http://www.neo-geo.com/forums/showthread.php?p=975663
http://www.neo-geo.com/forums/showthread.php?t=140668
http://www.neo-geo.com/forums/showthread.php?t=129063
A few years ago, after buying a system from me, the customer (who considered himself an audiophile) complained that he could hear a buzzing noise coming from his TV speakers as well as from the headphone jack. He also mentioned that he had a low serial system that didn't have that issue at all.
Since I had another (unmodded) system with the same PCB revision (3-4), I tested it immediately and much to my surprise it was having the same issues described by the guy. I couldn't believe I had never noticed it because once it was pointed out to me it was impossible not to hear it.
To get a better grasp of the problem I ran tests with as basic a setup as possible: I ran the system without hooking up video, only with headphones plugged in and another system (revision 3-6) running next to it, also with only headphones in and only the 3-4 was having the buzzing issue. Both system were powered by the same power supply (without step-down transformer) so the PSU was ruled out as a possible cause.
When the guy sent the system back to me along with his other "silent" system (that turned out to be the very first PCB type NEO-AES) the tests confirmed that 3-4 had the issue and that NEO-AES was buzz-free just like 3-6. The thing was all the more puzzling as NEO-AES and 3-4, despite being different revisions, have the same audio circuit (while on revisions 3-5 and 3-6 it's partly different).
I was starting to randomly put the blame on the different PCB layout of the 3-4 causing interference when I noticed the difference between the NEO-AES and 3-4 circuits: each audio IC (DAC, op-amps, headphone amplifier) on the NEO-AES has a decoupling (bypass) cap nearby (close to the VCC and GND leads) just like all the other ICs on the board which is what you would normally expect to see:
https://i.imgur.com/5RRBXCw.jpg
https://i.imgur.com/M92PyvG.jpg
https://i.imgur.com/W9UHCXp.jpg
The 3-4 system doesn't have them though. Not a single bypass cap in sight:
https://i.imgur.com/NOPozTd.jpg
https://i.imgur.com/kqToWc6.jpg
After adding 0.1uF (104 marking) caps across the 5V and ground leads of each IC the issue was gone:
https://i.imgur.com/dTJPkJE.jpg
The cap for the DAC (YM3016) seems to make the most difference.
According to the
datasheet, the bypass cap recommended for the headphone amplifier IC is a 100uF electrolytic cap. The value for this cap on MVS boards is 470uF. I put a 100uF (470uF is too bulky) in addition to the 0.1uF:
https://i.imgur.com/gM0YRsT.jpg
https://i.imgur.com/yFfolEn.jpg