• Please review our updated Terms and Rules here

8032 repair

Gary C

Veteran Member
Joined
May 26, 2018
Messages
2,256
Location
Lancashire, UK
Getting lots of PET practice

Now its time for an 8032 (not an SK this time) that was working but always a bit flakey

No chirps, no screen.

Removed all but the CRTC and the CPU and installed a tynemouth board with RAM and CPU mapped out and up it comes but with ! on every odd character, its a start :)

PETTEST installed and ROM's mapped in and its showing no problems in the display RAM's but to be sure, I extended the test to check every location of the 80 column screen and sure enough it wont complete. As its a ! rather than a space character, its obviously bit 1 thats stuck in the odd SRAM, so swapped and screen now working.

Now, with the normal ROMS in, nothing. Kernal rom has been butchered in the past and is missing two legs for which bits of leg have been jammed in. So wrote a 2532 with the kernal and now it gives the chirps but no basic welcome. Took all the basic ROM's out and verified them in my programmer and, damn, one has bit rot. Another 2532 programmed and now its working.

It came with a tynemouth 40/80 column switchable edit ROM installed so thats three of the original ROMs missing :(

But at least its working.
 
Surely the PETTESTER would have told you the state of the ROMs (with the exception of the EDIT ROM)?

I think it is data bit 0 that is possibly faulty in the ODD bank - but it is the same chip so who cares. V5 will fix this missing error detection mode. It is the too simple a data pattern I am using.

I am, however, surprised that extending the video RAM test showed this up (unless there are two concurrent faults with the same video RAM device - or a strange address fault internal to the RAM).

All Opex for V5!

Dave
 
Ooops, yes first bit as in bit 0.

Must take more care.

I did it in a bit of an odd sequence, using the tynemouth board (its really useful for seeing if the very basics of the PET are working) and PETTEST, and looking at what I wrote its a bit out of order I think, as the kernal ROM was missing legs I reprogrammed an EPROM and while I had the machine setup I checked all the others while I was at it.

But the SRAM test, yes thinking about it I cant work out how it passed the screen test with ! in every odd character but failed when I extended it (I had already done a mod for the problem I had with the SK, though its not perfect as it doesn't clear the screen properly before it moves on). But it must have worked as the test then passed when I swapped the SRAM chip.
 
Of course if bit 0 of the odd SRAM is faulty and stuck at 1, in an 80 column PET, the test won't detect it as it always sets the odd SRAM bit 0 to 1. I then put a &55/&AA test in which was failing.
 
I have found a much neater (and more comprehensive) test for memory, so I will be changing the page 0 and 1 and video RAM test in Version 5.

Dave
 
And now its crashing to monitor.

PETTEST back in and ROM checksums are showing F800 5960 A425 CF19
So it looks like that the Kernal ROM has now gone south too. I did have to solder legs onto it so it was a bit sick already.
 
Back
Top