Problem Williams Defender D8224 Sound problems

Elektraglide

Newbie
Credits
2CR
HI
My D8224 sound board plays the test sound when the test button is pressed, and if I pull J3 sound select lines low, 2 or 3 sounds can be made to play but most do not.
I've traced every sound select line with a logic probe thru the 4050 buffer and the 14068 8-way NAND to the PIA and all looks good.
I've swapped the PIA with a couple of spares and it makes no difference. I guess its possible I have 3 bad PIAs, but I *think* unlikely.

I've rigged up 4 buttons on the sound select lines 0-3 so I can manually trigger sounds - see pic.
I can get other sounds to play. But then pressing the same button (or buttons) has no sound. That makes me think the sound data, PIA and CPU etc are actually good but something stops sounds getting triggered - but other than pull up resistors and noiuse caps there is nothing else..

Any suggestions as to what to try next? My assumption is that if it can play any sounds, the RAM and CPU are working. Is that true?

IMG_5442.png
 
Last edited:

philmurr

Active member
vacBacker
Feedback
46 (100%)
Credits
2,338CR
Are you actually having sound problems when the sound board is in the cab, or when you're testing it on the bench and not connected to the other Defender boards?

If it's the latter then the above behaviour is perfectly normal . If it's the former you need to check the cable between the ROM board and the sound board, and the ROM board PIA.
 

Elektraglide

Newbie
Credits
2CR
yes, in cab the sounds tests mostly are silent - very occasionally 1 sound will work. And occasionally during gameplay I get a sound.
ROM board -> sound board cable is newly made. ROM board PIA was recently replaced.
I *think* I may have found the problem. I triggered on the falling edge of a sound line on CH2 and watched the 8 input NAND to CB1 on the PIA on CH1 (yellow).

It looks bad to me - but I am not a h/w guy.


NAND14068_risetime.png
 

Mc-Q

Active member
Credits
1,450CR
i would remove the 8 input nand and then test the board.
it works by the nand generating what is pretty much a strobe to read the other lines
so you can set binary numbers on the data - then strobe it in

and yes, the yellow trace looks bad, but not knowing the voltage or if the lower line is ground or floating makes it harder - turn some of the readings on on your rigol

one other observation - the yellow trace is filled with jitter.
did you replace the caps relating to the 5v regulator?
the williams soundboards dont use the cab 5v - they generate their own.
 
Top