Sorry to resurrect an old thread, but my midibox sid (modular, not wilba) is doing exactly the same thing. It’s really strange; if no leds are lit, there’s no hum, and also if all the LEDs in the matrix are lit, the hum disappears, but if even one led is on, I get this constant hum. It’s not 50/60 hz as far as I can tell. Depending on the number of leds lit, it sounds like it’s varying the pw of the hum too, or at least pushing forward different overtones. It’s loudest in the multi engine, and actually the different engines seem to invert the behavior. In the multi engine, I don’t hear any hum unless a note is played. In the lead engine, it’s the opposite! There is hum only if no oscillators are sounding.
This is really a difficult problem to track down. Mike (or anyone else), did you ever get this fixed/figure out what was going on?
(EDIT to clarify: no leds or all leds doesn’t produce hum, but any number in between does. Also, if I hold down one of the row buttons and cause an entire row to blink, there is no hum until I select which led in the row will remain lit using a column button, and then the hum starts. Holding down a column button to produce a blinking column causes hum immediately.)
Okay. I just spent a couple hours trying to sort this out. The noise seems to be emanating from the sid chips themselves. If you engage all (in my case) six oscillators, the hum completely disappears. if however you have even one oscillator disabled(so that when you play a note, only 5 out of six oscillators sound), the hum is present.
The good news is I found a *kind of* solution. I put one 2200 uf capacitor in line with the power input on each dout board, and this has dropped the hum by about 15-20 db, so it’s much better now, and really barely audible (most present on high notes, since lower frequency stuff completely masks it).
This isn’t a real solution however as the hum is still present. It’s also just confused me more, since it seems the hum is not directly emanating from the power (ie, not a simple ground loop or something similar), but is still somehow directly affected by changes to the power rail. It’s like some kind of complex cross talking going on. If anyone wants to chime in, I’d really appreciate it!
EDIT: another possible not-really-solution would be to implement a simple matrix-off mode in the sid application. It could be as simple as a third mode toggled with the M_Mode button. This way we could at least disable the whole thing when fidelity is really important, like making a studio recording for instance. Sadly I lack the assembly language skills to do even a simple mod like this.
I’m really wondering how widespread this issue is. It seems probable that it is somehow linked to the sid’s leaky DCA, so someone using an 8580 might not even hear it without *really* listening for it (Mike, were you using a 6581 or an 8580/6582?). Or maybe we just screwed something up. But the issue seems systemic to me, I just can’t make sense of its complex behavior.
)… does everone elses Wilba 6582 have this sort of problem?