Topic: Commodore PET 8032 Repair Log  (Read 7547 times)

Author Message

0 Members and 1 Guest are viewing this topic.

Offline Phu

  • RCM Workshop
  • Committee
  • Amiga 4000
  • *
  • Posts: 2084
  • Kudos 41
  • Gender: Female
  • Pay no attention to that PCB....
    • View Profile
    • ZX Spectrum Laptop Project
Commodore PET 8032 Repair Log
« on: June 01, 2011, 10:51:03 PM »
I thought I'd treat you to a Womble-style repair log on this fairly simple fix.

If you seen our PET at the recent RCM Weekender, you would have noticed it was a little ill. Its screen looked like this:



By itself this might not have been a problem, unfortunately Scramble reads the screen back to see if you collided with anything. Which meant unless you kept your finger on FIRE, you collided almost instantly.

Before showing the chips that failed, you might like to know how PET screen memory works. There are a set of 2114 RAM chips (1K x 4 bit) that are dedicated to the screen. The PET not only uses these to display, but because of the "type anywhere" nature of input, it uses it as an input buffer too. Which means when what you typed shows as "doad" the PET thinks you actually typed "doad" - which doesn't help.

The 8032 has an extra quirk. To maintain some sort of compatibility with 40-column PETs, the screen memory is in two sets. One set manages the even columns, and the other set manages the odd columns. If you look at the screenshot, you can see that the fault is at a seemingly predictable position. In fact, the faults are on exact multiples of 32 characters.

Remembering the odd/even nature, this means that the fault is possible every 16 bytes. We can also work out which half (thus chip) and which bit is faulty, however (as will be seen later) this academic. The code for * is 2A hex, where as the code for " is 22 hex. 2 and A differ by one bit... the "8" bit, or the top bit of a 4-bit block. Thus, bit D3 of one of the 2114s is faulty.

Looking at this image, you can see where this fault occurred before on a previous chip (and caused no end of lopd'ing fun):



Since this is the second chip to fail like this, and since spares are available I decided to go the whole hog and replace all three remaining 2114s. Incidentally, the chips have a manufacture date code of March 1980... they did well for 31 years of service (they are the same age as me!).

I used a hot air gun and solder sucker to desolder the chips. Unforunately, on a board this old a couple of other things occurred:



The capacitor above one of the chips snapped under heat stress. This is only a decoupling capacitor designed to smooth out any voltage ripples that may have arrived at the chip. Its marked 104M, which is a 100nF part. I also managed to knacker a nearby via that seems to have buckled under heat stress.

Some wire and a spare 220nF part later (value is not vital here) along with 3 18-pin sockets and 3 spare 2114 chips, the board looks good again:



The rest of the board seems to have held up well, though you can see the adapter board I had to make in its first repair to replace the character ROM:



It required a transistor/2 resistor NOT gate on it in addition to the socket adaption as Commodore used a 2364 mask ROM and activated it with an active-HIGH signal... something that 2764 EPROMs don't support.

Powering the old dear back on gives a nice clear display, and we see the DRAMs are still in fine working order:



I just need to take the keyboard apart and clean the contacts now so that we can type LOAD"",2 in less than a minute!

-- Richard
8 End of File, RCM:1

Offline Scarlettkitten

  • Enjoy the silence
  • Administrator
  • Amiga 4000
  • ******
  • Posts: 2025
  • Kudos 31
  • Queen of the 3DO people
    • View Profile
Re: Commodore PET 8032 Repair Log
« Reply #1 on: June 01, 2011, 10:57:33 PM »
Nice Richard, good to see the old girl fixed :)

Offline porchy

  • Moderator
  • Amiga 4000
  • ******
  • Posts: 825
  • Kudos 33
  • Gender: Male
    • View Profile
    • JAMMArcade
Re: Commodore PET 8032 Repair Log
« Reply #2 on: June 02, 2011, 06:41:54 AM »
very interesting read, cheers

Offline AndyRCM

  • >=))))º> GO FEED THE FISH! <º((((=<
  • Administrator
  • Amiga 4000
  • ******
  • Posts: 9660
  • Kudos 50
  • Gender: Male
  • Manic Jet Set Willy
    • View Profile
    • Retro Computer Museum
Re: Commodore PET 8032 Repair Log
« Reply #3 on: June 02, 2011, 07:58:20 AM »
Nice one Richard. Always good to see her working again! ;) A
"I could see the faces of those who led pissing themselves laughing" - Funeral Pyre by The Jam

Offline billdooruk

  • Administrator
  • Amiga 4000
  • ******
  • Posts: 2476
  • Kudos 26
  • Gender: Male
  • I could murder a curry
    • View Profile
    • Our site:-
Re: Commodore PET 8032 Repair Log
« Reply #4 on: June 02, 2011, 08:13:42 AM »
Great Work as ever Phu :)

Offline questor

  • user
  • Amiga 4000
  • *
  • Posts: 862
  • Kudos 38
  • Gender: Male
    • View Profile
Re: Commodore PET 8032 Repair Log
« Reply #5 on: June 02, 2011, 02:36:28 PM »
Great work and good call to replace and socket all the RAM - hopefully she will be stable for a while longer.

