Repairing an 8032SK... any help?

asbesto · Post by **asbesto** » Thu Feb 23, 2012 11:32 am

Hi there,

this is our second fix attempt, a COMMODORE CBM 8032SK. The problem can be seen at this URL:

http://museum.freaknet.org/index.php/Co ... CBM_8032SK

The CBM power on but it "chirps" only sometimes. When it chirp, a bright line with 40 "-" and 40 spaces is printed at the center of the screen. When it not chirp, nothing is printed on the screen.

We tested the 901447-10 PLA on a 4032 and it works well. The CPU also works and was tested. No other chips are on sockets...

I measured weird TTL levels on various pin of that PLA (about 1.5V) and this sound weird to me. Those weird levels also are present on UA2, the 74166, so I figured it was broken. I changed it, but no luck - and TTL levels there seem to be floating again, also on other chips around it.

More details will follow...

Any idea?

fachat · Post by **fachat** » Thu Feb 23, 2012 5:42 pm

I don't really have an idea right now.

One comment - the 8032 (even the -SK) does not have a PLA. All address decoding etc is done in TTL logic. The 901447-10 is a ROM chip - in this case the character ROM that produces the characters on the screen.

How did you measure the TTL levels? with a multimeter or a 'scope? With a multimeter you don't get real TTL values there, as the signal changes a lot. Oh I see you're using a scope.

But anyhow, the character ROM can not cause such problems. I would guess that either some RAM or I/O chip is broken. Check the DRAM chips for heat. Check the PIA and VIA chips for heat. Also check if the interrupt line is working correctly.

In fact, the chirp and the line on the screen show that the computer at least does some initialization - the chirp is created by the CPU using the VIA I/O chip (does not mean it's ok, it might be partially broken), and the CRTC video chip is initialized by the CPU (so that a picture comes up on the screen, which it does not when not being initialized).

You say that the chirp and the screen happen "sometimes" - what happens in the other cases?

For Commodore PET specific repair help, the vintage computer forum is probably a very good (if not better) place to ask:
http://www.vintage-computer.com/vcforum ... -Commodore

André

asbesto · Post by **asbesto** » Thu Feb 23, 2012 6:54 pm

fachat wrote:

One comment - the 8032 (even the -SK) does not have a PLA. All address decoding etc is done in TTL logic. The 901447-10 is a ROM chip - in this case the character ROM that produces the characters on the screen.

my fault!

I was confusing PLA with ROM

Quote:

Also check if the interrupt line is working correctly.

is the "Phi2" line? It's always 1

Quote:

You say that the chirp and the screen happen "sometimes" - what happens in the other cases?

it does nothing, a blank screen

I will post this also in that forum, thank you anyway!

BigDumbDinosaur · Post by **BigDumbDinosaur** » Thu Feb 23, 2012 6:55 pm

fachat wrote:

But anyhow, the character ROM can not cause such problems.

Also, those would be mask ROMs, which generally do not exhibit problems due to age alone.

Quote:

I would guess that either some RAM or I/O chip is broken. Check the DRAM chips for heat. Check the PIA and VIA chips for heat. Also check if the interrupt line is working correctly.

I'd be casting a suspicious eye on the DRAM.

BigDumbDinosaur · Post by **BigDumbDinosaur** » Thu Feb 23, 2012 6:56 pm

BTW, most of your circuit investigations can be sped along with a good logic probe. As Andre recommends, look at /IRQ for signs of life.

ElEctric_EyE · Post by **ElEctric_EyE** » Thu Feb 23, 2012 7:57 pm

I'm wondering if it uses the same DRAM as in the C-64. I think the part #'s were 4164, 64Kx1. They were very sensitive devices. Back when I was experimenting with hardware and did damage to the 64 my Dad's friend would fix it for me. When I got it back repaired, there was always at least a couple of them blown out. 2nd vote to replace all the DRAM.

BigDumbDinosaur · Post by **BigDumbDinosaur** » Thu Feb 23, 2012 10:10 pm

ElEctric_EyE wrote:

2nd vote to replace all the DRAM.

