• Please review our updated Terms and Rules here

DOA PETs. (2001-8, 4032 and 8032)

I washed my 4032's board today. I went with a hand wash in warm water with some soft-bristle toothbrush scrubbing, mainly because I actually don't have access to a dishwasher right now. (Kitchen is being rebuilt. Great timing!) I think I got all the sticky dust off, but ended up with only mixed results getting the rust off the IC pins. I also ended up hitting some areas with alcohol swabs... it's odd, but in particular areas that show signs of previous repair would end up with this cloudy, oxidized residue on them even after multiple rinses in distilled water.

Cleaning the board does give me some more idea what I'm up against. This machine has had a *hard* life. The 555 timer in the reset circuit appears to have been replaced by some sort of lower life form. I was going to say "an ape", but I think I'm looking at the soldering work of a creature that doesn't have a full compliment of fingers, let along an opposable thumb. (My soldering skills are pretty rotten but I managed to not turn the board brown anywhere I worked on it.) There are several more hackish chip replacements, and a number of spots where it appears traces have been repaired by running solder over cracks. (I wish I could get the hang of that.)

I'll let it dry for a few days and see what I get. I'm starting to wonder again about that ROM image that supposedly existed to use a Universal board in a 9" PET. After seeing all this damage... trying not to lose hope.
 
Sounds bad. Well at least you now know what you're up against.

One of the problems with buying these machines. You never know just how much they've been hacked around with.

I've now managed to get one of my 3032 PET boards going. However, if I was to total up the number of faults that board has had since I received it a couple of years ago it would come to six.

Originally it had...

1. A faulty F000 ROM
2. A faulty B000 ROM
3. A faulty 2114 video ram chip
4. A second faulty 2114 video RAM chip

And more recently....

5. A bad PIA (6520) socket
6. A bad E000 ROM socket

And that board had not been hacked around with!

These machines can certainly be a challenge, especially when they have multiple problems.

Tez
 
tezza said:
One of the problems with buying these machines. You never know just how much they've been hacked around with.
Click to expand...

I guess I'm glad I didn't *buy* the machine, but I suppose I am in it somewhere to the tune of $170 or so between buying parts I haven't needed and a soldering station. :^/

I really wish the problems with my machine were "just" bad components. These PCB issues are really annoying. And... maybe I *do* need to borrow a dishwasher, or something more radical. I was just passing by and decided to take a look at the board. Up in the upper right corner there's a hamfisted fix to the IC in UB2. It looks like the IC may of been replaced in total and on the upper right leg of the chip there's a large blob of solder on the component side that appears like it might of been put there to bridge over a bad spot on the trace leading to the chip leg. That's one of the spots where I noticed a mysterious white scale (like a hard water stain) immediately after washing the board. At the time I swabbed it off with alcohol and re-rinsed the board. Now, several hours later that same joint (and the leg of the IC) were *covered* with a white scale that looks an awful lot like what you get when a dry-cell battery leaks. WTF? There are several other white spots like that and I'm at a loss.

Could there be old flux on the board? Acid flux? I have no idea what I'm seeing here. Maybe this PET was a victim of a serious liquid incident at some point?
 
Eudimorphodon said:
it's odd, but in particular areas that show signs of previous repair would end up with this cloudy, oxidized residue on them even after multiple rinses in distilled water.
Click to expand...

That's flux residue. It's a sign that the board had quite a bit of flux left over on it after its previous repairs over the years :/ More isopropyl alcohol and/or flux remover should help.

Man, this board is a daunting task indeed. You know what someone needs to do? (someone who is not me! ;) ) A Pet Replica PCB, just like the Apple I Mimeo replica ( http://www.willegal.net/appleii/apple1.htm )
 
tezza said:
Sounds bad. Well at least you now know what you're up against.

One of the problems with buying these machines. You never know just how much they've been hacked around with.

I've now managed to get one of my 3032 PET boards going. However, if I was to total up the number of faults that board has had since I received it a couple of years ago it would come to six.

Originally it had...

1. A faulty F000 ROM
2. A faulty B000 ROM
3. A faulty 2114 video ram chip
4. A second faulty 2114 video RAM chip

And more recently....

