Mike's PCjr Page

I've recently been noticing flaky behavior from my PCjr and after playing around with the system tonight I'm fairly sure something is up with it.

Symptoms:
Certain games seem to get stuck when they switch video modes:
1) Thexder - original disks - system asks to insert the original disk....even though they are the original disks and are still in very good shape looking at a Kryoflux dump. Trying to get past this eventually just sits at a blank screen and doesn't do anything. Ctrl+Alt+Del sometimes works.
2) Thexder - non-copy protected copy - game attempts to load and then switches resolutions and displays a single white pixel on the screen. Then nothing. Ctrl+Alt+Del successfully reboots.
3) PCjr booter King Quest game - does one of two things, either keeps seeking over the same tracks over and over again on the disk, or gets further and eventually shows a few random white pixels on the screen, or random characters. Then nothing. Ctrl+Alt+Del successfully reboots.
4) Jumpman - tested because it is a different 16 color booter game...this one seems to work just fine.

I tested Thexder and Kings Quest with the following system configurations:
Base system + internal 64k + parallel port side car + jrIDE = see above
Base system + internal 64k + jrIDE = see above
Base system + jrIDE = system seems to get stuck in a reboot loop when it checks the first little bit of memory, repeats cycle over and over.
Base system = see above

Looking at the motherboard, the only thing that I noticed that was potential off was very slight corrosion on one side of a diode near the 10 uF capacitor on the left/center side of the main board. I'm not sure what this cap is for and have not stuck a tester on it to see if it might be having issues.

Any ideas for what to test or try to get more info on what could be wrong? I'm somewhat suspecting video issue, but the slightly weird drive behavior on occasion has me a little confused as well. I have the ATX power converter mod, so power should be good and seems more stable than it was with the long board.

You always need to have the 64K internal card in place, otherwise you will have a whole in memory from 64K to 128K. The BIOS memory walk will correctly stop, but the jrIDE assumes you didn't do something silly like remove the internal card, so it will override the amount of memory that the system detected and then bad things will happen.

You might have bad memory on the system, so get a memory testing program and let it run over night. Memory testing programs are limited in that they can't test all of the system memory, but if you have an obvious problem the code and point it out, or the machine will keep crashing. I really wish they had put parity memory on these machines.

I'll post "jrMD" later tonight - it's a memory tester specifically for the Jr.

EDIT: Find it under jrMD.zip at https://www.brutman.com/PCjr/pcjr_downloads.html

Thanks much! I'll try to give that a shot when I have time. (background: I have a 1.5 year old so I rarely get time to work on the machine).

I did manage to get some testing done today:

System + 64k internal card and jrATX power supply.
I ran the repeated POST test using the service plug in the PCjr Hardware Maintenance and Service Manual. Maybe about 15-20 times with no issues.

I then decided to try going back to the original long board since that was the most recent system change. I put that back in and after 2-3 reboots, King's Quest started reliably working....even several retries later it still worked. Thexder is still a no go though, same issue.

Back to jrATX power supply. Neither game works, same as before.

Back to long board. King Quest seems to work, Thexder no.

Back to jrATX...repeat.. neither game works.

I find this very odd and do not have a good explanation for it. Maybe some slight difference in the voltage output? I'll put a meter on the jrATX power supply next time and look at the voltages when I have time again, lol.

I am suspecting bad memory overall though. One time it was trying to boot kings question the drive kept heading up the tracks until it hit 41, then it kept trying to drive the head further...obviously I stopped that non-sense immediately. I'll give the memory test you linked a test next as well! Thank you so much for the ideas, hardware is not my forte.

Could it be one of the capacitors on the board, or do those not really do that much? Specifically the one that looks suspicious on the left center.

Brutman wrote: ↑Thu Nov 04, 2021 6:54 pm I'll post "jrMD" later tonight - it's a memory tester specifically for the Jr.

EDIT: Find it under jrMD.zip at https://www.brutman.com/PCjr/pcjr_downloads.html

I was attempting to run the jrmd tool today and ran into an issue. It says the range of memory couldn't be tested (it defaults to 20000 - 1FFFF). And then when I read the manual, I see the following statement on page 5 under the "F2: Set Address Range" section:

"Segments 0 and 1 (128K) are contained in the PCjr system unit, and aren't allowed to be tested by jrMD. (See the addressing map at the end of this document.)"

Since I am having issues with just a bare bones 128k system configuration that is what I'm attempting to test with.

