• Please review our updated Terms and Rules here

Commodore 8032-SK suspected RAM fault

P

Pondis

Member
Joined
Oct 19, 2013
Messages
11
Location
Leicester, UK
Hello all

I posted a few days ago about an 8032-SK that wouldnt turn on at all. Deep cleaning solved that, but now when booting up, it says "b432 bytes free" which someone pointed out isnt correct.

I tried to run a RAM test located here: http://www.commodore.ca/manuals/PET_RAM_test.htm and I got to line 23 when I got ?out of memory errors.

So it sounds straight forward enough, but how do I test to see which RAM chips are bad? Are they cheap enough just to replace the lot? Where would I get them from?

Importantly, as a newbie to this sort of thing, I really only have a multimeter, so is it even possible to do?

Thanks in advance!
 
So I was reading an old thread, and Dave_M (thanks!) suggested typing print peek (1024 +56) and showing the result. Heres mine:

IMG_20131023_185531[1].jpg

Now I am not really any the wiser, but I am hoping you guys can tell what might be going on.

The guy with the issue also said that he took RAM from a ZX spectrum and stuffed it in his 8032 and it worked, does that sound right? I do have a spare ZX spectrum I can cannibalize.
 
Pondis said:
The guy with the issue also said that he took RAM from a ZX spectrum and stuffed it in his 8032 and it worked, does that sound right? I do have a spare ZX spectrum I can cannibalize.
Click to expand...

Pondis you can do a check using the 4116 Ram chips found in the Sinclair spectrum.
It doesn't always work though it's one of those hit or miss things. It's called Piggy Backing!
Basically you put the Ram chip on top of the existing Ram chip with the computer switched off.
Then turn on the computer to see if it's done the trick. You can also physically check each chip
whilst it's on to see if it's running hot which can indicate a shorted Ram chip.
 
I think PET Rescue is right that low RAM is not working right. Problems with zero page and perhaps the stack area is keeping BASIC from running properly. Perhaps you can try and see if the machine language monitor is working well enough to display some RAM locations and then if you can tell what bits are not working to give a clue about what chips of the lower 8 4116's to replace.

Try 'sys 1024' again and then at the dot prompt, type 'm 0404 0460' to look at locations $0404 to $0460. They should have $AA in there. Of course we don't really know what location failed because the b432 on the display is not correct, but give it a shot. You can sample other locations also.
 
Thanks for the advice guys.

I tried the sys 1024 thing and got the following

IMG_20131024_180919[1].jpg

Not exactly what I was expecting, nothing happened. So does that mean the memory locations are dead? Or did I just do something wrong?
 
Pondis said:
Not exactly what I was expecting, nothing happened.
Click to expand...

From the one time you got to a break instruction and got a register display, it looks like the program counter (PC) was in the RAM area. This implies the system is very unhappy. It appears most of lower RAM is not working. If zero page RAM (0000-0100) is not functioning, BASIC will not run (and apparently neither will the monitor) as that's where are the system variables are stored.

We were hoping to get a clue as to which of the eight 4116 (16K x 1) chips in the lower 16K were bad, but did not get a clue.

Try PET Rescue's trick of piggy backing a 4116 chip one at a time on top of UA5, 7, 9, 11, 13, 15 ,17, and UA19 and reseting the PET by power cycle or better yet grounding the /RESET line for a second and hope that you see some sign of improvement. If one of the chip shows some difference then that is a chip to replace. Hopefully you will cure the RAM problem without changing all 8 chips! All this is assuming that the problem is a RAM chip and not a mux or other chip, but it's usually the RAM chips. Best of luck.
-Dave
 
Pondis you are in very good hands with dave_m he is very knowledgeable on these old beasts. If you follow his guidance you won't go far wrong. I wish I had his understanding of how computers worked internally!
 
That sounds like its my weekend planned!

Thanks so much for the help guys. Dave_M I have no idea how you learnt all this stuff, but you are a definitely good to have on side
 
Hi dave I've aquired an old systems control smart-arms robot arm. It is a robot arm fixed to a board with the interface underneath, It's circa 1982.
From the very scarce info I have found it was able to be used on Acorn Atom, zx80/81 and commodore pet.

Not my actual arm but the same model, mine has brown board not white.

http://www.flickr.com/photos/electric-monk/2480961001/in/set-72157604989439887/

2355379387_25fffd9880_z.jpg
 
Last edited:
Pet Rescue said:
Hi dave I've aquired an old systems control smart-arms robot arm. It is a robot arm fixed to a board with the interface underneath, It's circa 1982.
From the very scarce info I have found it was able to be used on Acorn Atom, zx80/81 and commodore pet.

Not my actual arm but the same model, mine has brown board not white.

http://www.flickr.com/photos/electric-monk/2480961001/in/set-72157604989439887/

View attachment 15663
Click to expand...

That arm is the sort of thing I would desperately want, and then when I have it have no idea what to do with it.
 
OK! I think we might be getting somewhere and you might have been right all along Dave.

Piggy backing every chip resulted in the same b432 error every time, though once across 15 I got ?syntax error in 1542yr sp, but it might have been me not setting the other chip correctly and it went back to b432.

Apart from UA19! There I got 72 bytes free. I tried sys 1024 followed by m 0404 0460 and the dot prompt and got this;

IMG_20131026_154611[1].jpg

Which looks promising!

So is it safe just to replace UA19? what else will i need to do to get all my RAM back?
 
Last edited:
Pondis said:
Apart from UA19! There I got 72 bytes free. I tried sys 1024 followed by m 0404 0460 and the dot prompt and got this;

View attachment 15677

Which looks promising!

So is it safe just to replace UA19? what else will i need to do to get all my RAM back?
Click to expand...

72 byes free indicates the first bad location is at decimal 1097 (1025 +72). This equates to hex 0449. Looking at the display at 0449 we see 2A instead of AA. This indicates the MSB (D7) is low when it should be high. I would replace that bit first which means chip UA5, but since piggybacking UA19 (D0) caused this to be seen, that part would be on the list to be changed also.
-Dave
 
Pet Rescue said:
Hi dave I've aquired an old systems control smart-arms robot arm. It is a robot arm fixed to a board with the interface underneath, It's circa 1982.
From the very scarce info I have found it was able to be used on Acorn Atom, zx80/81 and commodore pet.
Click to expand...

That video showed a lot of degrees of freedom on that arm. It seems very sophisticated. Does the PET model interface on the IEEE 488 bus?
 
Pondis said:
That arm is the sort of thing I would desperately want, and then when I have it have no idea what to do with it.
Click to expand...

Pondis that is where I am at no idea what to do next, I've cleaned it up and added a gripper that was missing.
But because I have no power supply or info I'm stuck at the moment.

Glad you are making progress on your pet!

dave_m said:
That video showed a lot of degrees of freedom on that arm. It seems very sophisticated. Does the PET model interface on the IEEE 488 bus?
Click to expand...

The one on the video is run via a microcontroller(pic) using a power mac.

The specs for the arm state it has features that include 8 TTL inputs, 4 relays for "process outputs", "2 kg cm" gripping torque, 200g lifting capacity, and angular resolution of 1/255.

Apart from that dave I'm lost for now but not beaten yet, just need to find some info.
 
Well, I did exactly as you said, replaced UA 5 and UA 19, and got this

IMG_20131027_174425[1].jpg

Dave, you are a scholar and a gentleman and couldnt thank you enough.
 
You must log in or register to reply here.
Back
Top