5. A bad PIA (6520) socket
6. A bad E000 ROM socket

And that board had not been hacked around with!

These machines can certainly be a challenge, especially when they have multiple problems.

Tez
Click to expand...
Actually, those are the classic 3 issues with PETs (crappy IC sockets, bad 2114 RAMs and bad ROMs), so I'd say you were actually quite lucky compared to some of the folks here with broken traces, video problems, etc. ;-)
 
Eudimorphodon said:
That's one of the spots where I noticed a mysterious white scale (like a hard water stain) immediately after washing the board. At the time I swabbed it off with alcohol and re-rinsed the board. Now, several hours later that same joint (and the leg of the IC) were *covered* with a white scale that looks an awful lot like what you get when a dry-cell battery leaks. WTF? There are several other white spots like that and I'm at a loss.
Click to expand...

Second reply, because I should have read this post first and consolidated!

Yeah, that's DEFINITELY flux residue. It is shockingly persistant, and when it's been partially removed, washed, and then exposed to air, it spreads out and turns into that white stain. It makes me wonder if some of the corrosion came from leaving the flux on it originally, back in the 80s when I presume it was first ham-fistedly repaired? Rosin flux isn't corrosive, but water soluble flux is very corrosive if left in place. I wonder if they used that? *shudder*

The only way to clean the really persistant stuff is more washing. Isopropyl alcohol will do it, but you need to keep applying, brushing, rinsing, applying, brushing, rinsing...

You can find commercial flux remover at Fry's, but I'm not convinced it's really any better than isopropyl alcohol + toothbrush. It's certainly fouler, and you definitely need gloves when using it. Apparently Jeri Ellsworth makes her own flux remover by combining isopropyl alcohol + acetone nail polish remover. Now that's good thinking!
 
Twylo said:
...Man, this board is a daunting task indeed. You know what someone needs to do? (someone who is not me! ;) ) A Pet Replica PCB, just like the Apple I Mimeo replica ( http://www.willegal.net/appleii/apple1.htm )
Click to expand...
Sounds like a great idea!

As to troubleshooting I still recommend the old NOP generator and systematically following all the address lines and looking at data lines on the various chips; that catches a lot of problems with bad sockets, broken traces, buffer chips etc.
 
No magical cures, but no worse than before...

No magical cures, but no worse than before...

I had a chance and deemed it dry enough so I smoke tested the PET board this morning. Nothing (yet) appears to have been harmed by the wash as it came up as it was, a non-moving screen of garbage.

Didn't have time to do much else with it, but I went ahead and socketed the tester ROM and did some poking with the voltmeter. Checked the obvious candidate first and found that RESET was stuck low. And lo and behold if I hold a jumper between 6502 pin 8 (5v+) and pin 40 PETTESTER runs.

When I get some more time I'll walk backward through the reset circuit and see what I see. Based on the fun Tezza had I'm wondering if this *could* be as simple as a bad socket, although based on the track record of this board it's probably a cracked trace. Whatever it is it's somewhat intermittent, because it did that "thing" where at one point it ran a few milliseconds worth of the test code while I was just randomly touching the board.

So I guess I'm not giving up yet. Yay.
 
I feel you are getting close. If it's intermittent and "comes right" with a board touch it does indeed indicate connectivity is the issue rather than any failed IC.

If the board has had a lot of repair work on it it's probably ALSO had a lot of chip swap-outs. Might be worth testing each of the relevant CPU lines at wherever they go to, just to make sure a socket connection is happening.

Stick with it.

Tez
 
Dove into the reset circuit today...

Dove into the reset circuit today...

For once the problem doesn't seem to be a bad trace. Although... I'll get back to that.

I've traced all the signal connections back from the reset pin on the CPU through A10, A3, and the output on A2 pin 3. Everything checks out. When I power up the machine the output pin on the 555 timer reads about 4.7v and stays there.

dave_m said:
It sounds like you have another intermittent signal. The /RESET signal is easy to trace. The source is the 555 timer A2 pin 3 which, based on a RC time constant, outputs a positive pulse for less than a second on power up only. That signal is inverted by A3 and buffered by A10 to give a negative going pulse on the /RESET line to give a clean start for the 6502.
Click to expand...

