PDA

View Full Version : CBM 8096 dead after long storage



Hedgehog
January 9th, 2014, 01:16 AM
Hi folks,
I've got a CBM 8096 that's been in my attic for the best part of two decades. I decided I was never going to use it again and got it out with the intention of testing it and putting it up for sale. After giving it a couple of days to acclimatise and checking/cleaning connectors I fired it up and... nothing. Not a sausage. I checked there is some power there, and on the tags of the transformer from left to right looking from the front of the case:
10.8v, 6.6v (not connected), 11.2v, 9.2v, 9.2v, 8.2v, grnd, 8.2v (all AC). These read the same if I disconnect the plug from the motherboard.
From a quick bit of research I suspect some of these should be more like 20v. When testing with my meter the voltage does read momentarily higher then instantly drop to the readings above, so it could be that my multimeter is not appropriate for testing this.
There appears to be no life in the monitor, no CRT scanning squeal.
The only components I know of that "go off" with age are electrolytic capacitors - is it worth just replacing all of them just as a matter of course, or is it likely to be something else?

Any ideas greatly appreciated! (it was working when I put it away in the mid-90s).
I don't want to throw money at it to get it working because I have no intention of keeping it, but obviously it'll be easier to sell it on if it's alive.

Thanks,
Rob

billdeg
January 9th, 2014, 05:14 AM
check voltages at the chip level. For example each RAM chip as a 5v pin, you can check similar chips look for variations. You can also re-seat the chips carefully to make sure they're getting a good connection to the motherboard. Check the picture tube for a glow on the inside behind the tube initiation point. Even if the system is not sending text to the screen, there should be a glow coming from the tube.
bd

dave_m
January 9th, 2014, 08:02 AM
on the tags of the transformer from left to right looking from the front of the case:
10.8v, 6.6v (not connected), 11.2v, 9.2v, 9.2v, 8.2v, grnd, 8.2v (all AC).

Assuming your transformer is the 'universal' design, see this link for terminal numbers: http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/8032/8032051-3.gif

AC voltages across pins:

across pins 9 and 11 s/b 21VAC for video board
across pins 4 and 6 s/b 22VAC for +/-9VDC unregulated, center tap pin 5 grounded for +/- 5VDC regulated
across pins 7 and 8 s/b 18VAC for +16VDC unregulated for +12V regulated

Hedgehog
January 9th, 2014, 09:08 AM
Thanks for the replies. I'll have another look this evening, try reseating the chips better (I did give them a quick squirt of contact cleaner and wiggle them about a bit but not much). My transformer doesn't look like the one in the diagram you linked to, dave_m. It's got all eight tags in a row rather than being in two rows of four, but I think they are numbered 4 to 11 in the same way. I was measuring them all relative to the ground pin (5) so that's why they made no sense - I'll do it again. I didn't actually look at the back of the tube to see if the heater was glowing - I expected to hear the squeal if the monitor was powered up. I'll probably be back later after I've had another look at it... thanks for your help.

Hedgehog
January 9th, 2014, 11:54 AM
Just had another quick look at it. The AC voltages seem about right now I'm measuring between the right pins :oops:. The chips have their 5v DC (actually more like 4.8v). With the monitor cover off the tube heater is glowing, so the monitor does have power. I'll try reseating all the socketed ICs when I get a chance - no more time tonight. In the mean time, any other suggestions as to what might have died during its 20 year rest?

billdeg
January 9th, 2014, 12:19 PM
If it worked before, the highest likelyhood is that you have a loose chip that needs to be pressed in more snuggly. Next most likely issue will be a cap along the RAM area, next a RAM or ROM chip will have failed.
bd

giobbi
January 9th, 2014, 01:12 PM
If it worked before, the highest likelyhood is that you have a loose chip that needs to be pressed in more snuggly. Next most likely issue will be a cap along the RAM area, next a RAM or ROM chip will have failed.
bd

I'm wondering if is there an easy way to check if the CRTC chip is working. The last 80xx (8032) I fixed, had a broken CRTC, and the previous owner swears it was ok when he put it away 20 years ago.

Can we assume if there are visible scan lines (*) the CRTC *should* (or at least *could*) be ok? Or, is there any other way to check it without to use a scope or a probe?
I mean, just a quick and dirty test for that people who hasn't a scope or a proble.

--Giovi

(*) -> I mean the diagonal ones usually visible if you set brightness to 100% with the CRT pot.

dave_m
January 9th, 2014, 07:19 PM
Can we assume if there are visible scan lines (*) the CRTC *should* (or at least *could*) be ok? Or, is there any other way to check it without to use a scope or a probe?
I mean, just a quick and dirty test for that people who hasn't a scope or a proble.

--Giovi

(*) -> I mean the diagonal ones usually visible if you set brightness to 100% with the CRT pot.

Giovi,
Yes, if there is a screenful (25) of horizontal retrace lines, then the 6545 CRTC has been properly initialized. If just dark screen then probably the CRTC is not initialized which means the 6502 is not running in program due to problems like bad ROM, bad zero page RAM, bad socket connection, etc.