Offline HeadingtonBard

  • C128
  • ****
  • Posts: 272
  • Kudos 12
  • Gender: Male
    • View Profile
    • Alex Aris Games
Re: Commodore PET 8032 Repair Log
« Reply #6 on: June 02, 2011, 07:12:59 PM »
" Dr. Phu ... Calling Dr.Phu ! We have a system that is starting to flatline ! "

Well done Mate ! Great job ! ;)
:: HeadingtonBard ::

[ http://www.alexarisgames.co.uk/ ]



Why not check out Google Play Store for more games to buy and for *FREE* with no adverts or nag screens !

Offline Prime

  • CPC 464
  • ***
  • Posts: 230
  • Kudos 12
    • View Profile
Re: Commodore PET 8032 Repair Log
« Reply #7 on: June 02, 2011, 08:15:30 PM »
Great work and good call to replace and socket all the RAM - hopefully she will be stable for a while longer.

I think I would have done the same, as I guess two chips have already failed, only a matter of time before the others do, so makes sense to socket them now. Also once they are in sockets I suppose if one of the chips goes again it's then just a matter of power down, unplug faulty one insert good one, power up again....

Cheers.

Phill.

Offline Phu

  • RCM Workshop
  • Committee
  • Amiga 4000
  • *
  • Posts: 2084
  • Kudos 41
  • Gender: Female
  • Pay no attention to that PCB....
    • View Profile
    • ZX Spectrum Laptop Project
Re: Commodore PET 8032 Repair Log
« Reply #8 on: June 02, 2011, 09:55:28 PM »
Well it seems that the RAM chips are only one half of the story.

Having cleaned out the keyboard (she types wonderfully now ;) ) a short test program reveals that when the RAM is stressed (e.g. scrolling lots of text) characters get miscopied.

I haven't looked in detail yet, but I remember each chip pair has a latch associated with it. I have a feeling those chips have snuffed it too.

Follow up log to come :)

-- Richard
8 End of File, RCM:1

Offline Prime

  • CPC 464
  • ***
  • Posts: 230
  • Kudos 12
    • View Profile
Re: Commodore PET 8032 Repair Log
« Reply #9 on: June 02, 2011, 10:02:19 PM »
Well it seems that the RAM chips are only one half of the story.

Having cleaned out the keyboard (she types wonderfully now ;) ) a short test program reveals that when the RAM is stressed (e.g. scrolling lots of text) characters get miscopied.

I haven't looked in detail yet, but I remember each chip pair has a latch associated with it. I have a feeling those chips have snuffed it too.

Or buffer chips perhaps between the 6502 bus and the 6845 bus ?

Cheers.

Phill.

Offline Phu

  • RCM Workshop
  • Committee
  • Amiga 4000
  • *
  • Posts: 2084
  • Kudos 41
  • Gender: Female
  • Pay no attention to that PCB....
    • View Profile
    • ZX Spectrum Laptop Project
Re: Commodore PET 8032 Repair Log
« Reply #10 on: June 03, 2011, 09:09:33 AM »
Looking at the circuit diagram its one of two places. Either the 74LS373 data latches that sit between the 2114 screen RAM and the character generator ROM, or the 74LS244 buffers between the 2114s and the CPU.

I'm guessing the latter for two reasons:

1) It seems to have no problem writing to screen RAM, just reading it back again (i.e. the read-write cycle of scrolling).

2) They are 74LS244s which I haven't got any of right now, so naturally it has to the most inconvenient part (I have 8 74LS373s lying around).

-- Richard
8 End of File, RCM:1

Offline porchy

  • Moderator
  • Amiga 4000
  • ******
  • Posts: 825
  • Kudos 33
  • Gender: Male
    • View Profile
    • JAMMArcade
Re: Commodore PET 8032 Repair Log
« Reply #11 on: June 03, 2011, 11:06:51 AM »
You want some 244's sent through fella?

Offline Phu

  • RCM Workshop
  • Committee
  • Amiga 4000
  • *
  • Posts: 2084
  • Kudos 41
  • Gender: Female
  • Pay no attention to that PCB....
    • View Profile
    • ZX Spectrum Laptop Project
Re: Commodore PET 8032 Repair Log
« Reply #12 on: June 03, 2011, 11:40:09 AM »
If you have some handy you don't mind sending my way...

I can get some from Farnell, but that incurs costs. I also don't like placing an order for just a few chips if I can help it :)

-- Richard
8 End of File, RCM:1

Offline porchy

  • Moderator
  • Amiga 4000
  • ******
  • Posts: 825
  • Kudos 33
  • Gender: Male
    • View Profile
    • JAMMArcade
Re: Commodore PET 8032 Repair Log
« Reply #13 on: June 03, 2011, 11:46:38 AM »
No problem, how many you want?

Offline Phu

  • RCM Workshop
  • Committee
  • Amiga 4000
  • *
  • Posts: 2084
  • Kudos 41
  • Gender: Female
  • Pay no attention to that PCB....
    • View Profile
    • ZX Spectrum Laptop Project
Re: Commodore PET 8032 Repair Log
« Reply #14 on: June 03, 2011, 11:49:44 AM »
For the PET, I need at least 4. But I'd love a few spares if you them to part with :)

Ideally around 20 or so would be fantastic, but I think also extracting the urine :)

-- Richard
8 End of File, RCM:1