• Please review our updated Terms and Rules here

DESPERATE CBM 8032 WITH BAD SCREEN

Desperado

Desperado

Veteran Member
Joined
Nov 25, 2017
Messages
6,827
Hi guys, i'm here again with my new Pet 8032.
I bought it from a gentleman who had it in the cellar, apart from a nice crack on the case, it lights up. However it shows a confusing screen with strange symbols. Maybe I'm wrong, but my experience leads me to think of bad video ram ... how about it dear friend Dave? Scope time?
Tks!!
 

Attachments

  • 1672690045964.jpg
    1672690045964.jpg
    201.9 KB · Views: 19
  • 1672690045975.jpg
    1.1 MB · Views: 7
  • 1672690045986.jpg
    2.4 MB · Views: 6
Hi - looks like a fault with the lower nibble even display ram chip (at least) which should be UC4 on the board (2114) - and the display address multiplexor for address line TA4, which is UC9 (74LS157)...

Start by swapping those two and see how it goes!

You will likely have more than 1 of the 4 x 2114 display RAM chips faulty, I would get 4, and some sockets to suit...
 
Last edited:
Radix said:
Hi - looks like a fault with the lower nibble even display ram chip (at least) which should be UC4 on the board (2114) - and the display address multiplexor for address line TA4, which is UC9 (74LS157)...

Start by swapping those two and see how it goes!

You will likely have more than 1 of the 4 x 2114 display RAM chips faulty, I would get 4, and some sockets to suit...
Click to expand...
Thanks so much, i have 4 video ram
 
If the characters are 'static' this indicates either the character generator or the parallel to serial converter on the output of the character generator.

If the characters are 'twinkling' then it could be the video RAM.

If you see static Klingon or Romulan text on the screen you should suspect the character generator (or the socket) first.

What are you running to get that display?

It is possible that there are multiple video faults, so you need to work on one suspected fault at a time - otherwise you will fix one problem, not realise you have fixed one problem and then need to work on another, and then get desperate my friend...

It may be better to post a video of what you can see on the screen to help further diagnose the (initial) fault to work on.

Dave
 
Last edited:
Now i have Pettester inserted but i have same problem also with original Rom.
 
Desperado said:
Yes Dave, characters are twinkling!
Click to expand...
So that would have been a useful thing to have stated in your post... You must state exactly what you can see and hear when describing a fault - not just post a photograph. If things are changing (as they are here) then you either need to describe it or post a video instead of a static photograph.

The fact that you get a different response with the original ROMs and with my PETTESTER leads me to the conclusion that the CPU side of things is working and that the problem is related to the video side.

I don't think (with your current oscilloscope skills) you will be able to diagnose this - so my advice is to replace the video RAM first and see if that fixes the problem. I don't think it will though - but that is an educated guess of course.

If not, we can then remove the video RAM (as you have it now in a socket) and we can force the data lines from the video RAM to test out the video circuitry after the video RAM (e.g. data latches, character generator and parallel to serial shift register).

An 8032 has two sets of video RAM (odd and even) and two video data latches. It us unlikely that both of them have failed in an identical manner - hence the fault may either be further back (the video address multiplexers UC8, UC9 or UC10) or further forward (the character generator or parallel to serial shift register).

I can see a lot of the correct characters on the screen, so poking around aimlessly with your logic probe or oscilloscope probe will probably not yield much in the way of useful information - you need to still learn how to do this...

However, my PETTESTER is not moving on from the video test so there is something likely wrong with either the video RAM, video address multiplexers or video data buffers.

We have to start somewhere though...

Dave
 
daver2 said:
So that would have been a useful thing to have stated in your post... You must state exactly what you can see and hear when describing a fault - not just post a photograph. If things are changing (as they are here) then you either need to describe it or post a video instead of a static photograph.

The fact that you get a different response with the original ROMs and with my PETTESTER leads me to the conclusion that the CPU side of things is working and that the problem is related to the video side.

I don't think (with your current oscilloscope skills) you will be able to diagnose this - so my advice is to replace the video RAM first and see if that fixes the problem. I don't think it will though - but that is an educated guess of course.

If not, we can then remove the video RAM (as you have it now in a socket) and we can force the data lines from the video RAM to test out the video circuitry after the video RAM (e.g. data latches, character generator and parallel to serial shift register).

An 8032 has two sets of video RAM (odd and even) and two video data latches. It us unlikely that both of them have failed in an identical manner - hence the fault may either be further back (the video address multiplexers UC8, UC9 or UC10) or further forward (the character generator or parallel to serial shift register).

I can see a lot of the correct characters on the screen, so poking around aimlessly with your logic probe or oscilloscope probe will probably not yield much in the way of useful information - you need to still learn how to do this...

However, my PETTESTER is not moving on from the video test so there is something likely wrong with either the video RAM, video address multiplexers or video data buffers.

We have to start somewhere though...

Dave
Click to expand...
I tried piggyback on all 4 video ram but i have always same bad screen....
 
Is there a way to test with oscilloscope Uc8-9 and 10?
Thanks so much!
 