Especially DRAM that is 30-some years old.

asbesto · Post by **asbesto** » Thu Feb 23, 2012 10:24 pm

Thank you all, I will check that tomorrow! They are 4116. I'm lucky to have a lot of those chips

(I also found a way to change the 4116 with 4164 but seem a bit complicate but interesting: http://www.osiweb.org/osiweb/misc/4116to4164.txt )

BigDumbDinosaur · Post by **BigDumbDinosaur** » Thu Feb 23, 2012 10:45 pm

asbesto wrote:

Thank you all, I will check that tomorrow! They are 4116. I'm lucky to have a lot of those chips

How old are those 4116s? If they are 1980s vintage I'd be a bit leery of using them.

Quote:

(I also found a way to change the 4116 with 4164 but seem a bit complicate but interesting: http://www.osiweb.org/osiweb/misc/4116to4164.txt )

Not something I'd approach lightly on 30 year old hardware. Besides, what would be the advantage to doing so on this particular machine?

asbesto · Post by **asbesto** » Fri Feb 24, 2012 8:42 pm

BigDumbDinosaur wrote:

asbesto wrote:

Thank you all, I will check that tomorrow! They are 4116. I'm lucky to have a lot of those chips

How old are those 4116s? If they are 1980s vintage I'd be a bit leery of using them.

I was wrong! I have TONS of 4164 but no 4116

I'm ordering them from a reseller here in Italy; they have many brands: NEC 100 and 200ns, Texas TMS4116/20JDL, ITC ad Toshiba... any hint about what can be the better one? The Texas seem very old, with the metallic plate on top... they seem cheap, about 1 or 2$.

fachat · Post by **fachat** » Fri Feb 24, 2012 9:35 pm

asbesto wrote:

Quote:

Also check if the interrupt line is working correctly.

is the "Phi2" line? It's always 1 :(

Ah, no, the /IRQ line is separate from Phi2, which is the system clock. Neither of them should be constant 1. Phi2 should show a 1MHz signal (system clock), /IRQ should be something like 50Hz (screen retrace interrupt)

André

BigDumbDinosaur · Post by **BigDumbDinosaur** » Fri Feb 24, 2012 10:21 pm

fachat wrote:

asbesto wrote:

Quote:

Also check if the interrupt line is working correctly.

is the "Phi2" line? It's always 1

Ah, no, the /IRQ line is separate from Phi2, which is the system clock. Neither of them should be constant 1. Phi2 should show a 1MHz signal (system clock), /IRQ should be something like 50Hz (screen retrace interrupt)

André

Asbesto, sounds like you may have a dead clock generator circuit. As André noted, Phi2 should exhibit a 1 MHz symmetric square wave. Without Phi2 being present, the 6502 will not do anything.

/IRQ (Interrupt ReQuest), an active low signal, should exhibit a very short, low pulse at a 50 Hz rate (in your case). This so-called jiffy IRQ is responsible for periodic housekeeping operations that must be completed, such as scanning the keyboard.

You can observe these signals with either a logic probe or 'scope. First thing is to get Phi2 running. Ain't nothin' gonna happen without it.

BigDumbDinosaur · Post by **BigDumbDinosaur** » Fri Feb 24, 2012 10:41 pm

asbesto wrote:

BigDumbDinosaur wrote:

asbesto wrote:

Thank you all, I will check that tomorrow! They are 4116. I'm lucky to have a lot of those chips

How old are those 4116s? If they are 1980s vintage I'd be a bit leery of using them.

I was wrong! I have TONS of 4164 but no 4116

I'm ordering them from a reseller here in Italy; they have many brands: NEC 100 and 200ns, Texas TMS4116/20JDL, ITC ad Toshiba... any hint about what can be the better one? The Texas seem very old, with the metallic plate on top... they seem cheap, about 1 or 2$.

All were about equal in quality. The NEC and Toshibas were second-source for Asian and European distribution. If you can, use the 100ns versions.

BTW, the TI DRAMs may be a military package design for heat sinking. Dunno for sure...I don't recall that stuff anymore.