• Please review our updated Terms and Rules here

Frank's Danish PET

carlsson

Veteran Member
Joined
Jul 30, 2003
Messages
6,274
Location
Västerås, Sweden
Today I've had a PET day. I brought out two 3032 from the basement: one known working and one labeled as "emits smoke".

Emits smoke was an understatement. The four big diodes on the power supply part of the motherboard were crackling, smell really badly and have almost corroded all the solder on the underside. However the monitor flashed a screen of garbage characters for a second before it went black.

However I wasn't sure if the major fault was with the transformer or the board, so I plugged in the faulty board in the other PET with the known good transformer/monitor. Slight crackling, then nothing happened. No display. I put back the known good motherboard, and still no action. Damn, did I break another PET by this move?

So I did what every daring hobbyist repairman would do: put the known good motherboard in the case where the bad motherboard originally came from. It can't get worse, can it? No, this time the PET powered on nicely. The monitor has a bit more burn-in than the other one but also brighter picture. The transformer is a bit cleaner wired. I had to swap keyboards though as one is very good and the other is in need for a complete overhaul.

It means I now have a PET 3032 with Basic 4 in a slightly dirty case smelling of smoke, with a monitor that shows signs of burn-in. I may try to swap parts over, but both the transformer and in particular monitor are a bit tricky to remove unless you really have to.

Ok, so what does this have to do with Frank's PET? Well, I had previously verified all his EPROMs to be OK but the last ROM (901465-23) didn't verify when read as a 2532. Therefore I swapped this ROM with my set and was greeted with a display of garbage! Back to the EPROM programmer, program a "fresh" 2532 with the same code and the PET boots again. I swapped in all the other EPROMs from his set, and my PET works OK with those except for his business keyboard layout doesn't match my graphics keyboard layout.

It means I'll send back Frank's set of EPROMs. Hopefully his PET will come alive after all.

While swapping chips, I got curious about the state of the ROM set on the board that had suffered smoke signals. It is a Basic V2 set with a PAICS ROM expansion at $B000. I put those chips into my working motherboard, and it promptly booted to the Basic V2 prompt. SYS 11*4096 also started the PAICS expansion. It suggests that even on a board where the diodes have been crackling and emitting smoke, the ROMs may still be OK. I didn't yet check the big chips: 6502, 6520 and 6522.
 
For what it is worth, after I replaced the blown fuse (!!) in the previously known good PET, it works again so I swapped back the motherboard and keyboard to original state. Eventually I'll dig into exactly what kind those blown diodes are and may attempt to replace them. The big electrolyte capacitor next to it looks fine and no other visual damage on the board so with a lot of luck it might be enough to replace four diodes to get the other PET back to life.
 
No, the blown fuse was found in the good power supply. So this is how it worked:

1. Good PET, everything OK
2. Bad PET, smoking motherboard
3. Moved bad motherboard to good transformer, fuse blew
4. Moved good motherboard to unknown transformer, working
5. Replace fuse in 3, moved back motherboard, now working again

You tell me why one power supply could withstand a bad motherboard but the other didn't. It should be said that the transformer whose fuse blew originally comes from a 8000 series PET but apparently works the same way with a 3000/4000 series machine, just a bit different appearance.
 
Well, it's not really a "power supply withstanding a bad motherboard" since most of the power supply is on the motherboard. Maybe the fuses were different sizes?

Sounds like what you call the "power supply", i.e. the transformer and filter caps, are fine in both systems but there's a short somewhere on the bad motherboard; could be a shorted diode (easy to check) but could also be any of a number of other parts causing a short.
 
Sorry. Yes, I've learned the difference between power supply and transformer and try to use the correct term as often as possible. :)
 
Upload_19_04_2011_075.jpg
[/url][/IMG]

This pic shows the screen without UC6 and UC7, and with the Eproms installed.
31743 bytes frue... Could it mean that the RAM are OK?

Upload_19_04_2011_076.jpg
[/url][/IMG]

This pic shows the screen with UC6 and UC7, and with the Eproms installed.
Do any here in the Forum have an idea of the problem?