Desperado said:
Is there a way to test with oscilloscope Uc8-9 and 10?
Thanks so much!
Click to expand...
There is, but you tried this last time using your oscilloscope probes but gave up. Are you going to have a serious attempt thus time? If so, you need to know how an SN74LS157 works by reading the data sheet. It contains four (4) individual electronic 1 pole 2 way switches. All if the switches have an A and B input and a Y output. Depending upon the state of a 'select pin' (pin 1) either the A input or B input is present on the Y output.

You connect the select pin to channel 1 of your oscilloscope (using this channel as the trigger) and then use channel 2 of your oscilloscope to probe the A and B inputs and Y output of each electronic switch.

In order to do this, you need to understand how the IC works though so, before you do anything, you need to read (and understand) the datasheet for the device you are testing.

UC 4 and 5 are the EVEN video RAMs.

UC 6 and 7 are the ODD video RAMs.

Dave
 
Radix said:
Hi - looks like a fault with the lower nibble even display ram chip (at least) which should be UC4 on the board (2114) - and the display address multiplexor for address line TA4, which is UC9 (74LS157)...
Click to expand...
HI, i changed UC4 and UC9 with two new ics but nothing change :(
 
Ok now i have a working machine! Thanks so much guys!
I changed: Uc4/Uc5/Uc6 2114 video rams!
 

Attachments

  • 1672760930583..jpg
    1672760930583..jpg
    1.2 MB · Views: 6
The picture in post 1 of the pet with almost exact repeating faults is usually caused by faulty video ram chips, as pointed out by Radix above. I would recommend MBH2114 Tesla video ram chips sold on ebay by ELEBOZ a seller in Chech Republic. Any I have bought have all worked 100%, and they are reasonably priced.
 
Desperado said:
Ok now i have a working machine! Thanks so much guys!
I changed: Uc4/Uc5/Uc6 2114 video rams!
Click to expand...
Desperado,

I mentioned on one of your other threads recently, that while I'm not keen on the idea of replacing IC's off the bat, in an attempt to repair faults, it does appear that the 2114 IC in the PET turns out to feature in most repairs, when there is a defect that suggests it could be a problem in the video RAM area.

Since there are a small number of 2114's in the PET, I think it is reasonable in these cases to simply remove the originals, socket them and fit known good IC's.

It would be a good time saver and also it eliminates the most common known cause of trouble in the video ram system. It is still possible of course to have other faults in the video RAM's control & interface system, at least if the video RAM is known good, it helps to track any remaining faults down with the scope.

I bought up a large number of different brand 2114 (and 4116 DRAM) IC's for testing, and have testers for them and have screened all of the stock, and even in these new old stock parts, the occasional one in 20 or 30 or so is defective . They certainly appear to be fragile and quite vulnerable to failure, perhaps more than any IC type I know.

It is interesting how the 2114's (and 4116's) fail, simply because they are aged rather than related to running time. Many old PETs that have been unused/un-powered for decades and when powered up, have defective 2114 and 4116 chips in them. Occasionally a VIA or PIA fails , yet mostly, the 74 series logic chips are ok, with the occasional failure and surprisingly, given the complexity, the CPU's appear generally robust.
 
Hugo Holden said:
Desperado,

I mentioned on one of your other threads recently, that while I'm not keen on the idea of replacing IC's off the bat, in an attempt to repair faults, it does appear that the 2114 IC in the PET turns out to feature in most repairs, when there is a defect that suggests it could be a problem in the video RAM area.

Since there are a small number of 2114's in the PET, I think it is reasonable in these cases to simply remove the originals, socket them and fit known good IC's.

It would be a good time saver and also it eliminates the most common known cause of trouble in the video ram system. It is still possible of course to have other faults in the video RAM's control & interface system, at least if the video RAM is known good, it helps to track any remaining faults down with the scope.

I bought up a large number of different brand 2114 (and 4116 DRAM) IC's for testing, and have testers for them and have screened all of the stock, and even in these new old stock parts, the occasional one in 20 or 30 or so is defective . They certainly appear to be fragile and quite vulnerable to failure, perhaps more than any IC type I know.

It is interesting how the 2114's (and 4116's) fail, simply because they are aged rather than related to running time. Many old PETs that have been unused/un-powered for decades and when powered up, have defective 2114 and 4116 chips in them. Occasionally a VIA or PIA fails , yet mostly, the 74 series logic chips are ok, with the occasional failure and surprisingly, given the complexity, the CPU's appear generally robust.
Click to expand...
Hi Hugo, yes i agree with you!
over time I learned that most of the ics MOS they go bad quickly.... also in this case, the video rams were Mos....I replaced them with the European Tesla...
 
Desperado said:
Hi Hugo, yes i agree with you!
over time I learned that most of the ics MOS they go bad quickly.... also in this case, the video rams were Mos....I replaced them with the European Tesla...
Click to expand...
Hi there, the ITT 4116 ram chips are probably the worst of all - I ended up with a 32KB static ram adapter board to replace the whole lot in one go - far quicker and more productive.
 
You must log in or register to reply here.
Back
Top