Audio Mixer Crosstalk

[tooki]

So I've been working on my headphone amp project, and one of the things I've added to the project scope is a mike preamp with zero-latency monitoring of the headset microphone, for audio recording (since the headset has extremely good passive sound isolation, you can't hear yourself at all). My Mac Pro has only line-level audio input, so a standard PC-style mike won't work without a preamp. The headset in question is a Beyerdynamic MMX 300, specsheet attached.

I've built it up on a solderless breadboard and for the most part it works as expected. (The grayed out blocks in the block diagram are not built.) The problem I've run into is crosstalk, where sound out from the Mac is leaking through the circuit into the mike output (line level). It's at a fraction of the signal level. I'm using an active mixer circuit with a virtual ground, which in theory has zero current at the mix point, so crosstalk should be impossible.

I've done a fair bit of testing and have figured out that the headphones themselves are inherent to the problem; if I unplug them (or completely mute them with their volume pot, R33), the crosstalk disappears. Inserting buffer amps (U5) or another output stage (U7) doesn't make any difference.

Is there something fundamentally wrong with this circuit? Or could the problem be an artifact of breadboarding it? (I'm loathe to solder the whole thing up just for testing purposes, never mind lay out and etch a board as an experiment.*) Or could it be component choice or quality? (I used NE5532 op-amps throughout instead of the various ones in the circuits I referred to, see below. I tried a few different NE5532s for U6, the active mixer recovery amp, but observed no change.)

I'm grateful for any help. I've included a functional block diagram as well as a schematic of the circuit as currently breadboarded.


Resources used:

• HP amp circuit: http://www.paia.com/prodimages/hdasch.pdf
• Electret preamp circuit: http://www.ece.ucsb.edu/Faculty/rodwell/Classes/ece2c/labs/Lab1b-2C2007.pdf page 3
• Active mixer circuit: http://sound.westhost.com/articles/audio-mixing.htm#s3

*I used EasyEDA for the schematic. I tried both EAGLE and KiCad, but found both to be far too obtuse in their user interfaces, as well as requiring ungodly numbers of clicks to do basic things, which unfortunately aggressively aggravates my hand pain, so I had to abandon them. So there's no quick way of just doing a quick autoroute for a PCB. :/ Since doing a PCB layout causes me physical pain, I want to do that only at the very end when everything is debugged and the physical envelope defined.

[Andy Watson]

How have you grounded it? In particular where does the ground for the mic connect with respect to ground(s) for headphone amplifier. Have you created any loops?

[alexanderbrevig]

This may be stupid but it could be acoustic feedback? Simply the Mic picking up audio from the speakers? 

[tooki]

How have you grounded it? In particular where does the ground for the mic connect with respect to ground(s) for headphone amplifier. Have you created any loops?

Assuming that I'm understanding the question correctly, as one of physical assembly: one of the buses on the breadboard is the ground, every ground connection (as shown on the schematic, as well as the sleeves of the TRS jacks used for in/out) has its own wire going directly to the bus (no daisy chaining or loops).

Though this project is intended for my Mac Pro tower, I just tested it using my old MacBook on battery, so there are no actual connections to earth ground whatsoever, same result. I also tried using an iPad as the line-in source, with the MacBook recording, same result again. In other words, the problem occurs on a completely floating system.

FYI, 50Hz hum isn't an issue.

[tooki]

This may be stupid but it could be acoustic feedback? Simply the Mic picking up audio from the speakers? 

I actually meant to address this in the original post and then forgot. No, it's not acoustic/mechanical feedback from the headset's drivers. The crosstalk occurs even with the microphone unplugged from the preamp. (I'm currently using the same headset on the Mac Pro using a simple single-transistor USB-powered preamp I cobbled together, and it works great, with no noise or crosstalk. Unfortunately the built-in sound card doesn't do zero-latency monitoring, so when it's in software it's got a slight delay, hence wanting to do it in hardware outside the computer.)

[tooki]

FYI also, if you refer to the functional block diagram, the intent is to have the actual loudspeaker output come only from the Mac's line-out, the Mac's line-in is supposed to be purely the microphone output, and the headphones get a mix of both. The intended use case is of speakers totally turned off during recording, since that produces terrible echo and feedback.

[singapol]

How have you grounded it? In particular where does the ground for the mic connect with respect to ground(s) for headphone amplifier. Have you created any loops?

