PDA

View Full Version : Wounded pet



RetroSweden
January 15th, 2014, 10:29 AM
Hello VC!

I'm newly registered on this forum, but i have been lurking the commodore section for a few months. I must say that you do a great job in supporting people with their computer problems!

I've been struggling with getting a CBM 3032 working for a few months, and i'm getting closer to actually getting it working. It's actually booting into BASIC 2 now. However, it seems as if it's having a bit of trouble spelling out the words.

The AC/DC bridge has malfunctioned for the previous owner, which resulted in a few copper traces burnt off around the diodes. New diodes had been fitted, but i think a couple of mistakes were made when repairing it, leaving the drams out of power. That should be fixed now since i'm able to boot it up. And i'm getting a screen full of "g"s when running the PETTESTER diagnostic ROM, which i previously didn't.

Character ROM has been tested in another 3032, along with the other ROMs, and they work OK.

I've borrowed two 2114 static rams from a working VIC 20, so they should also work OK.

The four 4116 drams far to the right has been replaced by four new ones.

All 74LS244 IC's has been replaced.

I have replaced more than the ones listed above, however i fail to remember what :-)

This is what the screen usually say after boot now.
http://imgur.com/ePfGmVq

Once about every fifth boot this happens instead.
http://imgur.com/mVyMr7O

Any help appreciated
// RetroSweden

giobbi
January 15th, 2014, 11:57 AM
I had the same problem with my 2001/8. See the picture (and don't question about the green screen: I was using a 3032 monitor! )
16870

After many nights replacing ICs without to get the rid of it, I discovered it was a bad socket problem. I'm totally sure at 90% ;-) it was the char ROM socket.
(I discovered it because while I was looking at the traces, I passed a finger on the solder joints, under the pcb, and suddenly the chars changed!)

Actually I don't know how many bad sockets I got until now, but I fixed more issues replacing sockets than replacing ICs :-)

Giovi

RetroSweden
January 15th, 2014, 12:31 PM
Thanks for the quick answer! I'll go look for a socket that will fit and let you know the result. Wish me luck!

RetroSweden
January 15th, 2014, 01:27 PM
OK, i didn't find a brand new socket, but i de-soldered one off a spare c64 main board and replaced the socket on the 3032 mainboard, and now i'm getting this http://imgur.com/ePfGmVq
I'm glad that the 2nd screen i posted seems to have disappeared. I restarted it about 15 times and it stayed the same. Perhaps i should replace the other sockets aswell..

tezza
January 15th, 2014, 01:28 PM
Thanks for the quick answer! I'll go look for a socket that will fit and let you know the result. Wish me luck!

Bad pet sockets have also tripped me up. Check the sockets holding the 2116s and around that video RAM circuitry.

Tez

RetroSweden
January 16th, 2014, 11:57 AM
I've replaced all the ROM sockets today (except for the unused ones). it didn't change much. Also i see that the "(" signs has come back, as in the second picture of my first post.

I'm also detecting a pattern. when the picture looks like the first picture and i write the alphabet on the keyboard, it will show as this:
ABCDEF@ABCDEFPRSTUVWPQRSTU...

And when it boots like the second picture:
HIJKLMNOHIJKLMNOXYZ─Í┼XYZ┼─Í

Soit looks like some address line isn't doing what it should.. maybe?

RetroSweden
January 23rd, 2014, 05:51 AM
I noticed that i had written the problem types wrong.

Problem type 1: @ABCDEFG@ABCDEFGPRSTUVWPQRSTU
Problem type 2: HIJKLMNOHIJKLMNOXYZ─Í┼XYZ┼─Í

So let's do a comparison between @ and H that should have followed after G in problem type 1.

Char Hex Bin
@ 40 0100 0000 <- was this
G 48 0100 1000 <- should have been this

And comparing X with the letter P,that should have followed after the second O in problem type 2.

Char Hex Bin
X 58 0101 1000 <- was this
P 50 0101 0000 <- should have been this

So i think we can point this problem to a specific address now. Faulty bit marked in red.