I checked out all the connections from the 555 to the analog components (Capacitors c66 through 68, resistors R15 and R16, etc.) All of that seems solid as well. My cheap-o DVM says the resistors are in the ballpark of their specified 1M rating. Said cheap-o DVM seems less useful for telling me if a capacitor is a specific value but none of them seem to be shorted or open. So I'm guessing the problem is the 555.

Here's where I'm I little unsure what to do. As I mentioned earlier, someone *really* butchered a soldering job on the 555 sometime in the past. The board has brown stains (burn marks?), corrosion, and I also just noticed that one of the trace pads on the component side seems to lifted at least partially off the fiberglass. (If my macro-lens digital camera were charged up I'd take a photo. It's quite a sight.)

Should I try to desolder and clean up the board to install a socket or would it be safer to try to cut this thing off, leave the pin legs in place, and solder a replacement onto those? I hate to replace an IC without putting in a socket, but this is a real piece of work here.

Yay.
 
You know it still may be a cap in that circuit. This sounds very similar to the problem I had with my Apple clone. It's detailed here.
http://www.classic-computers.org.nz/blog/2009-03-29-re-repairing-an-apple-II-plus-clone.htm

Here the reset line was also being help low and the 555 output was being held up at 4v. The problem wasn't the 555, but a capacitor in the circuit. This cap didn't show up as being shorted or open. In fact, the only reason I managed to track it down was because I had a "good" apple II+ board I could compare readings with. It's in the article. It could be the 555 but it sounds very similar to what I had...maybe it's the same cause?

Tez
 
If you do have to extract the chip, clip the legs off on the component side of the board. Heat the pins a bit and pull them out. Clean the holes from the component side of the board and then stop.

Glue the pad back down with some epoxy using the small end of a toothpick (the board is G-10 epoxy anyway) and let it dry. Don't get any epoxy on the solder side of the pad or in the via.

When it's dry, put your socket in as usual, just be very careful with any browned traces.

If the solder doesn't flow up the hole with the damaged pad, solder it from the component side and it will meld with the solder on the solder side.
 
Eudimorphodon said:
When I power up the machine the output pin on the 555 timer reads about 4.7v and stays there.
Click to expand...

Tez may be right, it could be the cap (C67). Scope A2 pin 6 on power up. If it instantly jumps to +5 and stays there, the cap may be open/high impedance. Pin 6 should start low since the cap is discharged and charge up to +5V within about a Second if the cap is good. If you are getting a good RC time constant, and the 555 output does not drop to ground, then the timer must be bad.
 
PET resetting again. So... now what...

PET resetting again. So... now what...

dave_m said:
Tez may be right, it could be the cap (C67). Scope A2 pin 6 on power up. If it instantly jumps to +5 and stays there, the cap may be open/high impedance. Pin 6 should start low since the cap is discharged and charge up to +5V within about a Second if the cap is good. If you are getting a good RC time constant, and the 555 output does not drop to ground, then the timer must be bad.
Click to expand...

I finally made it to the electronics store this week to get some goodies, and here's the results of this afternoon's poking around:

First thing I did is replace the Tantalum cap in C67. Unfortunately the results were negative. Both before and after replacing the cap I noticed that the tester ROM would intermittently and randomly run for a few milliseconds, so I wired up the oscilloscope to the output pin on A2 to see what was going on. On power-up I'd usually see A2 go high for about a second, as it's supposed to, but then about half the time it'd "twitch" off for just an instant. And if I left the machine on for a while in one of the power-up cycles where it twitched it'd often keep twitching at semi-regular intervals.

Anyway, just for the heck of it I decided to swap the other two caps in the circuit, C66 and 68. I replaced them with dipped monolithic ceramics instead of the original type, but followed the schematic with regards to value. And... now the system seems to be reliably resetting. Yay. I'm relieved I don't have to dive into the horror of desoldering the 555.

I repopulated the board with all the peripheral and ROM chips, and it's back at the BASIC prompt. One nice thing is it appears the original D000 ROM works. I'd been using an EPROM replacement since it'd seemed like there was a problem with that one, but perhaps it just wasn't seated correctly. So... for anyone keeping track the ROMs, another supposed common problem spot for PETs, are all good on this one. Yay?