Assuming that I'm understanding the question correctly, as one of physical assembly: one of the buses on the breadboard is the ground, every ground connection (as shown on the schematic, as well as the sleeves of the TRS jacks used for in/out) has its own wire going directly to the bus (no daisy chaining or loops).

Though this project is intended for my Mac Pro tower, I just tested it using my old MacBook on battery, so there are no actual connections to earth ground whatsoever, same result. I also tried using an iPad as the line-in source, with the MacBook recording, same result again. In other words, the problem occurs on a completely floating system.

FYI, 50Hz hum isn't an issue.

Case of "ground is not ground". Fatal mistake 1-- using ground bus ,you are actually creating many ground loops. The proper way is for each circuit to have it's own common ground ( connect all components of this circuit to this common ground) and from each common ground to the ground bus. In essence what you are using the bus for is a star ground and from this star ground to earth ground if you so choose which is optional (but standard practice).

Fatal mistake 2 -- If you see the original preamp circuit there is a decoupling capacitor (100uF) to plus/minus power rails. The purpose of decoupling cap is to isolate circuit from power supply and act as reservoir or reserve if there is demand from circuit. So what you should add are power decoupling caps to each of the opamps dual supply in addition to 1uF.It's also standard to add 0.1uF for high frequency noise especially if
using PC equipment.If you want to save costs then decoupling each section but since you want quality and it's for recording a few more caps is worth the trouble.Don't be penny wise ,pound foolish.

PS: A 100uF capacitor has lower esr/impedance than 1uf to shunt noise to ground for frequencies that 1uf can't do efficiently.

[StillTrying]

The problem I've run into is crosstalk, where sound out from the Mac is leaking through the circuit into the mike output (line level).

On Schematic-2016-07-23.png any noise or variations on the +9V line will be fed back into the mike input through R2, especially if the +/- 9V are batteries, and driving headphones. For testing you could temporarily remove R2, and permanently, split it into 2 resistors and decouple the centre of them.

There's too many op amps in the circuit, half of them don't do anything at all, other than add a small amount of noise and distortion.

[tooki]

Fundamental question from my original post: is there something fundamentally wrong with the circuit design, or is this a construction issue that would be resolved when built up as a proper soldered PCB instead of on a solderless breadboard?

Case of "ground is not ground". Fatal mistake 1-- using ground bus ,you are actually creating many ground loops. The proper way is for each circuit to have it's own common ground ( connect all components of this circuit to this common ground) and from each common ground to the ground bus. In essence what you are using the bus for is a star ground and from this star ground to earth ground if you so choose which is optional (but standard practice).

I thought having all the grounds come together only at one point was a star ground? (I can't make that point any smaller than the single strip on the breadboard, right?)

Where does one circuit end and another begin exactly?? I'm not exactly sure how I would breadboard what you're suggesting.

Fatal mistake 2 -- If you see the original preamp circuit there is a decoupling capacitor (100uF) to plus/minus power rails. The purpose of decoupling cap is to isolate circuit from power supply and act as reservoir or reserve if there is demand from circuit.

During experimentation I tried it with and without those caps, and with various values. None made any difference whatsoever so I pulled them for simplicity's sake, this being the proof-of-concept phase. (The schematic reflects the current state as currently breadboarded.)

So what you should add are power decoupling caps to each of the opamps dual supply in addition to 1uF.

To clarify, are you saying to add large 100uF to EACH op-amp (where there 1u's currently are), or to have them once on each rail of the PSU (akin to the linked HP amp circuit's)? (In my other thread about the amp, someone said to avoid exceeding about 500uF total capacitance on each rail, saying it could cause voltage regulator instability.)

It's also standard to add 0.1uF for high frequency noise especially if using PC equipment. If you want to save costs then decoupling each section but since you want quality and it's for recording a few more caps is worth the trouble. Don't be penny wise ,pound foolish.

I didn't realize I was cheaping out by not adding components that weren't specified in the schematics I was copying.

It's also standard to add 0.1uF for high frequency noise especially if using PC equipment.

PS: A 100uF capacitor has lower esr/impedance than 1uf to shunt noise to ground for frequencies that 1uf can't do efficiently.

Is this crosstalk considered "noise"?

On Schematic-2016-07-23.png any noise or variations on the +9V line will be fed back into the mike input through R2, especially if the +/- 9V are batteries, and driving headphones. For testing you could temporarily remove R2...

