• Please review our updated Terms and Rules here

PET 8032 display issue

You will find my PETTESTER and documentation (including the source code) here: https://drive.google.com/drive/folders/1fyLbr1kcG98a2FDOMo1H5pj9lIdJpHcx?usp=sharing

You can modify the source code and assemble it yourself (using an online assembler) should you wish. The information is in the manual.

My PETTESTER is also built into a number of diagnostic cards that you insert into the CPU socket. These (obviously) have a host of other features (such as a NOP generator and RAM/ROM replacement).

Any questions - just ask.

Dave
 
Thanks for that, Dave. It passes the tests - I think! Putting the old SRAM in halts the test at the first step but with the new ones it goes through to the DRAM test.
Seems as soon as I fix one issue with this machine another one crops up. Now the vertical hold has gone and so I get a repeated screen. At one point in the test the screen does go haywire and display random characters which seem to be redrawn so fast the bit patterns change to other characters partway through the raster scan.

This sync issue could be the faulty wires but I'm leaning toward the CRTC due to the randomness and the fact I have new SRAM and the tests pass. My guess is the CRTC is generating the wrong addresses but I don't have a CRTC chip to try (my other PET is a 2001) I'm not sure if putting the scope on it will tell me much since the display is sort of working.

The attached picture looks worse than with the naked eye. You can actually see the "32k dram memory test" several times down the screen but due to the headache-inducing flicker the camera missed it.
 

Attachments

  • IMG_20220311_203840.jpg
    IMG_20220311_203840.jpg
    344.5 KB · Views: 17
The PETTESTER obviously checks that the CPU can write to (and read from) the video RAM Ok, which it obviously can.

This leaves the CRTC and the video address multiplexers (UC8, 9 and 10).

It could be worth quickly checking the TAn signals from the CRTC to the address multiplexers to see if one (or more) are garbage.

The address multiplexers are known to go faulty.

Is the CRTC in a socket or not?

Dave
 
After life getting in the way I've finally got round to looking at this.

Looking at your suggestion, Dave, I put my scope on UC8, UC9 & UC10 and all the outputs seem to be outputting something. Nothing looks stuck.

I'm trying to get my head around what is going on. When I use the PETTESTER ROM the screen is rolling and filled with garbage but the RAM tests seem to be OK.
When I use the original ROM I get some garbage but a solid display. It's the same bit of screen with the same spurious characters.

So, I just tried an experiment. Starting with the initial screen which displays correctly, first I scrolled the screen down and as I did so "ready." was scrolled up 1 line but was replaced by "r@a@y@" and also some of the garbage which scrolled up was also changed. I continued to scroll everything and the screen eventually cleared.
Next, at the bottom left of the now cleared screen I typed a row of "q" characters. As I did so every other one caused a "!" character to be printed in the centre of the screen spaced a single character from the previous i.e.

Code: 
                    ! ! ! ! ! ! ! ! ! !






qqqqqqqqqqqqqqqqqqqq

I pressed cursor down to scroll this and the screen changed to this:

Code: 
                    ! ! ! ! ! ! ! ! ! !





 ! ! ! ! ! ! ! ! ! !
qpqpqpqpqpqpqpqpqpqp

So there appears to be a stuck lower bit in even character positions in certain situations plus some odd character copying to areas of the screen memory. Scrolling this down caused the lower row of "!" characters to scroll up by 2 lines each time so it got further and further away from the "qp" line each time. I'm not sure if this is just one problem or more.

Finally as I typed this up I switched the PET off and on again to check again and now it's dead. It did this before when I was probing and I thought it was me. After faffing about checking everything it worked again so I assumed I had a loose socketed chip. I'll leave it off for a while and try turning it on again later.
 
Robin - I have had my PET intermittently not power on and I have found it is the power connector to the motherboard.

This may not be your problem but it may be worth a try - with the power off, wiggle the power connector on the motherboard end of the wires (at the time I found slightly lifting the connector up a few mm fixed it) and see if that helps.

Of course it may well be more fundamental than that, but it's worth a try.

Colin.
 
Ok, wet bank holiday weekend so I finally have some time to look at this PET again. Power is all good again after a clean of the contacts. Hopefully that was all it was.

I'm getting consistent garbage printed in the lower half of the screen depending on what is appearing in the upper half of the screen. The garbage is every other character and changes consistently with the content of the upper half of the display. For example, typing a row of 1's will display
Code: 
[space]!
and 2's will display
Code: 
[space]"
My suspicions are UC8, UC9 and UC10 as Dave mentioned previously although none of the outputs appears stuck. I'm suspecting one of the chips is outputting when it shouldn't. I don't have any SN74LS157N chips in my spares box so it's going to have to wait until after the holiday.

The other possibility is that it's when the CPU addresses the display RAM. I'm trying to read the schematics but I can't see where this is done so if anyone can explain I'd be grateful.

The final possibility might be the CRTC. From what I understand it generates the display RAM addresses for each frame so it's possible that this isn't working properly and so in the lower part of the screen it is addressing the same part of RAM as the upper part although I'm confused as to why the characters are different. Perhaps there are multiple problems. I don't have spare CRTC chips but I do have an Amstrad CPC or 4 lying around - can a borrow a CRTC from one of those?

