Oric Atmos problems

If you want to ask questions about how the machine works, peculiar details, the differences between models, here it is !
How to program the oric hardware (VIA, FDC, ...) is also welcome.
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Oric Atmos problems

Post by Dbug »

Make sure the keyboard connector is not shifted by one pin, one side has one less pin, so it's easy to get it wrong.
RetroUser
2nd Star Corporal
Posts: 22
Joined: Sat Jun 25, 2022 9:14 pm

Re: Oric Atmos problems

Post by RetroUser »

Think I've found the problem and it's a bit of bad luck. Checked the VIA socket for continuity and found pin 32 of the VIA socket is not connected. The picture below is a close up and I've highlighted the problem pin. Looks like the solder didn't make it to the pin or the pin is just a little bit short. I'll try reflowing tomorrow.
VIA socket.png
Pin 36 looks like the same problem but I've checked continuity and it has good continuity with the CPU.
RetroUser
2nd Star Corporal
Posts: 22
Joined: Sat Jun 25, 2022 9:14 pm

Re: Oric Atmos problems

Post by RetroUser »

An almost there update. The keyboard problem turned out to be with the keyboard cable as some of the wires had detached from the keyboard PCB. Didn't take too much before the cable came away from the PCB. Trying to bare the wires on the cable was a nightmare and in then end had to splice a ribbon cable on to the old cable to restore the length so the cable could fold back on itself when plugged into the main PCB.

I ended up removing the VIA socket as it looks like the board is slightly warped in that corner. Put a new socket in pin by pin checking continuity as I went along. Took quite a bit of time to solder the socket.

The now computer passes all of the test ROM tests and I've returned the Oric ROM and reassembled the computer. However, the computer case doesn't go together cleanly anymore. Looks like socketing the VIA is preventing the case from being screwed together properly. If the case it screwed down this results in holding down the space key and spaces are input as cursor moves and can hear the keyboard click sound.

With the case screws loose the computer isn't stable. Sometimes when the computer is turned on a £ sign is displayed automatically under the Ready prompt. This can be deleted. Typed in the usual 10 PRINT "HELLO WORLD" 20 GOTO 10 program and run it but it's unstable. Sometimes it doesn't print HELLO WORLD and may get HELLO, HEL, HELLO W, etc. The program sometimes crashes with a syntax error in one of the lines and one of the lines of code is corrupt, e.g. 20 GOTO 10 becomes 20 10 . Sometimes the program stops printing HELLO WORLD but is still responding to keypresses as can hear key click when a key is pressed. Sometime the computer freezes and has to be turned off and on. I think maybe it's the case causing the problems by pushing on the PCB, need to do some more investigation.
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Oric Atmos problems

Post by Dbug »

RetroUser wrote: Thu Aug 11, 2022 9:25 pm With the case screws loose the computer isn't stable. Sometimes when the computer is turned on a £ sign is displayed automatically under the Ready prompt. This can be deleted. Typed in the usual 10 PRINT "HELLO WORLD" 20 GOTO 10 program and run it but it's unstable. Sometimes it doesn't print HELLO WORLD and may get HELLO, HEL, HELLO W, etc. The program sometimes crashes with a syntax error in one of the lines and one of the lines of code is corrupt, e.g. 20 GOTO 10 becomes 20 10 . Sometimes the program stops printing HELLO WORLD but is still responding to keypresses as can hear key click when a key is pressed. Sometime the computer freezes and has to be turned off and on. I think maybe it's the case causing the problems by pushing on the PCB, need to do some more investigation.
Looks like you have some intermittent contact somewhere, I'd try with the PCB flat on something like a soft mouse mat, and try to push on some of the components to see if you can reproduce the behavior.

Could possibly also try to reflow all the solders on the large components and connectors?
RetroUser
2nd Star Corporal
Posts: 22
Joined: Sat Jun 25, 2022 9:14 pm

Re: Oric Atmos problems

Post by RetroUser »

Tried the PCB out of the case and unplugged from the keyboard. Tried powering up the computer and get the following line displayed after 27-30 seconds (no pressing of board).

corruption.png

The line doesn't appear in the same place, sometimes it's above and below shown in the picture. Tried this six times and it appeared four out of six times. The other times the cursor disappeared and pressing the NMI button reset the computer so the computer was responding to NMI. The other time the cursor moved down a line and to the right a bit but kept flashing.

Noticed that the AY-3-8912 and RAM get hot really quickly. Not too hot to touch but noticeably hotter than the CPU, VIA and ULA. The CPU, VIA and ULA are cool in comparison. Wonder about the capacitors as did test a few and found one of the 47nf ceramic capacitors was reported as 71nf on my tester (that said, it's not an expensive tester so could be misreading). Tried a new 47nf ceramic capacitor in my tester and it was reported as 48nf so replaced the old capacitor with the new capacitor. Worth a recapping the entire board?
User avatar
Chema
Game master
Posts: 3013
Joined: Tue Jan 17, 2006 10:55 am
Location: Gijón, SPAIN
Contact:

Re: Oric Atmos problems

Post by Chema »

Don't you think this looks more and more like a RAM issue?

