SirNickity:
I think you're beginning to learn a painful lesson on why you don't produce a PCB in large numbers until you've had a chance to run it through the wringer for a while.
Well I didn't really have a choice.
I raised the $16K to produce the boards in May of last year (actually $14K since Kickstarter and Amazon took a chunk) and then I realized my original design wasn't going to cut it and redesigned the whole thing virtually from scratch, upgrading the 328P to the 1284P, increasing the available IO ports, changing the power LED to an indicator LED, redesigning the audio circuit to provide better isolation from noise in the power and ground planes (apparently I didn't do a good enough job), moving all the LED drivers off the main board (driven by the realization that with a proton pack, at least one on the board would end up being wasted, and it was more efficient to create modules that could allow ribbon cables to be plugged in), and finally just before shipping, I had to remove the voltage regulator from the board and replace it with the mosfet because I realized I'd screwed up the power dissipation calculations and the regulator was going to melt the board to slag.
I stupidly assumed that if I followed the recommended layout for the TLC5947 exactly that that would provide adequate noise suppression. And I didn't encounter anyone who'd mentioned having this sort of problem with them. And as for the audio circuit, I did the same, following the schematic, and looking at how others designed their line outs. I tried my best to isolate that portion of the circuit from noise, but information on how to properly isolate the audio portion of a circuit is hard to come by.
Anyway, I finished the redesign at the end of September and many folks were expecting the boards for Halloween and I thought maybe I could still put together something in that time, but obviously that didn't happen. Programming them and working out the kinks with the audio by moving the LED drivers to their own data bus, and then getting an SD card based bootloader to work took several more months, and I've finally just started to ship them. Yes, I've had to ship them with the audio problem. My customers are willing to purchase a ground loop isolator, but not willing to wait any longer.
Which is what they'd be doing (waiting) if I had not had the PCB house press full steam ahead in October because it took till now to find most of the issues and I still don't have a solution for the audio issue. I've got some suggestions I can try but I don't know they'll fix the problem for certain. And even if they do fix the current problem, what of all the noise that is generated when I power servos? Fixing the LED drivers isn't going to fix that. I haven't yet tested the servo noise issue with the ground loop isolator. I suspect it will greatly reduce the noise, but not eliminate it completely, as has been the case with the LED drivers. A combination of greatly eliminating and attempting to reduce it seems like the best course of action. It's not like I have tons of space on the board to add big capacitors.
So I really had no choice in the matter.
I agree with you that the noise is probably not caused by a ground loop. Trying to alleviate the noise using techniques intended to fix ground loops is just going to be a waste of time. There were some good suggestions a while back on trying to isolate the high current loops using capacitance and such. Did you try that? I know it's not what you want to hear based on board layout, but if you can, test the theory anyway. If for no other reason than to prove it wrong and move on to something else. Do whatever you have to do to get a large cap on the power rails near your LEDs and see what happens.
Oh, I believe them when they say it's the LED modules causing the noise, but modifying them at this point is not going to happen. That is something that will have to wait for a redesign. Right now, I am starting to ship the boards as is because with the ground loop isolator they work good enough for their intended purpose. Adding more capacitance is something I will try when I have some time and before I get the next batch manufactured, assuming there is a next batch.
But assuming it is high current loops in the LED drivers... What would the isolation transformer on the RCA line be doing to fix that, and how might I replicate that without a big transformer? (big relative to my tiny 2"x2" board I mean) If it were simply that the board's ground was at a different potential than the amp, and I'm sure that's true as well, I would have expected the capacitors on the RCA cable to do something, but they didn't do a thing.