And a big "thank You" to Anders Carlsson for the help and the new chip.
 
Last edited:
This pic shows the screen without UC6 and UC7, and with the Eproms installed. 31743 bytes frue... Could it mean that the RAM are OK?[/quote]Main RAM is almost certainly good. Don't know about the video RAM though; sure looks like a problem in the video department, especiallly with the screen split like that.

[quote]This pic shows the screen with UC6 and UC7, and with the Eproms installed.
Do any here in the Forum have an idea of the problem?[/QUOTE]I doubt that it has anything to do with UC6&7 but you could try swapping them just in case. Intriguing pictures though.

What do you see with the character generator ROM removed?
 
Which slots are UC6 and UC7? Are those the $9000 and $A000 utility ROM slots? I'm too lazy to go into the livingroom to check for myself.

It should be noted one of the unknown EPROM chips (outside the regular Basic V4 + char ROM set) Frank sent me had one pin (#13 IIRC) broken off so it may not address correctly. Would a such chip cause garbage all over the screen even when not initiated?
 
The UC6 is the chip for IEEE-488, and UC7 is the PIA. Both are 6520 chips!
Since the last activity I have found out, that at least one of the 6520 chips must be defect because it gives most garbage (see photo).
It gives the same result if I put it in UC6 or UC7... Does any have a good idea?
 
I had a 8032 (IIRC) with a broken 6520. It was all black when I powered it on. After replacing that chip, it booted up properly. I believe the 6520 falls in the same category as 6522, chips that can have various grades of failure, showing in different ways. Unfortunately I don't really have any known good spares.
 
With both UC6 and UC7 removed, you get the first almost-correct screen, and with both installed you get the second screen; correct? I'd say the ROMs are OK.

1 - Leave UC6 empty; Mark which 6520 is which and install each one into UC7. What happens?

A - They both change the screen as in the second picture.
B - One does not change the screen, the other changes it as in the second picture.
C - One does not change the screen, the other one does, but not like the second picture.

2 - Remove UC7 (and leave UC6 empty). Remove UF10 (Char Gen). What do you see?

3 - Replace UF10 (Leave UC6 & UC7 empty). Mark which is which and remove both UF7 and UF8 (VRAM). What do you see?

4 - Replace UF7. What do you see?

5 - Remove UF7 and replace it with the chip that was in UF8. Any different from #4?
 
Last edited:
Hi Mike.
The UC7 6520 placed in UC7 and nothing in UC6 gives a screen as the first photo, and I also get a flashing cursor. The keyboard also function well!
If I put the UC6 6520 in UC7 and nothing in UC6 it gives the second photo...

Frank
 
Hi Mike.
The UC7 6520 placed in UC7 and nothing in UC6 gives a screen as the first photo, and I also get a flashing cursor. The keyboard also function well!
If I put the UC6 6520 in UC7 and nothing in UC6 it gives the second photo...

Frank
OK, one problem found. Looks like the UC6 6520 is bad and the other one's OK; no problem, mark the bad one so you don't use it again and try the other tests.
 
Good work!

Replace UF7 and UC6 and you should be in business. You only need UC6 for IEEE I/O so you can just leave it out until you get a new one.

Good luck!

Mmmmm... Danish.... <HomerDrool>...
 
Last edited:
All right, but where do I buy a new UF7???
They're not really very scarce yet, despite the fact that they're notorious for failing. Look for a 2114 or an AM9114 (and buy a spare or two while you're at it ;-) ) Speed is not critical, but you might as well get the faster one if there's a choice.

I think this fellow is a member of this forum, and his prices look pretty good (if he has stock):
http://www.arcadecomponents.com/memory.html

Here's a place on your side of the big pond, but the prices don't look so good:
http://www.elettrosurplus.com/sh_index.php?c=64&p=341

I've got a few, but you'd be an old man if you wait for me to get around to finding and shipping them; However, maybe, just maybe, Anders can find one or two in his vast inventory of retired PETs and you can sweet-talk him out of a couple...
 
Back
Top