Next thing I did was try rotating another 2149 (supposedly a suitable replacement for 2114s, which I haven't been able to find locally) into both VRAM slots to see if it would solve the misplaced/garbage character problem. No luck. And just for the heck of it I made a vain attempt to "shotgun" a problem based on another recent thread by piggybacking a new 74LS157 one-by-one on the video address multiplexers at F3, F5, and F6... again, no change. I'm sort of stuck how to proceed at this point to clean it up. Maybe I can figure out some sort of test program to exercise lines with while monitoring with the logic analyzer.

The last thing I've attempted was to try wiring up a couple of the disk drives I picked up. Using the best-looking IEEE-488 cable I had on hand I tried an MSD SD-1 and and an 8050, as both were passing diagnostics. Here's where I have to ask some dumb questions, since I'm barely at all familiar with Commodore disk usage. When I type "CATALOG" or attempt to save a file to disk (with either DSAVE or SAVE "name",8) the PET is immediately returning to the READY prompt, without the disk drive spinning at all. Meanwhile, If I attempt to load a file again the drives don't respond at all and the PET simply spits out a ?FILE NOT FOUND error.

At first I assumed the drive I'd connected up must be having some sort of internal problem to just return a "I saved the file for you" status after not spinning up following a SAVE "name",8 command, but when the same thing happened on the next drive I decided to fire up VICE and see what happened if I typed the same commands on an emulated PET with no drive peripherals attached. The behavior seems to be the same! (IE, a SAVE will just happily disappear into the aether and CATALOG won't complain about about there being no drives to catalog.) Is this an accurate depiction of how a PET behaves? Clearly the PET is failing to communicate with the drives, but whether it's the cable, the PET, or the drives I have no idea because I have no working combination to swap with. I'm wondering if there's any way I can try to whittle down the possibilities. Is there some sort of loopback plug or something I could construct to troubleshoot the IEEE 488 port with?

I'd *really* like to get a working storage device. Neither of my tape drives work... or maybe it's the PET. I don't seem to be able to talk to my drives, which may or may not work... sigh. It's a serious disincentive to try to get diagnostic software onto the machine if I have to retype it every time I power the thing on.

Thus this remains a frustrating endeavor. :^/
 
Check out this user manual for disk syntax

Commodore BASIC 4 Users Manual

Have you tried to format a disk properly? For BASIC 4, use the Header command. I think the MSD uses double density and the 8050 may need a higher density.

And perhaps the MSD does not handle BASIC 4 commands. Try the older Commodore syntax to 'NEW' the disk for the MSD.

This book has all the commodore syntax: "PET/CBM Personal Computer Guide". find it on this page on bombjack
 
Last edited:
Eudimorphodon said:
Anyway, just for the heck of it I decided to swap the other two caps in the circuit, C66 and 68. I replaced them with dipped monolithic ceramics instead of the original type, but followed the schematic with regards to value. And... now the system seems to be reliably resetting. Yay. I'm relieved I don't have to dive into the horror of desoldering the 555.
Click to expand...

Congrats for another rung up the ladder to full workability. Slowly but surely...

Tez
 
Keep up the good work Eudimorphodon! On the 2114 front it may be worth picking up a broken Vic 20(early type) off ebay as they have ten 2114's in them.

On the drive front, does the drive reset when you turn the computer on and off? On one of my pets it would not recognise the drive and I found it to be a dirty edge connector on the ieee 488. Once cleaned with some isopropyl alcohol it worked fine.
 
Last edited:
I've got a extra dead vic motherboard here (Chattanooga, TN) that you can have cheap. (It has rust and crap on other parts of the board.)
Currently it's mounted to my wall. The ram is NOT in sockets.
Drop me a pm.

Later,
dabone
 
Dabone you can get good results on the Vic 20 using a heat gun. I've removed the 2114 out of my Vic 20 using the heat gun without damaging the board or the chips. If its a scrap board it's worth practicing on with nothing to lose. Use an embossing type heat gun though and not the paint stripping type they are less harsh but be warned can still do damage!
 
You must log in or register to reply here.
Back
Top