Just to make sure we're on the same page, R1 and R2 split the +9V rail to create the +4.5V bias for the electret capsule -- the op-amps have real +/-9V rails. With R2 absent, how does the 4.5V get generated? (Without R2, the electret measures at 0V.) Or is there a better way to provide the electret its bias voltage?

... and permanently, split it into 2 resistors and decouple the centre of them.

Huh? Can you explain what you mean?

There's too many op amps in the circuit, half of them don't do anything at all, other than add a small amount of noise and distortion.

As I said in the original post, I added the buffers (U5) as I'd seen suggested in various articles on crosstalk elimination, but indeed they do not appear to make a difference either way. (What weren't optional are the 3 individual preamp stages [U1-U3] -- trying to get the three outputs from one preamp stage performed extremely poorly in testing.)

As I understand it, the recovery amps (U6) in the active mixer section are inherent in making it an active mixer (as opposed to a passive mixer where crosstalk is to be expected).

Indeed, the output amp stage (U7) makes no difference now, but it is ultimately necessary since there will be 4 of them in parallel in the final product.

[tooki]

P.S. For experimenting, I'm using two 9V batteries to provide true dual-rail power, but my plan was to build a dual-rail linear supply (either 7809+7909 or some newer LDO regulators) for the real deal.

[Someone]

There's too many op amps in the circuit, half of them don't do anything at all, other than add a small amount of noise and distortion.

As I said in the original post, I added the buffers (U5) as I'd seen suggested in various articles on crosstalk elimination, but indeed they do not appear to make a difference either way. (What weren't optional are the 3 individual preamp stages [U1-U3] -- trying to get the three outputs from one preamp stage performed extremely poorly in testing.)

As I understand it, the recovery amps (U6) in the active mixer section are inherent in making it an active mixer (as opposed to a passive mixer where crosstalk is to be expected).

Indeed, the output amp stage (U7) makes no difference now, but it is ultimately necessary since there will be 4 of them in parallel in the final product.

The outputs of U1B, U2B, and U3B should all be identical. Since there is buffering from U5 to reduce the small amount of possible crosstalk (depends on the output impedance of the opamps) they are unnecessary. Or in the reverse direction U5 is unnecessary since the outputs of U2B and U3B are already low impedance.

P.S. For experimenting, I'm using two 9V batteries to provide true dual-rail power, but my plan was to build a dual-rail linear supply (either 7809+7909 or some newer LDO regulators) for the real deal.

The output impedance of a 9V battery is higher than the regulators, so they will be introducing more crosstalk. But its probably still not whats causing the large problems you suggest.

[C]

Note that the Mike and speakers in headset are separate.
The only place these two should be connected is at the computer. Not doing this creates a ground loop or crosstalk.
Think of the high currents on speaker common vs very small signal on mike common.

Change your thinking, you see virtual ground all over the place with people talking about op amps. The - input is a virtual 0V difference with the + input.

Remove all the ground symbols in your drawing and replace with labels or wires.

So between your two 9V batteries you have 0V

You have 0V, headset_mike_common. headset_speaker_common, computer_line_in_common & computer_line_out_common.
Care is needed when connecting these.

When you do this look at how that connection is used.

For example U1A+ needs to connect to J1-1, no ground or 0V needed here.
C3 & C4 are just pushing noise into ground next to U1. Using one cap here pushes noise into power rails.

Use just one mike preamp to get to line level. Then use a buffer to prevent feedback when needed.
So mike preamp and a bunch of parallel buffers.

A bunch of signals at same level have less crosstalk.

 You state that a MAC has no Mike input, only Line In.
To make this a good thing to have, you probably would like a Line In Input & a mixer that adds mike_in to this to give output for MAC Line in.
Buffers on box Line In (L & R) to ganged pot.
Mike preamp to pot.
Two mixers using a total of 4 resistors connected to the two pots.
The two mixers directly drive output for MAC Line in if possible.

Mac Line out circuit is a copy of above. Where the two mixers output drive volume pots for the many outputs.


Question
  When mounted does J1 connect case to pin 1?
 

[StillTrying]

(either 7809+7909 or some newer LDO regulators)

They should be fine, and 2 batteries will be fine for now. I can't see any ground problem at least on i]Schematic-2016-07-23.png[/i]  where all currents are quite low.

Almost everywhere in an audio circuit, a signal should pass from a low impedance output, to a high impedance input - so that the original signal is not reduced or distorted. The usual impedance ratio between the input and output is 10X, preferably more, almost everywhere in Schematic-2016-07-23.png the input/output ratio is back to front.