If random fails occur with time, it might not be correctly refreshed. The ULA is is charge of that.

If the RAM is not fast enough (the Oric needs faster ram than other 8bit systems) this could create problems too.

Recapping is a good idea, but I hardly ever saw an Oric needing that to work. Mine has the original capacitors.
RetroUser
2nd Star Corporal
Posts: 22
Joined: Sat Jun 25, 2022 9:14 pm

Re: Oric Atmos problems

Post by RetroUser »

It does look like it's something to do with the RAM. I replaced all of the RAM with 4164-15 which were bought from a reputable seller. I tested the RAM before installing the ICs and all was good. However, I guess the tester only checks it can read/write and doesn't check speed or maybe one of the 4164-15 RAM ICs was marginal as Adrian Black would say. Put the test ROM back in and didn't show any problems except later on when running the AY-3-8912 tests where characters changed randomly like Testing became T%sting whilst I was watching the screen. Could the RAM IC in IC13 (bit D6) have been flakey?

The irony is I put back all of the old INMOS2600 RAM ICs that were on the board except the one my tester found was faulty. To make up the last IC, I inserted a 4164-12 (only have six of these so couldn't replace all of the RAM). Powered up the computer and a { character appeared at the prompt like someone had typed in {. Reset the computer and left the computer on for 30 seconds but nothing appeared on the screen. Waited a bit longer and all seemed well. Have the computer on at the moment running a simple test program. The AY-3-8912 gets hot to the touch, a lot hotter than the other ICs, is this normal? Have checked the AY-3-8912 connections for any shorts but couldn't find any.

Did end up replacing a couple of 47nf capacitors because the packaging was coming away from the legs and they looked a bit dodgy. One of the capacitors broken (leg fell off) as I was removing it. The existing capacitors should be good but I guess with me working on the board and don't know who had worked on the board (computer arrived with case screws missing) the other capacitors had been moved too much putting stress on the legs/packaging.

Forgot to add, also changed R6 as the packaging near the C23 had come away from the leg exposing the end of the resistor.
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Oric Atmos problems

Post by Dbug »

RetroUser wrote: Sun Aug 14, 2022 1:38 pm later on when running the AY-3-8912 tests where characters changed randomly like Testing became T%sting whilst I was watching the screen. Could the RAM IC in IC13 (bit D6) have been flakey?
I've rarely seen rams have wobbly values that change randomly.

Indeed "e" is 01100101 and "%" is 00100101.

What about running a small BASIC test program:

Code: Select all

10 PRINT"FILLING RAM"
20 FOR A=#600 TO #9800:POKE A,255:NEXT A 
30 FOR L=0 TO 9999
40 PRINT"TEST PASS #",L
50 FOR A=#600 TO #9800:IF PEEK(A)<>255 THEN PRINT "FOUND ";PEEK(A);" AT ADDRESS";HEX$(A)
60 NEXT A:NEXT L
Each pass takes about 5 minutes, but basically if any value between #600 and #9800 (just before the charsets) changes, you'll see that.
User avatar
Silicebit.
Flight Lieutenant
Posts: 313
Joined: Thu Jan 12, 2006 10:18 pm
Location: Madrid, Spain
Contact:

Re: Oric Atmos problems

Post by Silicebit. »

Mmmm... It seem be a temperature sensitive semiconductor. To detect temperature failures I recommend a freeze spray like this or equivalent.
Oric user since 1984. YouTube
RetroUser
2nd Star Corporal
Posts: 22
Joined: Sat Jun 25, 2022 9:14 pm

Re: Oric Atmos problems

Post by RetroUser »

Thanks for the program and advice. I typed in and ran the program and it appears there are no further problems. I let the program run for 49 passes and no error was generated so conclude the RAM is okay. Unfortunately, I mixed up the RAM I removed, by accident, so not sure which of the 4164s is the problematic one. Could be the way the computer accesses the RAM caused the problem due to the RAM being marginal and the RAM will be okay in another computer.

Only problem left is closing the case as there's a gap due, I think, to socketing the VIA, maybe the RAM too. Can screw down the rear of the case (where the ports are) but can't screw down the front because doing so jams the spacebar. Will have to have a think about how to solve this problem.
User avatar
Dbug
Site Admin
Posts: 4437
Joined: Fri Jan 06, 2006 10:00 pm
Location: Oslo, Norway
Contact:

Re: Oric Atmos problems

Post by Dbug »

Just to be clear, this program only checked for the specific case you had, which was that one bit switched between 1 and 0.

What would make sense is to run a proper memory checker that tests various patterns over time.

Clearance issues when putting sockets is something that happens unfortunately, but I think that's mostly the VIA that cause problem, generally rams on socket are fine.
Post Reply