0
\$\begingroup\$

I've been a tinkerer and hobbyist for 20 years now, but mostly in software. I've done my fair share of simple soldering (replaced controller ports on an original Xbox with actual USB ports, for example) but only have the vaguest idea of how electricity actually works.

All that being said, I've got a virtualization setup that I love and for which I'd like to sort out an audio playback solution. Basically, I just want to be able to funnel two analog stereo sources into one amplifier and not damage any of my hardware.

The first thing I did was buy one of these and ran it backwards, effectively. The amp was connected "Amp Input" and my two sources were connected to the A and B channel outputs. That worked fine, but before long I wanted to eliminate one of the (now three) boxes on my desk.

After some googling, I wound up with this (likely often cited) article: http://www.rane.com/note109.html and set to building myself what I understood as the "right" circuit. Enter this fella:

Summing cable circuit

PC 1 is onboard audio on a fairly high-end motherboard, originating as a TRRS 1/8" connector. PC 2 is the 1/4" TS output from a Native Instruments Komplete Audio 2 USB audio interface. The amp takes two RCA inputs.

My solder job really sucked, but the connections were sturdy. I just used some high end(?) RCA cable I had lying around and a few other connectors. Once complete, I wrapped the whole shebang in electrical tape (no heat shrink at hand) and connected everything. And wouldn't ya know it - everything worked!

I have since discovered an issue, however, and I'm nearly certain it's my cable.

If I have the amp cranked, there is a very audible, pulsing hum. It sounds like cicadas, actually. It didn't take long for me to realize the pulses matched exactly with the "breathing" effect from my motherboard's LED controller. I've got two strips of white LED's pulsing, and the buzzing corresponds with the fade in either direction. When they're either fully off or fully on, there's no noise.

Predictably, when I remove the 1/8" TRRS connector from the motherboard, there's no interference. A little more interestingly, the signal noise also vanishes if I disconnect the outputs of the USB sound card.

So... what have I done? What might solutions be other than changing the lighting? ;) What did that little box do that I could replicate, perhaps?

Thank you all in advance for any help!

\$\endgroup\$
  • \$\begingroup\$ I am actually not sure what you want to do. Is it a mixer? Well, at least you are loading your outputs with way too much current. As far as I remember those outputs are typically specified for like 40 kOhms or so and you are loading them with a 2x470 Ohms voltage divider. Furthermore there might be all sorts of ground loops if both the outputs are powered indpependently. Better use a summation amplifier (OpAmp circuit) with reasonably high load resistors. \$\endgroup\$ – oliver Jan 5 at 20:19
  • \$\begingroup\$ Welcome to Electronics.SE, I really appreciate your effort and enthusiasm when writing you question and I certainly have seen worse. But, please limit your question to the relevant part (mainly the third last paragraph and the paragraph above the circuit diagram). your question appers to be about "EMI" add that tag too and add more information to the title (i.e. "Cause of humming sound in my audio signal mixer related to a LED strip operation"). Thank you and happy tinkering! \$\endgroup\$ – try-catch-finally Jan 5 at 20:21
  • \$\begingroup\$ @oliver Ah, yes. I was trying to build a stereo summing circuit. Also, I functionally understood very little of your answer, so I will have to get back to you on most of it. =) \$\endgroup\$ – musasabi Jan 5 at 20:28
  • \$\begingroup\$ @try-catch-finally Thank you kindly. I will make the changes. I suppose I was overly concerned with not posting a useless question, but ya know. Haha \$\endgroup\$ – musasabi Jan 5 at 20:29
2
\$\begingroup\$

This is not really an answer, just help in identifying the problem.

the signal noise also vanishes if I disconnect the outputs of the USB sound card.

This means it could be a ground loop issue. So, please try this:

enter image description here

This means you desolder the end of each 470R resistor from the wire that goes to the signal wire of your TRS jacks, and solder them to the ground wire of the same jack instead.

Since signal is connected with ground, no sound should come out of the amplifier, and you certainly won't hear any PC audio.

If you still hear your LED PWM in the loudspeakers with this arrangement, it means the potential of "ground" is different between all the points labeled "ground". So "ground" which should be 0V contains audible signal voltage instead. The cause can be high current (like your PWM LED current) flowing through ground wires, which do not have zero resistance, and this creates noise voltage.

If this is the case, then the problem must be fixed at the source, and the chopped LED current should be routed away from the sensitive parts. How to do that depends on how your LEDs are wired, how the controller is wired etc, so you will have to post more info and schematics.

But at least this simple test can narrow it down to "ground loop" or "something else".

| improve this answer | |
\$\endgroup\$
  • 1
    \$\begingroup\$ I would add USB to the troublemaker candidate set. I have met noise problems in systems where a computer has an analog audio connection to the same device which is already connected to the computer via USB. There was no such thing as silence before either audio cable or USB cable was disconnected. Data signal leaked to audio via a ground loop. Today there's available USB isolators to break the ground connection via USB cable. I haven't one, I simplified my system. \$\endgroup\$ – user287001 Jan 5 at 22:07
  • \$\begingroup\$ @user287001 yeah... and since you mention USB, I wonder if OP's LEDs are powered by 5V from a USB connector, maybe that could inject noisy current into ground near the USB port used by the soundcard. \$\endgroup\$ – bobflux Jan 5 at 22:32
  • \$\begingroup\$ Some common audio hobbyist's claims 1) connecting the grounds of the devices with a short sturdy copper wire reduces substantially the noise which leaks to unbalaned audio from USB. 2) A big radio noise supression ferrite belt around the USB cable does the same 3) beware cheap USB isolators. They are advertised as USB 2 compatible but in reality they do not support high speed bidirectional traffic needed by multichannel audio interfaces., they operate properly only in USB1-like mode.. \$\endgroup\$ – user287001 Jan 6 at 0:55
  • \$\begingroup\$ There are very few real USB2 isolators and they're all extremely expensive. \$\endgroup\$ – bobflux Jan 6 at 1:05
1
\$\begingroup\$

Exposing myself to the danger of probably missing your intentions, my suggestion would be to use an OPAmp summation amplifier (or rather two OpAmps for the two of your channels) instead of the voltage dividers comprised of the two 470Ohms resistors.

Because the input impedance of an OpAmp is very high (at least megaohms), you can choose pretty much higher resistors for your voltage divider. You could for example use 47 kOhms (x2) instead (in the linked article R1...R2; twice of course).

The summation amplifier is based on the fact, that the inverting input is always kept at ground potential (the one connected to the non-inverting input of the OpAmp), so the currents through R1 and R2 sum up, and ultimately flow through Rf. If you choose Rf = 47k also, you will get a unit gain mixer.

Apart from that: why do you want to build such a really truly basic circuit yourself? Why not just buy a simple mixer if all you want is a working solution. For example something like this one (really only a random pick, no recommendation).

| improve this answer | |
\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.