Hedgehog
January 10th, 2014, 12:45 AM
Wow, pretty active forum, this :) Thanks again for the replies.
I couldn't shift some of the large chips with the trusty biro lid, so I'll have to make up a better tool and try again.
If any of the RAM chips are dead I'll feel a bit of a fool... I got rid of a load of random ICs from that sort of era that were taking up space in my toolbox a couple of years ago, which could well have contained some of those! Oh well...
The only test gear I have now is a multimeter, so if I can't trace the problem using that and guesswork I'll probably just pass it on as a project, which would be a shame when it was working fine when it was put in the box all that time ago :(

dave_m
January 10th, 2014, 07:07 AM
I couldn't shift some of the large chips with the trusty biro lid, so I'll have to make up a better tool and try again.


Are any of the ROM chips on sockets? Use your voltmeter on DC to see if Sync signal has about 1.25 VDC average. This would imply that CPU is running if Sync has a 25% duty cycle. Sync is pin 7 on the 6502.

Does anyone have schematics of the 8096? Is it the same board as the 8032 with an additional RAM board? I am not familiar with it.

daver2
January 12th, 2014, 06:29 AM
Maplin do a cheap-and-cheerful IC extractor tool (FD54J at £2.39 each).

According to Andre's 6502.org - the 8096 was an expanded 8032 with an add-on memory board which was later combined to become the 8296 (http://6502.org/users/andre/petindex/boards.html).

Dave

Hedgehog
January 12th, 2014, 08:31 AM
OK, I've pulled out and reseated all the chips now, still nothing on the screen. Turning the brightness right up makes no difference - absolutely nothing on the screen.
I'm getting 3.6v on pin 7 of the 6502 (assuming they're numbered conventionally anticlockwise from the dot).
I've attached a couple of pictures of the inside of my 8096 - no separate memory board as far as I can see.

Hedgehog
January 12th, 2014, 09:04 AM
Here's a high-res version on Flickr:
http://farm3.staticflickr.com/2823/11911117163_082fedb12e_o.jpg (http://www.flickr.com/photos/52886372@N03/11911117163/)
Commodore 8096 Motherboard (http://www.flickr.com/photos/52886372@N03/11911117163/) by Rob Northcott (http://www.flickr.com/people/52886372@N03/), on Flickr

daver2
January 12th, 2014, 09:16 AM
That looks like a single board with the dual 64K blocks of memory (standard plus expansion) in a line at the bottom of your picture. It is possible that some 8096's from Commodore had a single board rather than an 8032 main board + memory expansion board - or that at sometime in the past it has been repaired and the repairer changed the entire board set for expediency?

The magic assembly number (top edge next to the battery) seems to be 8032030-6 if my eyesight isn't too bad and I can still read upside down! It looks like this to me:

http://zimmers.net/anonftp/pub/cbm/schematics/computers/pet/8032/8032030-6.gif

The fact that pin 7 of the 6502 (yes, you are correct that they are numbered anticlockwise from the dot or indentation) is 3.6 Volts only means that it is probably oscillating and not a static logic level. Unfortunately, inferring the dynamic characteristics of a signal from the integrated d.c. reading of a multimeter is not possible. The characteristics of multimeters are different so the integrated d.c. voltage measurement on one instrument would be different from that indicated on another.

If you measured the integrated d.c. voltage of a signal with a multimeter on a known good PET - and used the same multimeter on the same signal of another PET you may be able to infer that the signal was the same if you got the same reading (assuming the signal was something that was fixed - like a sync pulse - as opposed to a variable signal - such as a data, address or control line).

Some people try a technique to piggy-back a known good chip with a suspect chip to see if that helps. Sometimes it does, sometimes it doesn't and sometimes it may result in further damage. But if you don't have any spare chips then you will have to purchase some to see if this helps isolate the fault or not.

Dave

daver2
January 12th, 2014, 09:24 AM
Looking at your hi-res image - some of the IC pins look decidedly dirty and others relatively clean.

I obtained an Apple IIe a year or so ago that was faulty and had a disk controller card which had some clean ICs and some dirty. The disk controller (and the disk drive as it turned out) was faulty. I could see that the Apple was misreading the PROM contents of the disk controller - and this was an IC with dirty pins. Fortunately, the dirty ICs were in sockets so I decided to remove the chips and clean their pins. When I removed the chips, two of them (the PROM and a 74LS74) left a couple of pins behind in the sockets! The pins had been completely eaten through and the pin was no longer making with the integrated circuit at the body of the IC so it was effectively floating. I replaced the two chips and everything started to work (once I replaced the disk drive). There was also a weird fault on a DRAM chip that remained hidden for a while - but that's another story!

Just a though for you from my experience.

Dave

dave_m
January 12th, 2014, 10:26 AM
I'm getting 3.6v on pin 7 of the 6502 (assuming they're numbered conventionally anticlockwise from the dot).
I've attached a couple of pictures of the inside of my 8096 - no separate memory board as far as I can see.

This looks like Commodore took a 8032-sk board and used 64K X1 RAM chips to make this 8096. It also looks like all the ROMs are on sockets. Is that correct? If so that is a lucky break for replacement.

