulse (slow)
26. D7 - High/Low pulse (slow)
27. D6 - High/Low pulse (slow)
28. D5 - High/Low pulse (slow)
29. D4 - High/Low pulse (slow)
30. D3 - High/Low pulse (slow)
31. D2 - High/Low pulse (slow)
32. D2 - High/Low pulse (slow)
33. D0 - High/Low pulse (slow)
34. N/C
35. HALT - High/Low pulse (slow)
36. R/W - High pulse (fast)
37. Phi0 - High/Low pulse (slow)
38. S0 - Low
39. Phi2 - High/Low pulse (slow)
40. RST - High

6520 PIA

1. Vss - Low
2. PA0 - Low
3. PA1 - Low
4. PA2 - Low
5. PA3 - Low
6. PA4 - Low
7. PA5 - Low
8. PA6 - Low
9. PA7 - Low
10. PB0 - High
11. PB1 - Low(*)
12. PB2 - High
13. PB3 - High
14. PB4 - Low(*)
15. PB5 - Low(*)
16. PB6 - Low(*)
17. PB7 - High
18. CB1 - High
19. CB2 - High
20. Vcc - Low (***)
21. R/W - High pulse (fast)
22. CS0 - High
23. CS2 - High pulse (slow)
24. CS1 - High
25. Phi2 - High/Low pulse (slow)
26. D7 - High/Low pulse (slow)
27. D6 - High/Low pulse (slow)
28. D5 - High/Low pulse (slow)
29. D4 - High/Low pulse (slow)
30. D3 - High/Low pulse (slow)
31. D2 - High/Low pulse (slow)
32. D1 - High/Low pulse (slow)
33. D0 - High/Low pulse (slow)
34. RST - High
35. A0 - High/Low pulse (slow)
36. A1 - High/Low pulse (slow)
37. IRQB - High/Low pulse (slow)
38. IRQA - High
39. CA2 - NO SIGNAL (**)
40. CA1 - High

(*) Signal was very, very weakly low, as if there wasn't quite enough to trigger a logic 0. According to the 800XL schematics in seen in Sam's Computerfacts, I think these lines are unconnected and should be High (logic 1). Not sure if the 1200XL is identical in this regard.

(**) This seems worrisome. All the analogous schematics I've seen for other machines show this as being High (logic 1). Unfortunately, my multimeter picked today to go kaput (cheap leads pulled right out of the probe and no way solder them back together. So I don't really know where to start checking this. Schematics show that the next component up the line is R52, and I get a Logic High at the positive end of that component, but no signal at all on the side that leads back to the PIA. I don't know if that's what's expected or not.

(***) This is very troublesome; it means there's a short somewhere; it's supposed to be +5V (Logic High). Schematics show it's isolated from Ground by C60. Assuming I can verify continuity from +5V to the pin, is it possible the capacitor has failed closed and shorting the line? Thanks for any pointers here.

ANTIC

1. Vss - Low
2. AN0 - High/Low pulse (fast)
3. AN1 - High/Low pulse (slow)
4. LP - High
5. AN2 - High/Low pulse (fast)
6. RNMI - High
7. NMI - High pulse
8. REF - High pulse
9. HALT - High/Low pulse (slow)
10. A3 - High/Low pulse (slow)
11. A2 - High/Low pulse (slow)
12. A1 - High/Low pulse (slow)
13. A0 - High/Low pulse (slow)
14. R/W - High pulse
15. RDY - High/Low pulse (slow)
16. A10 - High/Low pulse (slow)
17. A11 - High/Low pulse (slow)
18. A13 - High/Low pulse (slow)
19. A14 - High/Low pulse (slow)
20. A15 - High/Low pulse (slow)
21. Vcc - High
22. A11 - High/Low pulse (slow)
23. A9 - High/Low pulse (slow)
24. A8 - High/Low pulse (slow)
25. A7 - High/Low pulse (slow)
26. A6 - High/Low pulse (slow)
27. A5 - High/Low pulse (slow)
28. A4 - High/Low pulse (slow)
29. Phi2 - High/Low pulse (slow)
30. D0 - High/Low pulse (slow)
31. D1 - High/Low pulse (slow)
32. D2 - High/Low pulse (slow)
33. D3 - High/Low pulse (slow)
34. Phi0 - High/Low pulse (slow)
35. FPhi0- High/Low pulse (slow)
36. RST - High
37. D7 - High/Low pulse (slow)
38. D6 - High/Low pulse (slow)
39. D5 - High/Low pulse (slow)
40. D4 - High/Low pulse (slow)

