• Please review our updated Terms and Rules here

Cbm pet 8296-d black screen

Desperado

Veteran Member
Joined
Nov 25, 2017
Messages
6,776
Hello guys,
how are you? I'm back here because i bought 8296-D cbm pet but it doesn't work ....i can see black screen and green led on dual floppy drive.
This is the first time that I have this model....
I tried to write the pla u5 and u6 on two eproms but nothing changes...
What can i do please??
Thanks
 
Hello again my friend.

The methodology for fault finding on the 8296 is very, very similar to that you used for the PET.

Start with the power supply rails, the clocks, the CPU signals (/RESET, /NMI, /IRQ, SYNC) etc.

Dave
 
Hello again my friend.

The methodology for fault finding on the 8296 is very, very similar to that you used for the PET.

Start with the power supply rails, the clocks, the CPU signals (/RESET, /NMI, /IRQ, SYNC) etc.

Dave

Hi Daver2, how are you?
Please what's correct romset for this machine?
Thanks!
 
Hi, please
is there a way to test the outputs of the two pla ue6 and ue6 and see if they are working? Thanks!
 
Please any suggests? I removed 6520 and 6522 chip, i checked all 16 ic ram but i have always black screen :(
 
I burned last version of dave's pettester and i inserted on UD7 socket...i have only this screen...why??
Thanks!
 

Attachments

  • Schermata 2021-12-27 alle 11.54.27.jpg
    Schermata 2021-12-27 alle 11.54.27.jpg
    102 KB · Views: 23
The only schematics I can find are for the 8296 at http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/8296/324644-5of9.gif.

The initial test sets the characters in the video RAM to the sequence @abcdefghijklmno... to test the read, write and display of the video RAM. As the test is ‘stuck’ here, it clearly can’t perform this test correctly.

If you look at the characters displayed on the screen, E has become an A and G has become a C. Likewise, the pattern repeats for M/I and O/K.

When I write the values out in binary (remembering that there is still an odd and even latch) I see that one possibility is data bit 1 from the ODD latch. This would be pin 5 of UC6 or UC7 (on the 8296 schematic). I am not sure if the 8296D has the same IC numbering. They are both 74LS374 octal latches.

This would account for the visual aspect, but not why the test got stuck. This implies that what the test program wrote into memory was not being read correctly.

The other possibility is an address fault on the DRAM multiplexers.

I would check with an oscilloscope to make sure that all of the address and data lines of the DRAM (http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/8296/324644-4of9.gif for the 8296) have oscillating data on them (i.e. not stuck at a ‘0’ or a ‘1’ or have any non-TTL voltages) and see where that gets us.

Dave
 
Sorry sorry!!!
I was wrong, this is the test performed on another 8032 that doesn't work !! I was getting confused ..
 
Ah, well I can fix your other 8032 then!

Odd Video RAM data bit 1 or the corresponding data bus buffer (UC6 or UB6 respectively).

Dave
 
Thanks so much Daver2!!
So can i check every UC6 and UB6 pin with logic probe?
 
No need to check every pin. I mentioned data bit 1 (so look for the symbol OSD1 on schematic http://www.zimmers.net/anonftp/pub/cbm/schematics/computers/pet/8032/8032029-09.gif).

You should check UC6 pin 13 and UB6 pins 4/5. They should be oscillating. I suspect you will find they are stuck at ‘0’.

Check the other side of the buffer (UB6 pins 15 and 16). These should be oscillating. If not, the problem is further back somewhere.

If OSD1 is stuck at ‘0’ we don’t know whether it is the buffer that is faulty or the RAM that is faulty. Toss a coin as to which one to replace!

If the buffer is faulty, it may be always writing ‘0’ into the RAM and the RAM is Ok.

If the RAM is faulty, the buffer may be trying to write ‘1 into the ram on OSD1 (so you will see some oscillating activity) but the RAM may not store it, or may always read ‘0’.

The buffer has both a read and write half, so one side could be dead and the other not. This is why a logic probe doesn’t always tell you the truth, and an oscilloscope is better - we need to look at the data bit in both read and write modes.

Pin 1 of UB6 is low for a read whilst pin 19 of UB6 is low for a write.

Dave
 
Thanks Daver2, so these are results:

UC6 PIN 13= PULSE

UB6 PIN 4-5=PULSE
UB6 PIN 15-16= LOW

Unfortunately now i can't see again pettester characters's screen but i see horizontal green line at a middle screen :(
 
Check UC2 pin 6 (SN7486) for a vertical drive signal. If there is a signal here, the vertical drive circuitry in the monitor has died.

It is interesting that UB6 pins 15 and 16 are low though...

Dave
 
I have noticed though that this board has been modified or repaired .... it has a track at the top cut off and two wires underneath ....
 

Attachments

  • Schermata 2021-12-27 alle 16.21.37.jpg
    Schermata 2021-12-27 alle 16.21.37.jpg
    119.3 KB · Views: 8
  • Schermata 2021-12-27 alle 16.21.56.jpg
    Schermata 2021-12-27 alle 16.21.56.jpg
    92.2 KB · Views: 8
UC2 pin 6 being HIGH indicates the problem is on the logic board and not the monitor.

Check UB13 pins 39 and 40 for signs of activity.

I also made a mistake above when I referenced OSD1. It should have been OSD2, and the pin numbers should be different. But let’s not worry about that now until you get a display again!

I shouldn’t worry too much about the modifications/repairs at this stage.

Dave
 
Either the CRTC is dead or the ROMs are not programming the CRTC correctly (i.e. the CPU/ROM interface is not working.

Is the CRTC in a socket by any chance?

Dave
 
Yes Daver mos6545 is in a socket and i tried to change with another one....but nothing change :(
 
You need to go back to basics then and check the CPU pins - /RESET, /NMI, /IRQ, PHI2 and SYNC for activity.

/RESET should go low at power-up and then high.

/NMI should be high.

/IRQ should be high.

PHI2 should be oscillating (1 MHz).

SYNC should be pulsing to indicate the execution of instructions.

Dave
 
Back
Top