I also stuck a spare 64k expansion card I had laying around in the system and get the same issues as before. So I suspect it is might be one of the onboard 64k memory chips. Does that seem reasonable? How would I find out which chip is the issue? And can you still get the parts for them?

Thanks again!
Jason

I think your RAM is probably OK. If you ran the POST loop 15 to 20 times with the service plug and didn’t get any errors, that’s a good sign. To test further, plug in the “JJ” service plug, press Ctrl-Alt-Ins to get into the diagnostics, and let the RAM test run in the background for a while. You’ll see it counting up at the bottom of the screen repeatedly, and it will stop with an error code if it runs into a problem. The virtue of the built-in RAM test is that it runs from ROM and thus occupies very little of the RAM it’s testing. Furthermore, in service mode it includes a refresh test and will give you error codes that pinpoint any bad chips.

While you’re at it, you should run the diskette drive diagnostic as well. Put a blank, known-good diskette in the drive, choose the diskette test from the built-in diagnostics (it’s the first one), and type “MPNP” to confirm. This will help rule out the diskette drive as the source of your difficulties running games.

Go through the video diagnostic, too—number 8.

As for the power supplies, certainly measure the voltages from each one while they’re plugged in and the machine is on (so the supplies are under load). Another difference between the picoPSU and the original PSU card: the picoPSU and the brick feeding it 12 VDC are switching power supplies, while the original board is a linear power supply. Switching power supplies tend to produce more electrical noise on the power rails and thus could cause flakiness if the filtering weren’t adequate.

Regarding the capacitor: Most of the 10 µF electrolytic capacitors on the motherboard are for filtering out power supply noise. If they fail, they won’t be good filters, and noise on the rails could cause flakiness and crashes. However, these caps generally aren’t under a lot of heat stress and seem to hold up pretty well as far as I’ve seen. Plus there’s some redundancy.

All of that said, the cap in question, if I understand its position correctly, is a coupling capacitor in the audio section. Its failure would affect your audio output, not noise filtering on the power rails. The corrosion of the diode next to the cap suggests that the cap might have leaked its electrolyte. You should probably replace the cap and clean up the corrosion and surrounding area as best you can.

I usually suspect RAM when a system is unstable, but diagnosing RAM on the motherboard is pretty difficult without removing it. It's also not worth the risk of damage.

Based on that, go with what DoctorOctal just wrote ... use the built-in diagnostics to see if you can force the error.

Bad RAM is insidious because it can corrupt data, cause reboots, or just go unnoticed. It's often related to power, heating, and usage patterns. And without parity bits on the Jr, it is more difficult to detect because the system can't do anything to detect it.

DoctorOctal wrote: ↑Mon Nov 08, 2021 10:45 am I think your RAM is probably OK. If you ran the POST loop 15 to 20 times with the service plug and didn’t get any errors, that’s a good sign. To test further, plug in the “JJ” service plug, press Ctrl-Alt-Ins to get into the diagnostics, and let the RAM test run in the background for a while. You’ll see it counting up at the bottom of the screen repeatedly, and it will stop with an error code if it runs into a problem. The virtue of the built-in RAM test is that it runs from ROM and thus occupies very little of the RAM it’s testing. Furthermore, in service mode it includes a refresh test and will give you error codes that pinpoint any bad chips.

While you’re at it, you should run the diskette drive diagnostic as well. Put a blank, known-good diskette in the drive, choose the diskette test from the built-in diagnostics (it’s the first one), and type “MPNP” to confirm. This will help rule out the diskette drive as the source of your difficulties running games.

Go through the video diagnostic, too—number 8.

As for the power supplies, certainly measure the voltages from each one while they’re plugged in and the machine is on (so the supplies are under load). Another difference between the picoPSU and the original PSU card: the picoPSU and the brick feeding it 12 VDC are switching power supplies, while the original board is a linear power supply. Switching power supplies tend to produce more electrical noise on the power rails and thus could cause flakiness if the filtering weren’t adequate.

Regarding the capacitor: Most of the 10 µF electrolytic capacitors on the motherboard are for filtering out power supply noise. If they fail, they won’t be good filters, and noise on the rails could cause flakiness and crashes. However, these caps generally aren’t under a lot of heat stress and seem to hold up pretty well as far as I’ve seen. Plus there’s some redundancy.

