This suggests on primary has been wound the opposite way to the other, such= that the line side of one transformer is against the secondary winding, wh= ile the other has the neutral side against the secondary. -----Original Message----- From: piclist-bounces@mit.edu On Behalf Of Bob Bl= ick Sent: 03 July 2019 18:35 To: Microcontroller discussion list - Public. Subject: Re: [EE]: Audio amplifier power supply problem Hi David and all the other responders, Thanks for your input. The thing that really made this weird is that it's j= ust two single channel amps that only share a chassis and a power cord, wit= h a couple of resistors and capacitors tying the two grounds(from the main = filter caps) to chassis, which is a pretty typical thing in dual-monaural a= mps. The hum only happens when the two amps are connected together. Last night I followed up on my thought that the two transformers were diffe= rent from each other and generally pretty crappy with regards to common-mod= e isolation. Jim suggested I check the lot numbers, and indeed, they are fr= om different production batches. =20 Pulling the chassis connections, I tested the ground voltage of the two amp= s referenced to AC neutral, and one side was about .7V and the other about = 2V. Reversing the AC line I got about 70 and 90 volts. It looks like the tw= o transformers are definitely different, and there is enough common-mode cu= rrent to make some hum, especially when the AC line is reversed. Looks like there's not much to be done except to put a polarized line cord = on the amp. Since I cut those two .01 caps off the input to the transformer= s the hum is pretty much gone. I really have no idea why that would make a = difference, but I'm not going to drive myself crazy trying to figure that o= ne out. Thanks again! Bob ________________________________________ From: piclist-bounces@mit.edu on behalf of David = Reaves Sent: Wednesday, July 3, 2019 6:14 AM To: piclist@mit.edu Cc: Reaves David Subject: RE: [EE]: Audio amplifier power supply problem You may be over-analyzing the problem as to it being somewhere in the trans= former/power supply. It may be an unseen ground-loop problem. I have found, by strictly following high-quality analog audio wiring practi= ces, that having a **single point ground** and *any and all* ground connect= ions being single-wire "star" connections to that ground (usually located a= t the low end on the largest filter capacitor), that in almost all cases hu= m problems do not show up in the first place (Don't use loops, either)! Eve= ry subassembly's ground connection should have its own individual wire to t= hat point, and if it's a relatively high-current-demand section of the circ= uit, the wire should be substantial and direct, to keep the impedance low. = This philosophy is similar to the one-connection-only practice where the di= gital and analog grounds on a D/A or A/D converter have to be tied together= on a circuit board. Also, not using the outer shield to carry the low side of the input signal = may be of help, with the shield being connected to "ground" only at the end= with the highest noise susceptibility (receiving end, usually), and the ac= tual circuit return being brought in on a separate, but closely adjacent, c= onductor. Two-conductor shielded cable used for balanced circuits works for= this technique, too. Good Luck David Reaves -- http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/chang= e your membership options at http://mailman.mit.edu/mailman/listinfo/piclis= t --=20 http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist .