1-2-1 Memory Error on Bubbles board

YellowDog

In Memoriam

Donor 2011, 2013
Joined
Apr 12, 2008
Messages
9,851
Reaction score
158
Location
Humble, Texas
I am trying to sort out a 1-2-1 memory error on a Bubbles MPU board. I can't tell you what it is, but I can tell you what it isn't:
(1) It is not the RAM chip. I swapped chip 21 with chip 22, chip 11 and chip 38. The error did not move.
(2) It is not the 74153 chips. I replaced all 4 of them
(3) It is not chip 3H (74LS373) I replacd that chip as well.
(4) It is not the ROM card. I swapped 3 different ROM cards. Same symptoms.
(5) It is not the I/O card. 2 different I/O cards and no I.O card. Same symptoms.
(6) It is not the socket. I checked and all paths show the correct continuity and no stray paths.

Before I start shotgunning chips (worse than I have) I thought I would check to see if anybody else has run into this on any other Williams MPU boards and has any hints.

Also about once every 10 boots it comes up without the RAM error. If I warm reboot it (via reset switch), it will occasionally show a 1-3-1 error.

ken
 
Last edited:
Hmm, the last time I had a similar issue where I tried everything, it turned out to be a broken trace UNDER the socket on the top side, but as you have continuity, that's not it.

Have you tried a new 6502 just for kicks?
 
I had a Bubbles doing this exact same thing. The +5vdc was failing, but the LED was lit. Put a new power supply in it and it was good for a few days, and the problem came back. Four different power supplies later, it is rock solid.

Was not able to find a problem on any of the power supplies that I used.
 
Hmm, the last time I had a similar issue where I tried everything, it turned out to be a broken trace UNDER the socket on the top side, but as you have continuity, that's not it.

Have you tried a new 6502 just for kicks?

When I have an issue with RAM that doesn't move, the socket is always suspect #1. In this case the continuity is solid.

It's a 6809 and I did reseat it, but I have not tried swapping it out. I'll probably try that tonight.

I had a Bubbles doing this exact same thing. The +5vdc was failing, but the LED was lit. Put a new power supply in it and it was good for a few days, and the problem came back. Four different power supplies later, it is rock solid.

Was not able to find a problem on any of the power supplies that I used.

I am running this board off a switcher as it is on my desk right now. Tonight I'll try putting it on a Williams PS setup and see if I get different results.

I did try dialing the +5V up to 5.10V and down to 4.80V and it showed the same results consistently.

I had this same problem on a Robotron MPU board about a year ago. The really strange thing in both cases is that if I put an HP10529a logic analyser on any of the 74153's the problem goes away and it boot with no issues. That is why I replaced the 74153's. Now I am starting to think that it is something else that is failing and bleeding onto the data buss and putting the probe on it changed the electrical characteristics just slightly.

Thanks for the suggestions guys.

ken
 
Well, we are 0 for 3. I tried replacing the CPU with a known good one. No change. I also plugged it into a known good Williams power supply with no change. I even swapped out all the RAM in bank 2 just for grins and no change. Well, actually when it was on the Williams PS the sporatic 1-3-1 issues when it was soft booted stopped, but I am not sure if that happened before after the RAM swap or not. I also swapped two other Bubbles ROM boards in and had the same issues.

Oh, well. I guess I need to look for other chips on the data buss and start shotgunning....:(

Thanks for the suggestions though. I may need to back burner this board for a while and get on with some other projects.

ken
 
Well, we are 0 for 3. I tried replacing the CPU with a known good one. No change. I also plugged it into a known good Williams power supply with no change. I even swapped out all the RAM in bank 2 just for grins and no change. Well, actually when it was on the Williams PS the sporatic 1-3-1 issues when it was soft booted stopped, but I am not sure if that happened before after the RAM swap or not. I also swapped two other Bubbles ROM boards in and had the same issues.

Oh, well. I guess I need to look for other chips on the data buss and start shotgunning....:(

Thanks for the suggestions though. I may need to back burner this board for a while and get on with some other projects.

ken

you could try some temporary pullup resistors on the inputs to the 74153s - might help to highlight if the 74153's are getting good inputs or not.
 
Last edited:
...

Ive had that problem before. Kept getting ram errors. For me it turned out to be a fault power supply. The +5v wasnt constant and would drop below +4.8 and reset the game w/ram error.

Jon
 
Ive had that problem before. Kept getting ram errors. For me it turned out to be a fault power supply. The +5v wasnt constant and would drop below +4.8 and reset the game w/ram error.

Jon

Power supply was one of the first things I tried. Power at the RAM chips meters out good.

you could try some temporary pullup resistors on the inputs to the 74153s - might help to highlight if the 74153's are getting good inputs.

I have been looking at that. Also noticed that one of the boards (Rev B, I think) had capacitors across the 74LS374s. I may look at that a little closer and try it. I have nothing to lose at this point.

ken
 
i notice you said 74ls374's , emphasis on the 374. the schematic i am looking at shows 1H, 3H, 4H are 74ls373's not 4's presuming those are the ICs we are referring to.

there is a small diff between the two components. as the 374 is edge triggered rather than level triggered like the 373 so maybe you dont have a clean sharp edge transition from 6J, which the Logice Analyser may have conveniently corrected, also if the leading edge of the transition from 6J is rolled over it may also affect timing of data transfer (ie delay), meaning data may not be clocked through properly or at all. in which case i would suggest change the 374 back to a 373.
 
Last edited:
I'll have to check the part numbers, I was going from memory when I posted yesterday.

ken
 
Interesting, on a Robotron MPU board (Rev B) the schematics show 1H, 3H & 4H as 74LS374 chips. The Joust schematics show 1H, 3H & 4H as 74LS373. I wonder if I swap the 74's out for 73's if that would fix it..........

ken
 
Back
Top Bottom