Author Topic: Best way to debug a stuck bus line?!  (Read 2556 times)

Elgen

  • Member
  • **
  • Posts: 104
  • Vanløse, Denmark
    • View Profile
    • elgensrepairs.blogspot.dk
Best way to debug a stuck bus line?!
« on: January 26, 2013, 11:56:47 AM »
Okay, here's the situation:
I'm currently working on a bootleg 1942 where the only initial problem was that some of the controller lines was stuck low causing the plane to go left and use all rolls at start-up. That fix was easy: Changed a couple of fried 367's in the input section.
However while working on it, I suddenly lost all sound?! I don't think it's related to the input problem at all, just a coincidence.
The sound CPU is a Z80 controlling 2 AY-3-8910 via the Z80's databus. On the same bus is also an SRAM and the sound ROM. Now one of the 8 lines is stuck low (one right in the middle). That is it fluctuates between GND and 0.5V, so it looks to me like something is trying to drive it. It also have direct contact to GND (used continuity tester). But I can't seem to find the component that's pulling it down. Tried to cut the relevant pins on both the AYs, the ROM (just pulled it from the socket), the RAM, the CPU, and also a 374 I found to be on the bus. I would rather avoid desoldering the CPU and the AYs.
As bonus info I have a working SolomonsKey with 3 of the same sound chips, and here ALL datalines are active.
By the way, the way I tried to find ICs and components that connects to the stuck, is by starting at the AYs and beeping the healthy bus line just besides the stuck one around the PCB and then looking if data pins 'close to' it on the ICs I found were stuck low.

Where would go next? Is it desoldering the AYs and CPU to see how the circuit is under them, or is there more to check before that?

Any help could be useful };-P

PaulSwan

  • ArcadeLifeStyler'
  • ***
  • Posts: 786
  • ZZZ
    • View Profile
    • The Zaccaria WWW Site
Re: Best way to debug a stuck bus line?!
« Reply #1 on: January 26, 2013, 06:45:04 PM »
To be more sure you have a stuck data line you need to rule out that he CPU is just not crashed and stuck. Since any such crash like this is likely over in milliseconds after reset what you see is the state after the crash. In these sorts of cases a pulse injector is your friend - I have one I set to 400pps on the reset line to keep the CPU being reset 400 times each second. On a scope, you can see exactly what the initial activity is. It may or may not shed more light on the issue - in 50% or so of cases it might be possible to see a bad address, data or control line. In other cases it all looks fine and is a dead end.

If you have a Fluke, of course, then you can just pop out the CPU, socket, and investigate from there. The sound circuits are so simple with a Fluke it'd take just a few minutes to figure out what was really wrong :)

Paul.

ZZZ

Elgen

  • Member
  • **
  • Posts: 104
  • Vanløse, Denmark
    • View Profile
    • elgensrepairs.blogspot.dk
Re: Best way to debug a stuck bus line?!
« Reply #2 on: February 04, 2013, 08:49:17 AM »
Oops, forgot to follow-up on this one. I found the short; it was a solder-blob
https://www.dragonslairfans.com/smfor/index.php?topic=4246.0