• Please review our updated Terms and Rules here

Commodore 4032-12 dead screen

Here are my results of the 6502 readings with the multimeter:

AB00 - pin 9 - 2.05v
AB01 - pin 10 - 2.05
AB02 - pin 11 - 2.05
AB03 - pin 12 - 0.82
AB04 - pin 13 - 0.82
AB05 - pin 14 - 0.82
AB06 - pin 15 - 2.42
AB07 - pin 16 - 3.76
AB08 - pin 17 - 0.04
AB09 - pin 18 - 0.04
AB10 - pin 19 - 0.04
AB11 - pin 20 - 0.04
AB12 - pin 22 - 2.32
AB13 - pin 23 - 5.10
AB14 - pin 24 - 2.12
AB15 - pin 25 - 0.00
 
rschissler said:
Here are my results of the 6502 readings with the multimeter:


AB08 - pin 17 - 0.04
AB09 - pin 18 - 0.04
AB10 - pin 19 - 0.04
AB11 - pin 20 - 0.04

AB15 - pin 25 - 0.00
Click to expand...

These address lines look suspicious especial the A15 (Most Significant Bit). This means your PET is executing code in the RAM area and not in ROM. Let us thing about this. In the time being, with power off, ohm the AB15 line to ground in case it is shorted. That would explain everything.
 
I can't get any ohm reading with AB15. I rechecked the voltage on AB15 and it fluctuates between 0.00-0.01 volts, but I guess that is still zero.
.
 
Oops! I just realized that I screwed up taking the voltage readings of AB12-AB15. I forgot that the pins traveled counter clockwise to the next bank, as instead I was reading 40,39,38,etc. So, here are the correct readings for AB12-AB15:

AB12 - pin 22 - 0.04v
AB13 - pin 23 - 3.68
AB14 - pin 24 - 3.68
AB15 - pin 25 - 3.68
 
rschissler said:
Oops! I just realized that I screwed up taking the voltage readings of AB12-AB15. I forgot that the pins traveled counter clockwise to the next bank, as instead I was reading 40,39,38,etc. So, here are the correct readings for AB12-AB15:

AB12 - pin 22 - 0.04v
AB13 - pin 23 - 3.68
AB14 - pin 24 - 3.68
AB15 - pin 25 - 3.68
Click to expand...

OK, much better. It looks like the code is struck in the SEL E (1110) area of ROM. That is the Editor area where the video and other things should be initialized. Perhaps a bad 901499-01 ROM? I think that chip should be on a socket. Carefully reseat the chip to rub off any oxidation. Let me know if I called out the correct chip part number and I will send you a replacement for a quick check if the reseat does not work.
-Dave
 
I found the 901499-01 and reseated it, with no change. All the pins looked pretty shiny, and I didn't really notice any oxidation. I also reseated 901447-10.
 
rschissler said:
I found the 901499-01 and reseated it, with no change. All the pins looked pretty shiny, and I didn't really notice any oxidation. I also reseated 901447-10.
Click to expand...
the 901447-10 is the character generator and you'll need the screen up to see if it is working properly.

Send me a mailing address by private message and I'll burn a 2716 EPROM to replace the 901499-01. Are you lucky to have any other ROMs on sockets?
 
rschissler said:
I don't have any other ROMs on sockets. PM sent.
Click to expand...

Too bad about the socketed ROMs. Most 4000/8000 PETs only have the Editor ROM on a socket. I have burned the EPROM and will mail on Monday.

If the EPROM does not fix the problem, it might be a good idea to have a few dynamic memory chips (4116 or equivalent) on hand to try the piggyback test on the lower bank RAMS (UA5, 7,9,11,13,15,17, and UA19). What you do is piggy back a 4116 on one of the RAM chips at a time making good contact on all the 16 pins. You might need some help as then you would cycle power and look for any better result (some change on the screen or beep, etc). If no difference go on to the next RAM chip until something better happens. If piggybacking one of the chips makes a slight difference or works, then replace that chip. Go on with the test until you tried all 8 memory chips in low RAM. A fault on zero page will cause the PET not to boot.

The piggyback test is not fool-proof as it can not find all errors such as hard shorts, but it may minimize the memory chips that will have to be removed and a new one soldered. It's a better method than blindly replacing all 8 chips although you may wind up doing that.
-Dave
 
Last edited:
"Go on with the test until you tried all 8 memory chips in low RAM. A fault on zero page will cause the PET not to boot."

Ok, thanks. What does "A fault on zero page" mean?
 
It means a RAM problem (stuck or random bit) in the address range from $0000 to $00FF. It's called "zero page" because the "page", the first two digits of the address, are 00.

Immediately following "zero page" is "one page" at $0100 to $01FF but no one calls it that because that's where the 65xx stack resides, so it's the "stack page".
 
