• Please review our updated Terms and Rules here

CBM 8032, working but not

Well, assuming that the screen should be blank then piggybacking UC6 has actually incorrectly added $0F to the correct $20, whereas what we want is to subtract $0F from the slash $2F turning it into $20, no? Or have I got something wrong?

Piggybacking a good RAM on top of a good RAM should do nothing. Piggybacking on a bad RAM might do something (any change is good indication).

@Brian: When the screen alternates / and space, can you tell whether the space is at the beginning or end of the line?

This would be good to know. But it may be hard to tell now that I think about it, as there are blank spaces before and after the actual lines. I'm not sure why I thought it was the odd addresses that were bad. Piggybacking UC4 may give us more info.
 
Okay, taking it from the top...

On startup with the main board untouched, the resulting image is this.
IMG_2995.jpg
- This is also the way the screen looks when a chip is piggybacked on UC4. There is no visible change from the unaltered board.
- powering off/on from this produces a garbage screen for probably a half second, then back to alternating slash-space

On startup with the main board having a chip piggybacked onto UC6, the resulting image is this.
IMG_2997.jpg
- powering off/on from this produces a garbage screen for probably a half second, then back to all slashes.

I put some painters tape on the screen, as you may note, and the photos actually misrepresent what happened because of the angle of the camera, but it seems clear to me that with the unmodified board in the first photo, the last character displayed on each line is a space, not a slash. Whereas with the UC6 chip piggybacked, the last character displayed on each line is a slash. Because of the angle of the camera this is not really apparent in the photo.

Incidentally, this is what was displayed when I tried to piggyback UC6, but I probably did not have all the pins lined up correctly.
IMG_2996.jpg
- added this just for fun, and the off chance that it might indicate something interesting to anybody.


Thanks again for everybody's help!
 
What about the video RAM data bus buffers UB4/UB5 (EVEN VRAM) and/or UB6/UB7 (ODD VRAM)? These are also 4-bit devices.

I am just thinking that you may have two coincident faults - one a RAM fault so the latch gets 0x0 despite the fact that the data bus buffer is faulty and trying to write 0xF into the RAM (instead of the correct 0x0 from the CPU).

Piggybacking the faulty RAM causing it to work now latches 0xF into it resulting in the '/' rather than the expected <SPACE>.

It would be interesting to see what the bit state is coming out of each of the four RAM devices for the given display (for example, if the RAM is outputting 0x20 - but the latch is outputting 0x2F this would indicate a potentially (half) faulty latch).

What do you then get by piggybacking the RAM that gives a screen full of '/'.

Just thinking you may have more than one fault and we need to identify the best way of tracking the first down with some degree of precision.

Dave
 
I had another thought on the drive home...

Put a permanent RESET on the CPU (to prevent it clearing the RAM on power up) and look at the random characters displayed. Form a table of the ODD and EVEN RAMs from the characters you see and look for patterns that are not random (for example, do all the characters displayed in the EVEN RAM positions all have their lower four bits set)? This would be a pretty good indication of faulty video RAM or latch without assuming RAM write or CPU faults etc.

Dave
 
Put a permanent RESET on the CPU (to prevent it clearing the RAM on power up) and look at the random characters displayed. Form a table of the ODD and EVEN RAMs from the characters you see and look for patterns that are not random (for example, do all the characters displayed in the EVEN RAM positions all have their lower four bits set)? This would be a pretty good indication of faulty video RAM or latch without assuming RAM write or CPU faults etc.

Good idea. It will separate the CPU path into the video RAM from the screen refresh logic. It may give us great clues.

Brian: temporarily ground expansion connect J4-pin22, or UD15-pin6 (tack solder a wire to UD15-pin7) to hold reset down. Then post a photo of garbage screen for clues on the even and odd patterns.
 
Good idea. It will separate the CPU path into the video RAM from the screen refresh logic. It may give us great clues.

Brian: temporarily ground expansion connect J4-pin22, or UD15-pin6 (tack solder a wire to UD15-pin7) to hold reset down. Then post a photo of garbage screen for clues on the even and odd patterns.

Sorry but I'm a bit confused about your instructions. Total NEWB here! :D

Are you saying that I should solder a wire from pin6 to pin7 on the UD15 chip?
If that is the case, is that better to do on the top side of the board, or the bottom side of the board? What temp range should I be ideally doing this at so as to not harm the chip or the board?
 
Are you saying that I should solder a wire from pin6 to pin7 on the UD15 chip?
If that is the case, is that better to do on the top side of the board, or the bottom side of the board?

We are trying to temporarily ground the Reset line for the 20 minutes it takes to get photos. So a jumper is needed. The easiest way might be to use a DIP clip, but if one is not handy, get a skinny wire 24 gage or finer (30 gauge wire wrap wire is best), and tack solder it across a convenient place. UD15-pin 6 to adjacent pin 7 is convenient. I would use the legs of the IC itself. Tack solder means to use pin head size drop of solder at low temperature. We are not looking for a great solder joint. If you have jumpers with wire wrap sockets shown below, use the J4 expansion connector to get to the Reset signal (pin 22). The adjacent row should all be ground pins.

MFG_PRT-08430.JPG 70198253_large.jpg DIP Clip
 
Last edited:
In the old days we had large reels of stuff we called 'blue daisy chain' jumpers for use on wire wrap posts. Very handy for connectors with square posts like the PET expansion connector.

The are almost impossible to find now. Two years ago I found a tiny reel of the stuff with 50 daisy chained jumpers for $150(!) from Digi-Key (V01065-ND).

T120-1^50.jpgI found the data sheet in DigiKey. Part is now called Slip-On Contacts from Vector T120-1
 
Last edited:
Those female/female Dupont connector cables you show in post #29 are generally around $1.00 for 40 or so on eBay; I'd order a set of each (F-F, M-F, M-M).
e.g.: eBay 131529665203

But we don't want to wait for eBay, so just wrap a short single-strand wire around J4 pin 22 (4th from the rear, on the inside) and touch the other end to any of the outside pins on J4 when you want to reset or, if you've got a spare 2-pin jumper, stick it on J4-22 and the outside pin beside it to reset.
 
if you've got a spare 2-pin jumper, stick it on J4-22 and the outside pin beside it to reset.

Mike,
That would be the ideal solution for shorting the pins for the 10 minutes needed for the photo. I have a small bag of them somewhere, but can never find them when I need one...:(

5.jpg
 
I found that I already have a few goodies in an Arduino kit that a friend gave to me. I placed a F-F jumper at the J4 location on pin22 (4th one in from the back, on the inside) and just connected it to the adjacent pin22 (4th one in from the back, on the outside) and the screen went blank. All this was done with the UC6 chip piggybacked. After a few minutes just off to the center of the screen it seemed that a few pixels lit up very brightly. Not any kind of character. That lasted for about 15 seconds, then faded to a completely blank screen again. Gave it about 10 minutes total, and the screen stayed blank. Pulled the jumper off, and the screen went back to all forward slashes.
 
I'll own up to that bad idea since I started it (but everyone else followed a foolish man)...

You need the CPU running to initialise the CRTC to get a video signal; but you don't want the CPU running to initialise the video RAM! Catch 22!

OK, forget that bad idea of mine and I will think again...

Dave
 
Back
Top