Related to Step 22: Building a linear PSU
Update:
The new PSU had - depending on week day and weather :-) - some small noise issues... it was never really bad... but perceptible... like a 50hz hum and some infrequent "crackles"... I really wanted to solve the problem and was wondering why the new PSU wasn´t better in terms of noise than the old PSU... it has quality components all over... so... back to the assembly table :-)
After bugging Antix (because he built an identical PSU) and getting lots of feedback from orange_hand... the following tuning steps were performed...
a) Installation of a second 2200uF capacitator in parallel to the first one... which improves ripple filtering (thanks for that tip, Antix!).
b) Very close installation of the 100nF filtering capacitators to the 78S05 voltage regulator (you can directly solder them on the 78S05 pins, these capacitators are tiny) - this helped a lot... any connection longer than a few centimeters seems to act like an antenna and picks up transformer hum and switching psu noise from nearby synth switchers (thanks a lot for that tip, orange_hand!)
I did some A-B recording tests (using an old C64 PSU as a sparring partner) with the same unplayed patch ("silence" output from the MB6582 SID pair 1) at the same recording levels using max line-in gain on a tascam dr-100 audio recorder, recording 24bit, 48khz wavs...
And yes... there is a difference... see photo 1... after amplifying both waveforms once again with 45.5db in audacity, the lower waveform (old C64 "elephant foot" PSU) reached the maximum amplitude... the upper waveform (new linear PSU) was more silent (and has 2.5db more room to the amplitude max after amplification). It is a barely audible difference when using headphones during A-B tests, but every little bit of noise reduction counts... it is very well audible, when amplified... so... mission accomplished :smile:.
Apart from the improved protection of the SIDs, it might make sense to build such a PSU just because of the lower noise level.
Greets!
Peter