All of that said, the cap in question, if I understand its position correctly, is a coupling capacitor in the audio section. Its failure would affect your audio output, not noise filtering on the power rails. The corrosion of the diode next to the cap suggests that the cap might have leaked its electrolyte. You should probably replace the cap and clean up the corrosion and surrounding area as best you can.

Thanks for all the advice and the additional info. Here's what I did:

Bare system + 64k board + floppy drive + ATX power adapter:
Floppy test - passed, no issues
video test 8 - passed, no issues
Memory test with JJ service plug - ran for 2.5 hours with no issues.

So apparently this is going to be a fun one to track down.... If I get time, I'll try to check ESR values on the capacitors that are on the board, I assume they should be reasonably low.

Any other ideas on what to check? I'm at a loss here. The system was working great about a year ago, so I'm not sure what's happened.

Thanks for the suggestions!

Brutman wrote: ↑Mon Nov 08, 2021 6:18 pm I usually suspect RAM when a system is unstable, but diagnosing RAM on the motherboard is pretty difficult without removing it. It's also not worth the risk of damage.

Based on that, go with what DoctorOctal just wrote ... use the built-in diagnostics to see if you can force the error.

Bad RAM is insidious because it can corrupt data, cause reboots, or just go unnoticed. It's often related to power, heating, and usage patterns. And without parity bits on the Jr, it is more difficult to detect because the system can't do anything to detect it.

Out of curiosity, what's the risk of damage? It seems like it wouldn't be too hard to de-solder and solder in a riser to make chip replacement/testing easier. Are the chips sensitive to heat? I've never replaced IC parts before.

Take my advice - avoid it like the plague. Do not start desoldering things for fun, only darkness and despair lies on that path. (1)

• Chips are heat sensitive.
• Motherboard traces can be lifted and damaged.
• It's a labor intensive process and you run a high risk of damaging something.
• Even just getting the board out of case may result in broken plastic. And you'll have to flex the board quite a bit.

(1) Professionals have desoldering tools. Desoldering 8 chips with a wick or a manual suction plunger device is just painful, even with a good temperature controlled iron.

Brutman wrote: ↑Tue Nov 09, 2021 4:55 pm Take my advice - avoid it like the plague. Do not start desoldering things for fun, only darkness and despair lies on that path. (1)

• Chips are heat sensitive.
• Motherboard traces can be lifted and damaged.
• It's a labor intensive process and you run a high risk of damaging something.
• Even just getting the board out of case may result in broken plastic. And you'll have to flex the board quite a bit.

(1) Professionals have desoldering tools. Desoldering 8 chips with a wick or a manual suction plunger device is just painful, even with a good temperature controlled iron.

Thanks! Yeah, I've had a little experience trying to fix some old Heathkit equipment and have had traces lift...it's no fun.

Okay, so I pulled out my spare motherboard that I've had laying around and have never tested so that maybe I could do some A/B component testing. I have a spare 64k memory card, floppy controller, and a spare floppy drive (although I need to figure out it's issues from a different thread here). So I have basically 2 of everything.

After a bit of A/B testing...I'm even more confused... the spare system exhibits VERY similar issues as the first system...even with no shared components! I tried the spare system with the only working 5 1/4" drive I have and the jrIDE with no floppy drive.

So I need to do some more testing and see if I can figure out what is going on and possibly post some better notes as to what all I've learned once I can get organized (maybe this weekend). I was able to get Kings Quest running with the jrATX power supply..plus all my old sidecars..I could not getting it working with just the jrIDE...

In any event, I have the following questions if you guys don't mind answering some really basic questions since I'm at a bit of a loss at this point and am trying to think of any common threads.
1) For testing, I've had my monitor about 6-10 inches to the left of the system. Is there any reason to believe there could be any interference here? I am using the PCjr color monitor.
2) Could the pico ATX power supply be doing anything weird by not having enough load on it? To some degree the system seems happier when I have more sidecards plugged in, therefore I assume more load is on it. Here is the exact power supply I have (https://www.mini-box.com/picoPSU-90-100W-power-kit).
3) Is there some basic autoexec.bat/config.sys setup I should be using with the jrIDE? I hadn't had time to set up a basic config for it yet and I didn't think it would matter for what I'm testing, but maybe I'm mistaken. This would not affect testing King's Quest since it is a booter, but maybe the Thexder testing I am attempting to do as I forget what it's original requirements were.

Thanks for any suggestions, right now I wouldn't put it past myself to be overlooking something really simple as the results I'm seeing just don't make much sense. The odds of both sets of parts having the same issues seems highly unlikely.