3.6VDC on the SYNC signal does not seem good. It should be about a 25% duty cycle (about 1.25VDC) to have normal instruction fetching. It appears the CPU may not be running properly (due to reading bad ROM?) although as daver2 mentioned perhaps that can be affected by the type of voltmeter.

Make sure that you have +12V and -5V on the dynamic RAM chips (4164). Without zero page RAM, the computer will not run. If power is OK there, as another step, I can send you replacement firmware using 2532 EPROMs assuming the images are on Zimmers.net somewhere.

giobbi
January 12th, 2014, 12:37 PM
This looks like Commodore took a 8032-sk board and used 64K X1 RAM chips to make this 8096. It also looks like all the ROMs are on sockets. Is that correct? If so that is a lucky break for replacement.

3.6VDC on the SYNC signal does not seem good. It should be about a 25% duty cycle (about 1.25VDC) to have normal instruction fetching. It appears the CPU may not be running properly (due to reading bad ROM?) although as daver2 mentioned perhaps that can be affected by the type of voltmeter.

Make sure that you have +12V and -5V on the dynamic RAM chips (4164). Without zero page RAM, the computer will not run. If power is OK there, as another step, I can send you replacement firmware using 2532 EPROMs assuming the images are on Zimmers.net somewhere.

Dave, it seems the motherboard is using 4116 RAM and it is the 8032030 assy: I think it's a standard 8032 motherboard. I have two of them here (same assy) and they are the same motherboard. Also the romset is the same. Maybe somebody put an 8032 motherboard inside an 8096 case? I can't see how it could have 96 kB...

If you're planning to send him a replacement for the firmware, remember this assy need a fix (patch) or it won't work with the zimmers images (do you remember, Dave? My 8032 didn't work until I patched the pcb: http://www.vintage-computer.com/vcforum/showthread.php?39910-CBM-8032-video-issue).

Since the last 8032 I fixed had a broken CRTC, if I was in your shoes I would try to change the CRTC (UB13): it's easy because it's socketed, but of course it means you must have a replacement there. I found them some months ago on eBay for a little price. The fact you haven't scan lines make me think the issue could be there.

Also: in my very little experience, I fixed more issues changing old sockets than changing ICs :-) -- Commodore sockets from that era are often unreliable.

A question: does it make the "chirp" at boot?

aaah, try to remove the two 6520 and see if it boots: you don't need them to boot (but of course they're needed if you want to use the computer!).

--Giovi

dave_m
January 12th, 2014, 09:43 PM
Dave, it seems the motherboard is using 4116 RAM and it is the 8032030 assy: I think it's a standard 8032 motherboard. I have two of them here (same assy) and they are the same motherboard. Also the romset is the same. Maybe somebody put an 8032 motherboard inside an 8096 case? I can't see how it could have 96 kB…

You are right. This board is just a 8032 with 32K RAM.




Since the last 8032 I fixed had a broken CRTC, if I was in your shoes I would try to change the CRTC (UB13): it's easy because it's socketed, but of course it means you must have a replacement there. I found them some months ago on eBay for a little price. The fact you haven't scan lines make me think the issue could be there.

Yes, but it could also be that the CPU is not running in program properly due to bad ROM or RAM, and so the 6545 CRTC is not being initialized and then there will be no vertical or horizontal timing signals. This will result in a dark screen with a good 6545.

Hedgehog
January 13th, 2014, 01:26 AM
Thanks again for all the replies. I'll test the voltage on the RAM and try cleaning all the pins and sockets a bit better. If it still doesn't work I don't know if I really want to get into buying parts to test it with - I'll probably just pass it on to somebody who wants a project. I really only got it out to assess it prior to selling/moving it on.

giobbi - thanks, I'll try removing those 6520s and see what happens. There is no chirp at boot at the moment - no anything in fact :)

It's interesting that the motherboard looks like an 8032 - I only assumed it was an 8096 because that's what the badge on the case says - I can't honestly remember how much RAM it had. I rescued it when a place I worked at was throwing it away, and really only used it a bit to see if it was working before putting it away (I've got a twin floppy drive and printer as well, which were all working at the time). Perhaps the motherboard had been replaced at some time, or possibly the top part of the case has been swapped? Or the extra RAM board was removed...

daver2
January 13th, 2014, 11:47 AM
I couldn't read the numbers on the DRAM chips so I made the assumption myself that the motherboard contained 96K (as per an 8096).

I agree - it looks like a standard 8032 with the memory expansion module missing (which should fit on J4/J9). I assume that the memory expansion module died at one point and the simple fix was to remove it (or someone had another use for it).

Badges don't necessarily tell the truth... I saw a PET for sale once that claimed to be 'x' but it only showed an outside photograph of the case. If you knew what to look for it didn't actually contain a motherboard! Needless to say the purchaser wasn't too happy with his very expensive case...

Dave

Hedgehog
March 8th, 2014, 10:54 AM
OK, I completely failed to revive it. It's now gone to a new home with daver2, who has some proper test gear and hopefully he'll have more luck with it. I'll keep an eye out to see how it goes. Hope you can get it going Dave.