• Please review our updated Terms and Rules here

Just broke the PET

...From my look it seems that 40 column mode only uses the two even RAM chips, but having all four 2114 chips should not create a problem if all the jumpers are set properly since the odd chips will never be enabled.

This Commodore control logic is very confusing. Mike please double check when you get a chance.
I think you're right and that the 5 extra chips don't affect 40col operation, but since the OP was having problems I thought it might be a good idea to completely restore it back to a FAT40 and start from there. I'll try to find a little time and check to make sure.
 
No, just swapping the jumpers and the EPROM gives pretty well exactly the display the OP shows in post #4. If you're going to leave the extra 5 chips installed, try disabling the Odd video latch UB8 ('373) by cutting the trace to pin 1; works for me.
 
Just checked and Pin 1 of UC3 is at grd so looks to be correct if I am reading the schematic right. I'll pull that IC and run a vector check on it to see if it is good or failed itself. Thanks so much will update as I continue.
OK, you are on the right trail.

I don't know why Commodore needed so many jumpers and convoluted logic to switch between a 40 column mode with 1K video RAM and an 80 column with 2K RAM.
 
im guessing you found my problem. (hence why I was looking for instructions in the other thread) I did not remove anything just changed jumpers.
Sorry; I really have to remember to test assumptions. Until you mentioned even & odd I'd assumed that when you went 40>80 you'd put the added chips into sockets and then also removed them again when you switched the jumpers and ROM back.

Hope you get that new problem sorted.
 
OK, you are on the right trail.

I don't know why Commodore needed so many jumpers and convoluted logic to switch between a 40 column mode with 1K video RAM and an 80 column with 2K RAM.
Things are not always as simple as they seem, nor can we always appreciate saving 25 cents production cost.

How would you go about it?
 
No, just swapping the jumpers and the EPROM gives pretty well exactly the display the OP shows in post #4. If you're going to leave the extra 5 chips installed, try disabling the Odd video latch UB8 ('373) by cutting the trace to pin 1; works for me.

I think I see what you mean, the LSD bus going to the character generator is still enabled for all RAM even if the jumpers disable the odd RAM to the data bus. The guy who designed this control logic should be fired. :)

With all the jumpers used, why wasn't one used to control pin B8 pin 1 for 40/80 mode?

Are the five chips on sockets? If not, your mod would be ideal for being able to switch the hardware back and forth.

I take it you modified one of your PETs as an experiment. You are good!
 
...With all the jumpers used, why wasn't one used to control pin B8 pin 1 for 40/80 mode?
Well, in the FAT40 UB6, 7, & 8 and UC 6 & 7 are not installed so it doesn't matter, and I imagine Commodore wasn't too concerned with folks who wanted to convert the more expensive 8032 into a cheaper FAT40.

Are the five chips on sockets? If not, your mod would be ideal for being able to switch the hardware back and forth.
No, the unit I'm playing with is an 8032 with the chips soldered in, but I'd made the jumpers easy to change so it only takes a minute to switch from 80 to 40; I just never found (or bothered to patch) a 40 column CRTC ROM for a business keyboard.

I did this many years ago, and at the time just paralleled an external graphics keyboard, which took care of the problem and also had all the keys in the right place. I'd always intended to make it switch- or even soft-selectable but never got around to it (and just got a real FAT 40 instead ;-) )

Someone on here has done a switchable 40<>80 adapter though, but I don't recall who or whether it's written up anywhere.
 
Well, in the FAT40 UB6, 7, & 8 and UC 6 & 7 are not installed so it doesn't matter, and I imagine Commodore wasn't too concerned with folks who wanted to convert the more expensive 8032 into a cheaper FAT40.
There couldn't have been more than a few bucks difference between them. Did old Commodore charge a lot more for 8032 than for a 4032?

A more elegant design could have done both jobs without all the jumpers. Heck the labor cost of inserting the umpteen jumpers, if they were done by hand, would have been pretty costly.

But I know what you mean, large production commercial solutions need to count every penny.
 
OK, made some good progress. Replaced a ton of crap from shorting the board to ground and am now at this point. Working but shows on 39 bytes available. Looking for if I am correct that it just needs some memory chips now 4116's or is there usually another problem causing this? Also the cursor sometimes doesn't blink correct or pauses off.

thanks

39.jpg
 
Last edited:
Congratulations, looks like you're getting there.

I would say this would be caused by one or more faulty memory chips. I've had similar issues with my System 80 and OS Challenger 1.

Tez
 
Congratulations, looks like you're getting there.

I would say this would be caused by one or more faulty memory chips. I've had similar issues with my System 80 and OS Challenger 1.

Tez

Thanks can't wait to get this checked off, and my parts supply for the 2001 to get here so I can get it up and going also. I'll socket the memory tomorrow, I only have 4 new chips so hopefully some of what i pull will seat in a socket.
 
OK, made some good progress. Replaced a ton of crap from shorting the board to ground and am now at this point. Working but shows on 39 bytes available. Looking for if I am correct that it just needs some memory chips now 4116's or is there usually another problem causing this? Also the cursor sometimes doesn't blink correct or pauses off.

Yes, I agree with Tez, it is bad RAM in the lower 16K. the memory tests starts at the beginning of basic 0401H and tests until it finds a bad cell. It did not get very far (39 bytes) which indicates there may also be bad memory in zero page (0000-00FF) which would effect operating system variables.

Run the machine language monitor (SYS 1024). Examine and read/write some low memory locations with the .m command and see if there are stuck bits so you know what exact chip to replace. Use test patterns like 55H and AAH
 
For lower RAM how many bytes should be displayed? If just acting like a 16K machine. making progress but still some issues.
 
Back
Top