Question for SID specialists

[Xanxi]

Hi there.

I have here a SID 6581 marked as 2084, which gives bad results on some SID musics i have been testing.
It seems something is missing from the music, compared to the same one played on a XSID (Audio Boost) in 8580 mode and compared to two real 6581 (3984 and 2084) playing together in stereo.
I don't have other chips at hand right now.

I have used a test program SID Tester by Andrew Challis and it appears that all three channels are working with all waveforms, but when it comes to filters, High Filter = 15 makes that i can't hear anything in Low Pass and sound is very weak with Band Pass. All other tests are fine and audible.

ChatGPT says to me that it could be normal with this filter due to the chip sensibility but this AI has been proven to tell me a lot of bullshit before.

Indeed, i think i got this one from Ebay a couple of years ago or similar and there are often defect chips.

So, is it a defective chip or not?

 

[miggymad]

Fun Fact: The unique and iconic sound of the MOS 6581 SID chip wasn't solely a result of its advanced design. In fact, due to manufacturing inconsistencies and variations in the production process, no two SID chips sounded exactly alike! These slight variations in the analog components of each chip meant that the same piece of music could sound subtly different from one Commodore 64 to another. This unintentional quirk added to the charm and individuality of each machine, making every Commodore 64's sound output as unique as a fingerprint!

Watch and listen - Same music played thru 4 different chips ...

 

[Xanxi]

Yes i have already heard myself this kind of differences. But in this case, the musics i am testing simply seem to lack a part.
I expected that a channel out of 3 to be missing, but according to my test software it is not the case.
The SID chip seems to be quite complicated, i need to read again some old books about computer music to understand fully how it generates sound as it associate waveforms with filters and filters seem to be very important in the result.
Anyway, i am still under the impression that this chip is faulty.

 

[Bladejack]

This is the best reference I found to really testing Sid's -- Sid Test

 

[Xanxi]

Thanks for this link.
It leads to another webpage https://www.6581-8580.com/socse/ind...rd=off&add_fields=&query=bionic&term_counts=0 with MP3 of several different SIDs playing the same tune, and there is Bionic Commando from Patfield Kent which is one of the tunes causing issues to my under investigation SID. From about 50 s, the tune has a main beat and higher pitched notes rapidly playing on top of it. Those notes are missing on my SID (they are played with my other SIDs) and yet i can hear them with all 4 types of SIDs recorded on the website. That's why i thought i had a missing channel.

 

[Xanxi]

So i have swapped my suspicious SID for an ARMSID and i have noticed that my 6581 had pin 13 bent and out of socket.

According to this reference, pins 9 to 13 do :

9-13

A0-A4

in

Address bits 0-4 for selecting one of the 29 registers.​

So perhaps the SID was fine but this pin caused the problem. Now it seems tricky to unbend this pin without breaking it. I think i will wait for now.

 

[Xanxi]

I have tried to straighten the pin but can't do better than that and seems fragile, bent and rotated on its axis.
Maybe it would be better to stick that chip into a socket and solder a wire to bypass that pin.

 

[Xanxi]

So i haven't forgotten this chip.
I have installed into a socket in order to solder a wire from this pin to the socket pin in case it breaks.
The pin 13 is still bent and i was unsure if it was doing contact with the socket or not and i checked with a multimeter before soldering and continuity checks between the pin 13 and the pin of the socket. So i guess it is inserted enough.
Yet, it didn't solve my sound issue and there is still a part missing or unaudible in Bionic Commando.
I conclude the chip is indeed partially faulty unfortunately.

 

[levo-ace]

As happens with so many sid's , part of the analogue circuit has gone faulty .