We can conclude that the address line isn't stable, since it's set to either 0 and 1 during boot, causing either first and third(problem type 2), or second and fourth(problem type 1) byte from the character ROM to be displayed incorrectly.

I'm not at home at the moment so i cannot look at what chip/trace that could be broken.

Anyone have any suggestions?

Best regards
RetroSweden

RetroSweden
January 24th, 2014, 11:05 AM
Bad pet sockets have also tripped me up. Check the sockets holding the 2116s and around that video RAM circuitry.

Tez

Hi Tez,

Earlier i've replaced the soldered 2114's and the four 4116's far right and instead soldered sockets + replacing them with tested chips. So the problem must be somewhere else, but at the moment i don't have a clue where to start looking.

Best regards
RetroSweden

giobbi
January 28th, 2014, 07:42 AM
Just in case: have you an eprom programmer? In other words, can you program an eprom with the pettester image (look here (http://www.verrua.org/pet_repair/index.html) for more details) and with the original ROM images ?

RetroSweden
January 28th, 2014, 09:31 AM
Just in case: have you an eprom programmer? In other words, can you program an eprom with the pettester image (look here (http://www.verrua.org/pet_repair/index.html) for more details) and with the original ROM images ?

Hello Giobbi,

Yes, i've tested it, and it will type out the alphabet just like when i type it in the basic-prompt. i.e. @ABCDEFG@ABCDEFG(..), or sometimes HIJKLMNOHIJKLMNO(..), so it's just repeating the even or odd bytes in the character ROM.


//RetroSweden

KC9UDX
January 28th, 2014, 03:27 PM
My 2001N-32 did this.

17017

I replaced all the ROM sockets, and a handful of non-ROM chips. But, nothing I did cured the problem. Every day it gradually got better and better until the point where it just plain stopped doing it. I initially suspected power supply noise, but the power was clean. The only thing I can say is it must have been a resistive path in a logic gate somewhere, because I can simulate the problem by floating logic low signals in the display circuitry.

The really odd thing is just the other day my C64C did it. I've never before seen a C64 do anything of the sort. It only did it once upon power-up.

17018

MikeS
January 29th, 2014, 09:57 AM
Hi Tez,

Earlier i've replaced the soldered 2114's and the four 4116's far right and instead soldered sockets + replacing them with tested chips. So the problem must be somewhere else, but at the moment i don't have a clue where to start looking.

Best regards
RetroSwedenJust in case, try swapping the two 2114s.
You're looking at screen codes so the upper nibbles are zero, but you're right, bit 3 appears to be stuck low for some reason. Assuming you don't have a 'scope, try checking continuity between these IC pins:
E8-8 and F8-11
E8-8 and F9-17
F9-16 and F10-2

RetroSweden
February 11th, 2014, 12:34 PM
Just in case, try swapping the two 2114s.
You're looking at screen codes so the upper nibbles are zero, but you're right, bit 3 appears to be stuck low for some reason. Assuming you don't have a 'scope, try checking continuity between these IC pins:
E8-8 and F8-11
E8-8 and F9-17
F9-16 and F10-2

Thanks for the suggestions. I tried swapping the 2114:s, and it looks about the same. It didn't change the repeating byte problem. Im running the PETTESTER kernal now, just to see the problem more clearly.

Continuity looks OK as far as i can see. I'm suspecting a bad chip.

Best regards
RS

RetroSweden
February 11th, 2014, 12:37 PM
My 2001N-32 did this.

17017

I replaced all the ROM sockets, and a handful of non-ROM chips. But, nothing I did cured the problem. Every day it gradually got better and better until the point where it just plain stopped doing it. I initially suspected power supply noise, but the power was clean. The only thing I can say is it must have been a resistive path in a logic gate somewhere, because I can simulate the problem by floating logic low signals in the display circuitry.

The really odd thing is just the other day my C64C did it. I've never before seen a C64 do anything of the sort. It only did it once upon power-up.