KC9UDX said:
It means a RAM problem (stuck or random bit) in the address range from $0000 to $00FF. It's called "zero page"
Click to expand...

KC is correct and the only thing special about page zero is that there are special instructions for the 6502 that imply the address is on page zero and they save time and instruction memory as only two bytes need to be fetched for these instructions. The PET operating system takes advantage of this and stores all the variables and parameters (pointers, etc) it needs to work. If that area of RAM is messed up, nothing much will work and generally you will get the dreaded blank screen.
 
Thanks for the 901499-01 you sent, I'll have to owe you. However, installing it produced no change. I guess next I should get a couple memory chips (I saw some on ebay), and try the piggyback method you described?

If one of the memory chips is bad, nothing will work, right? I guess the hope is that only one of them is bad, as it seems that it would be rather difficult to piggyback multiple chips at a time.

Could there be anything else missed? I've read here about capacitors needing to be replaced, and the crt itself is connected to an entirely different board in the computer.
 
rschissler said:
If one of the memory chips is bad, nothing will work, right? I guess the hope is that only one of them is bad, as it seems that it would be rather difficult to piggyback multiple chips at a time.
Click to expand...

Yes, one at a time and look for something to change. You might consider buying 16 pin sockets to install as you replace the RAM chips to make it easier the next time.

Could there be anything else missed? I've read here about capacitors needing to be replaced, and the crt itself is connected to an entirely different board in the computer.
Click to expand...

If the power supply voltages are correct (+5, -5, and +12V) don't worry about capacitors. Also until you get a good Vertical Sync, don't worry about the CRT board. You need a good computer board booting up and initializing the 6545 CRT Controller. Remember you already replaced the 6545 so it is good but probably not being initialized by the CPU due to bad RAM, bad ROM, etc.
 
I have 16 ram chips, so I guess that means they are 2K each. The numbers on the chips are "ITT 8139" and "4116 3N"

What are the odds that more than one is bad? The point I was trying to make is that the piggybacking seems only feasible if one chip is bad. So, if I try piggybacking each one and get no change, do I then go with the assumption that two or more are bad, and then start unsoldering and replacing them until all 16 are possibly replaced?
 
Piggybacking is almost always successul for this. You're not looking for a complete fix, you're looking for a change. Almost certainly, piggybacking one chip will cause some kind of change. That is a dead giveaway that you found a chip that behaves differently than your test chip. Presuming your test chip really isn't bad, the chip being tested therefore is bad.

Once you replace that chip, either the system will work correctly, or, it will continue to work incorrectly, in which case it is likely you have one or more additional bad chips. Piggyback them one at a time again to find the next one, the same as you did for the first one. Continue the ptocess over and over until you don't observe a change in operation due to piggybacking..

I may have some dead 4116s around here. If I get a chance, I can install several of them on a board and demonstrate finding them by piggybacking.
 
KC9UDX said:
Piggybacking is almost always successul for this. You're not looking for a complete fix, you're looking for a change. Almost certainly, piggybacking one chip will cause some kind of change. That is a dead giveaway that you found a chip that behaves differently than your test chip. Presuming your test chip really isn't bad, the chip being tested therefore is bad.

Once you replace that chip, either the system will work correctly, or, it will continue to work incorrectly, in which case it is likely you have one or more additional bad chips. Piggyback them one at a time again to find the next one, the same as you did for the first one. Continue the ptocess over and over until you don't observe a change in operation due to piggybacking..

I may have some dead 4116s around here. If I get a chance, I can install several of them on a board and demonstrate finding them by piggybacking.
Click to expand...
You do realize that the screen is totally dead, there is nothing, like it doesn't have any power or anything. So, to me, any change would mean something on the screen materializes. That's why I can't understand how piggybacking can work with two or more bad chips, unless I start replacing them as well.
 
Any change of any kind is what you're looking for. In the time it takes to piggyback those chips one at a time, if nothing seems to change, you haven't lost much. :D

If you get multiple good chips, go ahead and piggyback as many as you can. If you get 16 good chips (not a bad idea), piggyback them all, and remove them one at a time until you see a change.

I tend to take the opposite approach. I desolder all the RAM chips, install sockets, and known good chips, then I substitute the old chips back in one at a time until I detect a memory error. Then, I mark the chip bad and put it aside or discard it. I tend to put old, known working, chips back into use. I can get away with this because I can desolder the chips without damaging them. I'm not saying you can't, I hope you can; that's up to you.

I've never messed around with a CRTC PET. Don't they make some kind of sound on startup? Even if you have dead video for reasons other than RAM, maybe you can get the boot code to get far enough (correctly or otherwise!) to make sound.
 
You must log in or register to reply here.
Back
Top