Finally, the brightness control pot is on the way out. If I put some slight pressure on it I get a display otherwise I get a blank screen - well sometimes - it seems to be behaving again now it's warmed up. I've had a look and there are no cracked solder joints so it must be internal. Can anyone tell me what is a suitable replacement part? Thanks
 
Yep, all wet in Worcester as well!

Don't forget that the RAM is interleaved on the screen. The first character position on the line is EVEN RAM followed by ODD RAM.

It could be a duff RAM chip or the associated data latch.

Is your 'duff' character the first or second position on each line?

If it was an address line, it shoukd make both odd and even both faulty. You could also have two faults...

Dave
 
daver2 said:
Yep, all wet in Worcester as well!

Don't forget that the RAM is interleaved on the screen. The first character position on the line is EVEN RAM followed by ODD RAM.

It could be a duff RAM chip or the associated data latch.

Is your 'duff' character the first or second position on each line?

If it was an address line, it shoukd make both odd and even both faulty. You could also have two faults...

Dave
Click to expand...
It's the second character so that sounds like the odd RAM addressing. However, I can't fathom what's going on. Looking at the PETSCII table it appears that whatever character I type on an odd RAM address the 2x version of it appears in an odd RAM address lower down the screen. e.g. 1,a,q->!, 2,b,r->", 3,c,s-># and so on. This seems to be a decimal shift but makes no sense from a binary perspective. Just scratching my head over that as a mental exercise!
BTW, the display RAM was bad and I replaced all 4 2114s previously.
 
I would like a little video of what you observe with my PETTESTER ROM.

Otherwise, there are too many variables.

If it passes the initial video tests, the RAM must write and read correctly.

If you 'frame step' the video, does it display correctly though? Version 5 has a much longer delay for this purpose!

Dave
 
I've just put the ROM I burnt previously in and it does nothing so I expect it's gone bad. I can't burn another one as to do the original I had to supply a separate 21V as my TL866 can't do that and my bench supply had to be returned under a safety recall!

When I manage to burn another one I'll update you with the results.
In the meantime I swapped UC6 and UC7 and I get the same fault but different characters. I think that might mean my new RAM is bad.
 
>>> I think that might mean my new RAM is bad.

I would actually swap UC6 and UC7 (ODD) with the other pair (EVEN) to see if you can make your problem move over to the EVEN bank.

Dave
 
After a long, long time I'm returning to trying to fix this PET. On first power up the screen started with some garbage in the same places are before then the whole screen filled with random characters. Subsequent power cycles just kept the garbage to the usual garbage in the lower half of the screen. Following Dave's advice above I swapped UC4+5 with UC6+7 and I saw the similar symptoms. Then after a minute or two the screen filled with garbage again.
Checking the before and after swap screen output I could see the problem shift from the even to the odd columns. After a bit of swapping I found that one of the new RAM chips was the problem. I'm just erasing an EPROM and then I am going to burn a fresh diagnostic ROM and check everything.
 
Fresh PETTESTER ROM burnt and I get the rolling screen again. After a bit of RTFM I realised I had to edit the ROM for the correct CRTC parameters. New ROM burnt and I watched it get to the 32K RAM test. Made a cup of tea, came back and everything's dead. There is no 5V on the board any more. This was the original fault I fixed with new voltage regulators so I'm hoping they are OK and have gone into short-circuit protection mode. Regulated Rectified voltage is getting to them so now have to try to find which component has gone bad.
 
Last edited:
Robin Elvin said:
Fresh PETTESTER ROM burnt and I get the rolling screen again. After a bit of RTFM I realised I had to edit the ROM for the correct CRTC parameters. New ROM burnt and I watched it get to the 32K RAM test. Made a cup of tea, came back and everything's dead. There is no 5V on the board any more. This was the original fault I fixed with new voltage regulators so I'm hoping they are OK and have gone into short-circuit protection mode. Regulated Rectified voltage is getting to them so now have to try to find which component has gone bad.
Click to expand...
With analog regulators they rarely if ever fail, though when the 5V goes to zero, they often get the blame.

When this happens, it usually means there is a short on the 5V rail, and it triggers the regulator to go into a low current shutdown mode. It is actually one of their great features, so you don't end up with things smoking and/or fused pcb tracks.

One clue when they are in this mode, they can get quite hot, even on a large heat sink. The reason is , that although the current is low, the entire input voltage is dropped across the regulator's in-out terminals, so the current x voltage product (=power) can be quite high.

The combination of no 5V output and a hot running regulator has tricked at least one forum member to go ahead and replace the regulator.

The usual cause for a short on the 5V rail is a shorted Tant capacitor, with those though it is unusual for them to have a short that comes and goes. But possibly there could be more than one which has failed.

The main initial thing to check is the input voltage to the regulator, if that is there, and there is no output voltage, its almost certain to be a short on the regulator output, not a defective regulator IC.
 
I checked the board everywhere for shorts and found none. Scratching my head I plugged it back in and there was a brief flash of life from the screen. After a lot of odd symptoms like the screen flashing up occasionally, machine resetting, and no 5V it turns out it is the power connector. It is very flaky and loses a good connection just by warming up. I had suspected this previously so I removed all the contacts, cleaned them, closed them slightly, and put them back but the orange plastic is very brittle and some of the wire clips broke off in the process. If I press down hard on the connector the screen stops flickering and I can get through the dram test without failure.

Is there a suitable replacement available? I don't mind replacing the whole thing including the pins on the board if something modern will make a good solid connection.
 
You must log in or register to reply here.
Back
Top