What's new

YesAffinity

Enthusiast
Joined
May 20, 2016
Messages
234
Reaction score
143
Location
Escondido, CA, USA
I had a thread going over at Arcade Otaku Forums, dating back a year and a half. I finally fixed this one, and thought the summary would be welcome here as well.

Symptom:
CPS1 'A' Board running Street Fighter 2 World Warrior B/C set, would lose audio after an indeterminate amount of time. The issue seemed to be somewhat correlated to ambient temperature. If it's hotter outside, the audio loss occurred after a minute or two. On cooler days, the board could be happily producing audio in attract mode for upwards of 15-20 minutes before it cuts out. On one particularly cold winter morning, the board was seemingly unphased. I took my hot air station to it, warming up different parts of the board, and within a couple minutes, audio was lost. It would typically die with audio playing fine one second, would then produce a scratch or a crackle, and then silence. Less frequently, the audio would hang on the final note or tone being produced when failing. Upon power cycling the board, audio would be normal again, but usually would fail much more quickly.

Solution:
I finally tracked the problem down to a resistor on the line between audio oscillator (3.579mhz) and z80 pin 6 (z80 clk) - R53. I believe that's a pull-up resistor? That resistor is supposed to be 560 ohm, but when pulling and measuring it, it measured at 860 ohm. I then pulled the same resistor off a parts CPS Dash board, and it too measured at 860 ohm. :/ Checking the labeling on both resistors (561), it is indeed supposed to be 560 ohm. So I pulled yet another R53 off of a working board, and it measured at 660 ohm. Well, if these things are 20% tolerance then that is probably within spec. I swapped it onto the original problem board, and it is now working with no disruption to audio.

Cause?
Theoretically, it seems this location has pretty high tolerance, and presumably at 860 ohm, the circuit was still functioning normally, but once heat was introduced through operation, it would reach some threshold of increased resistance that would result in failure of the z80 clocking.

Interesting that 2 boards had the same resistor failed. Hopefully this helps someone else down the line. Now to order some 560 ohm resistors for my stockpile.
 

Attachments

  • cps1 r53 failure.png
    cps1 r53 failure.png
    891.8 KB · Views: 591
Last edited:
Just like to say Thank you for this post. I had a similar problem where I would lose audio and if I let it rest it would come back. In my case the R53 was good but I did trace it back to the crystal
 
Back
Top