17018

That's really strange. In one way i'd be glad if this problem disappeared from my PET, but then i wouldn't know what the problem is the next time it appears. I hate not knowing :-)

RS

RetroSweden
March 7th, 2014, 02:13 PM
Just in case, try swapping the two 2114s.
You're looking at screen codes so the upper nibbles are zero, but you're right, bit 3 appears to be stuck low for some reason. Assuming you don't have a 'scope, try checking continuity between these IC pins:
E8-8 and F8-11
E8-8 and F9-17
F9-16 and F10-2

I've bought myself a scope now. Any hints on where to start looking? And what to look for? I'm reading about troubleshooting with scopes, but i've got a long way to go to get past the noob phase.

//RS

carlsson
March 10th, 2014, 12:40 AM
I confess to being the previous owner of this PET, and the one who replaced the diodes. I was meaning to provide some support during the fall, but I got occupied with other matters all the time. :-(

Gabriele72
March 10th, 2014, 08:59 AM
I've bought myself a scope now. Any hints on where to start looking? And what to look for? I'm reading about troubleshooting with scopes, but i've got a long way to go to get past the noob phase.

//RS

I would monitor supply voltages for a while, then, if OK, check the data lines from the video RAM and char ROM.
Trace the signals from the ICs back towards the CPU, to see if there are stuck lines or some bad support logic.
It's just a quick check to begin with.

daver2
March 10th, 2014, 11:26 AM
The pattern appears to be data bit 3 from the video ram through the 8-bit latch to the character generator.

In the first instance data bit 3 appears to be jammed low (or at least never gets driven high) thus the character 'O' (hex code = 4F binary code = 0100 1111) becomes a 'G' (hex code = 47 binary code = 0100 0111). The rest of the faulty characters appear to follow this rule (although I haven't checked out the '&' after the READY).

In the second case, data bit 3 appears to be somewhat erratic. The code for a SPACE (hex 20 binary 0010 0000) seems to become an '(' (hex 28 binary 0010 1000). So I would concentrate specifically on the schematic diagram concerned with this specific data bit.

If you have checked the static ram, the character generator and the sockets - the only thing effectively left is the 8-bit latch or the PCB tracking.

You might also like to consider using a hair dryer and a can of freezer spray. This should identify a faulty chip or suspect socket. Heat up a small section of the board using the hair dryer and then spray one of the chips with the freezer spray to rapidly cool it down. Look for changes in the screen display.

Dave

RetroSweden
April 6th, 2014, 05:03 AM
It turns out that the trace between UE7-11 and UE8-11 had an intermittent connection. I've soldered a wire on the back of the pcb, and after that the pettester program lists the complete alphabet.

Though now it won't boot correctly. Hopefully it's just a bad solder under one of the rom sockets i've replaced. In that case it'll be an easy fix.
the ### Commodore ### line won't show, instead directly in the top left corner "READY" will be displayed, with the cursor blinking over the "R".

if i press return, the computer crashes to TIM.



I confess to being the previous owner of this PET, and the one who replaced the diodes. I was meaning to provide some support during the fall, but I got occupied with other matters all the time. :-(
No worries, you gave me the chance to practice troubleshooting, and that's worth alot! :-) Since then i've also bought myself an EPROM programmer + Eraser, a scope, a buttload of EPROMS, a lot of 74** spare chips so i'll be better equipped for the next challenge!
//RetroSweden

RetroSweden
April 6th, 2014, 01:02 PM
I got it to boot properly after pressing on a few chips. I probably made a bad job at one of the chips i've socketed (which could be almost any chip on the mainboard, since there are only a few soldered ic's left). I'll have to wiggle the chips to identify which one i screwed up.

If i want to upgrade to basic 4, what kind of eproms will i need? I've got several HN462532G's. Will that do? I've programmed a basic v2 kernal with success on that kind of chip.

//RetroSweden

carlsson
April 7th, 2014, 01:40 AM
Great to hear you're getting some life out of this one!

As far as I know, the 2532 chips will do well. You probably need the full set of BASIC, EDITOR and KERNEL, if I understand correctly five chips of 4K each.