GTIA

1. A1 - High/Low pulse (slow)
2. A0 - High/Low pulse (slow)
3. Vss - Low
4. D3 - High/Low pulse (slow)
5. D2 - High/Low pulse (slow)
6. D1 - High/Low pulse (slow)
7. D0 - High/Low pulse (slow)
8. T0 - High
9. T1 - High
10. T2 - High
11. T3 - High
12. S0 - High
13. S1 - High
14. S2 - High
15. S3 - Low
16. N/C
17. DEL - High
18. AN0 - High/Low pulse (fast)
19. AN1 - High/Low pulse (fast)
20. AN2 - High/Low pulse (fast)
21. COL - High pulse
22. L1 - High/Low pulse (fast)
23. L2 - High/Low pulse (fast)
24. L3 - High/Low pulse (fast)
25. CSYNC- High pulse
26. HALT - High/Low pulse (slow)
27. Vdd - High
28. OSC - Low pulse
29. FPhi0- High/Low pulse (slow)
30. Phi2 - High/Low pulse (slow)
31. L0 - Low
32. CS - High pulse
33. R/W - High pulse
34. D7 - High/Low pulse (slow)
35. D6 - High/Low pulse (slow)
36. D5 - High/Low pulse (slow)
37. D4 - High/Low pulse (slow)
38. A4 - High/Low pulse (slow)
39. A3 - High/Low pulse (slow)
40. A2 - High/Low pulse (slow)

[+DrVenkman]

... it looks like a busted socket or broken trace on the board ...

 

Jon, you're a genius. I already discovered a bad 6502 socket, then a bad POKEY socket the other day which I replaced this morning. After fixing that, more testing with the logic probe and ... well, once again, we'll let the video show you.

 

 

EUREKA!!!

 

P.S. I don't know who had this machine before me, or what they did to abuse it, but I've never had a machine with THREE messed up sockets, plus the busted joystick port (which I'll fix sometime down the line; it's only Player 2.

[+Stephen]

Those single wipe sockets can sometimes only be good for one or two insertions (insert Beavis giggle).

[flashjazzcat]

Well done. These faults are sometimes very hard to track down (even once we know it has to be a fault on the board), and your troubleshooting was nothing if not exhaustive!

[+DrVenkman]

It gives me great satisfaction to run these issues to ground and get them figured out.

I was searching the forum for info about some of this stuff a few nights ago and ran into my own post from 3 years ago when I pulled these machines from storage. As it happens, this particular board with the broken joystick port and 30 years worth of grime was actually inside the "pretty" case (complete with a "Property of Mendocino Schools" inventory tag. ) It had the same problems then as it did until last night - it would boot but eventually crash, some messed up graphical glitches, and wouldn't go to self-test. The other board was originally in the yellowed case with the cracked acrylic, etc. It would boot but had odd graphical issues. I fixed those by salvaging POKEY and ANTIC from a donor 600XL. I swapped that board into the nicer case and dubbed it "Big Sexy." That was the one that caused me to start this thread a few weeks ago when it suddenly wouldn't boot.

THIS board, now living in the Ugly Duckling case, is coming back to life. The case has been cleaned up and thorougly retrobritened with peroxide, I have a piece of replacement acrylic from Jon to go over the LEDs and it all appears to boot and run normally again (or will, once I get that 6520 socket replaced). A joystick port repair will follow in a couple weeks once I get some replacements.

[flashjazzcat]

However you go about it, there is no feeling quite like getting a dead or problem board fully working again. Especially when it's a 1200XL.

[toddtmw]

Well done. These faults are sometimes very hard to track down (even once we know it has to be a fault on the board), and your troubleshooting was nothing if not exhaustive!

I've been following this from the beginning, I know I'M exhausted! ?

[+DrVenkman]

I've been following this from the beginning, I know I'M exhausted!

 

Yeah, me too!

 

I wanted to document the process as I went along, including all the missteps, false alarms and dead ends I went down. This site's archives are a treasure trove for maintaining these old machines and tracking down mysterious issues. It's instructive to read old posts and follow peoples' trains of thought as they work through the problems. Hopefully this topic and my wrong steps and issues will help others figure out similar problems in the future.

