EVALUATING THE 6581 SID (SOUND) CHIP
              latest updates and/or corrections: 7-15-2018

     When I repair a C64 or 128 computer, it's often necessary to test or 
just replace a bad or failing SID chip. Since I have no spare SIDs left, I 
ask the owner of the computer to locate a replacement for that particular 
IC to save them the expense of having me do it. However, when I get any IC 
from an unknown source, I have to make sure it's working properly before 
I install it and return the computer to the owner. I recently got some 
unusual results when testing a pair of replacement 6581 SID chips. One 
worked perfectly and the other garbled voices 2 and 3. I'm not talking 
about the minor differences of voicing between versions of the SID. 
     My most critical test for SID chips is the original Commodore Test 
CART V586220. It came with an "octopus" cable that plugs into all the 
external ports to do a complete evaluation of a C64 board. However, that 
cable is not necessary to test all computer functions, one of which is 
the SID chip. The SID test is the last one performed by the cart program.
It checks all three voices with multiple tones and its noise output as 
well as the SIDs A/D converter via the POT lines. If any of the tones it 
plays are missing or garbled, the SID is suspect. Now the problem...
     One of the two SID chips I tested seemed to be bad... voices two 
and three were garbled. Listen to the audio clips of good and bad SIDs 
to note those test results. I always use the same C64 board for tests 
as all its chips now have sockets. Now, why did that "failing" IC work 
-perfectly- in another board of the same type (both # 250407)? I could 
only assume that the IC was marginal somehow, perhaps failing, but now 
I just don't know. I thought back to my own experiments with replacement 
PLA chips (using an EPROM) and that system timing is a factor that can 
make such a replacement IC marginal or non-functional. Now I'm starting 
to think the SID can suffer the same marginal operation because of system 
timing differences between C64 boards. I tried that same SID in another 
C64 board (same 250407) and it produced a less glitchy output, different 
than my other test board. Lastly, because it worked perfectly in the 
owners board, we agreed that the chip, although possibly marginal, would 
be OK as a replacement. I ran the all-function test for over an hour to 
make sure the board was stable. 
     The SID is one of a few IC's in a C64 that runs hot and consequently 
fails often. That's why I routinely install heat sinks on that IC, the 
PLA and the CPU when I get one of those computers in for repair. 
Commodore knew about the problem and installed a metal shield in their 
later version C64's. The metal shield's "fingers" touch half a dozen IC's
and act as heat sinks. For earlier boards, it's advisable to add sinks 
to the the chips that run hot to make them last longer. See the article 
on this site regarding heat sinks for the C64. Most people don't have 
access to the test cart I use, so I wrote a simple BASIC test program 
that will provide the user an alternate way to test their SID chip. It's
not as inclusive of all faults but it will tell if the chip is working
at all or not. The best dynamic test I found is the disk program: Music 
Construction Set fo the C64. It will sometimes find a problem when all 
other tests show OK including the Commodore diagnostic cart. I use this 
program as a final test before sending a repair job out the door.

Ray