[ACML]

Makes you wonder what could have caused three sockets to have issues. You said the case it came with was cosmetically in good shape, but was there evidence that someone had opened it up and tinkered with it? You know, missing a few shield rivets, oxidized finger prints on the RF shielding or did it look untouched? Hope it all goes back together Code 1. I too share the feeling of accomplishment you get when you have a tricky issue that finally gets resolved.

 

Teddy Roosevelt; "Any work worth doing, is worth doing well".

Edited July 2, 2017 by ACML

[+DrVenkman]

Makes you wonder what could have caused three sockets to have issues. You said the case it came with was cosmetically in good shape, but was there evidence that someone had opened it up and tinkered with it? You know, missing a few shield rivets, oxidized finger prints on the RF shielding or did it look untouched? Hope it all goes back together Code 1. I too share the feeling of accomplishment you get when you have a tricky issue that finally gets resolved.

 

Teddy Roosevelt; "Any work worth doing, is worth doing well".

 

The case was in good shape. The shielding looked somewhat corroded though not terrible; I think a prior owner in the 80's or 90's had stored it somewhere with a lot of humidity, maybe in a basement or barn. There was lots of crud inside - animal fur (cat/dog hair? mouse hair? *shrug*) and general dirt. Many of the original chips have visible oxidation on the leg. One thing I did before I got the logic probe was remove the chips, clean the existing socket wipes with alcohol and the legs with a "magic eraser" type sponge to get them clean and ensure good contact. I didn't examine each socket under magnification until I got the logic probe and realized some of the signals on the board were missing or intermittent. In retrospect, when the wipe broke on the CPU socket, it should've been my cue to remove every chip on the board and look at each wipe carefully.

The good news is, the sockets for everything else look pretty good all things considered. The wipes all look to have good spring tension, no corrosion visible at the top bend (where the break was on the 6502 socket). All of this socket corrosion seems confined to left of center on the board, by and large. When I removed the existing socket for the POKEY yesterday, many of the socket legs just broke right off when I started to remove them with tweezers from the molten solder. Similarly, today when I pulled the 6520 socket, 5 or 6 broke as well. Fortunately, the rest came out intact cleanly, which makes it much easier to clean out the holes for the new one.

The board itself cleaned up pretty well. Once I've got the new 6520 socket in and have verified all is working well, I'm going to clean it thoroughly again and then reassemble it into the case. I'm debating whether to install my now-surplus Rev C UAV board from Bryan before putting it all back togher; stock video on these is really bad.

[+DrVenkman]

And done - UAV installed, +5V restored to the SIO port, case reassembled (thanks again, Jon!). The Ugly Duckling is gone replaced by The Swan.

 

 

And here they both are - thanks again to all of you who offered encouragement, suggestions and advice. It's been fun!

 

[toddtmw]

Fun?

 

Glad you thought so. I'm jealous. Enjoy the 1200's!

 

Todd

[Kyle22]

Did you ever de-solder, socket and replace the C060472 delay chip? They don't fail too often, but they can and do fail.

[+DrVenkman]

Did you ever de-solder, socket and replace the C060472 delay chip? They don't fail too often, but they can and do fail.

Nope. As it turns out once I did a more careful and thorough investigation with the logic probe, I uncovered first the bad socket wipe at Pin 7 of POKEY and after replacing that, found that Pin 20 of the 6520 socket was bad as well. In my initial tests before I got the hang of the logic probe, I was sort of pushing the probe against the shoulder or sides of the chip legs, ascribing poor signal as just surface corrosion or something. Later I realized that I was actually facilitating contact with a bad socket wipe by pressing against the chip leg. So when I measured everyhing a second time, I discovered that the wipes at the suspect points had lost whatever spring tension had once been in the metal and were no longer making adequate contact with the chip.

 

After I replaced the 6520 socket, the machine booted right up, went into self-test correctly with the HELP key, and passes all self-tests and all applicable tests on the SALT 2.05 and SuperSALT carts. I think it’s probably in pretty good shape now. I can’t say there’s not some other, even more subtle problem not easily uncovered but for now, I can’t locate anything else. It’s certainly in better shape than at any time since I acquired it in the early 00’s.