DAT_MMX300_EN.pdf shows that the mike's impedance is about 1K5, which means it should drive an impedance of 15K+, but instead it's driving 3 2K2 in parallel = 0.73K, - 66% of its output voltage just gets lost in itself. It's similar where a 100K volume control is driving a 10K input, 100R driving 32R headphones, there's to many to list.

All 8 op amps on the top half of Schematic-2016-07-23.png can be replaced with 1. 

[Audioguru]

The bias voltage for an electret mic should be fed from an RC filter, not directly from the noisy power supply voltage.

I agree that it is crazy to use a low input resistance inverting opamp to kill much of the mic output level then crank it up with more gain and more noise from the opamp. A non-inverting mic preamp opamp is much better since it has a high input resistance that does not reduce the mic output level.

[tooki]

Thanks to everyone who has taken time to reply. It's a lot to take in and try to implement, but I'm trying!

(either 7809+7909 or some newer LDO regulators)

They should be fine, and 2 batteries will be fine for now. I can't see any ground problem at least on i]Schematic-2016-07-23.png[/i]  where all currents are quite low.

Almost everywhere in an audio circuit, a signal should pass from a low impedance output, to a high impedance input - so that the original signal is not reduced or distorted. The usual impedance ratio between the input and output is 10X, preferably more, almost everywhere in Schematic-2016-07-23.png the input/output ratio is back to front.

DAT_MMX300_EN.pdf shows that the mike's impedance is about 1K5, which means it should drive an impedance of 15K+, but instead it's driving 3 2K2 in parallel = 0.73K, - 66% of its output voltage just gets lost in itself. It's similar where a 100K volume control is driving a 10K input, 100R driving 32R headphones, there's to many to list.

All 8 op amps on the top half of Schematic-2016-07-23.png can be replaced with 1. 

Thanks, I built up your circuit and it works beautifully.

The bias voltage for an electret mic should be fed from an RC filter, not directly from the noisy power supply voltage.

That's what StillTrying's circuit above does, correct? (The topmost 47K and the 47u to ground.)



I did a bunch more experimenting based on your replies, and seem to have narrowed it down a lot (I'm sure this is obvious to you): if the microphone preamp shares the same power supply as the headphone output amp, then the problem appears. If I move either one to a separate power supply (e.g one on battery, the other on bench PSU), the problem disappears entirely. (The mixer doesn't care who it shares its power supply with.)

Now, correct me if I'm wrong, but that's what I thought the decoupling caps are there to alleviate. What's weird is that in testing, leading up to the schematic in the original post, it made no difference. Just now, I tore down most of it, and built it using a non-inverting configuration for the headphone amp, and lo and behold, the decoupling caps make a day and night difference.

Right now it's:
- StillTrying's preamp circuit, but with a trimpot as the feedback resistor
- unchanged mixer circuit from the original post
- A non-inverting config for output (1uF input blocking cap, then 100K volume pot, 10K input resistor, 10K to ground from the inverting input, 100K feedback resistor, and 100R output resistor)

With this, using a bank of caps on each rail (220u, 10u, 1u, 100n, 1n) seems to completely solve the problem. It's a simple solution, I'm just somewhat baffled as to why it made no difference the first time around.




I'm still trying to wrap my head around impedance -- I think I understand how to manage input impedance, but output impedance is still kinda fuzzy, and what I totally don't get is how pots inserted in between to adjust levels affect impedance. It's unclear to me whether the value of the potentiometer counts as output impedance or input impedance, and thus how to choose a value. Maybe someone can explain this to me?


Thanks again to everyone, I really appreciate it! Every step brings me closer to a finished design, but most importantly the learning on the way!

[StillTrying]

Thanks, I built up your circuit and it works beautifully.

I'm glad you didn't think I was just being awkward for the sake of it. 
You still might have to tweak the gains, and what not. I'll have ago at the "active mixer", which is a "virtual ground summing amp" to me, in an hour.

The bias voltage for an electret mic should be fed from an RC filter, not directly from the noisy power supply voltage.
That's what StillTrying's circuit above does, correct? (The topmost 47K and the 47u to ground.)

4K7 for the top one, followed by 5K6 for the second one would be better. Have you measured the voltage across the mike when it's connected, it's should be close to 5V which is fine.

I did a bunch more experimenting based on your replies, and seem to have narrowed it down a lot (I'm sure this is obvious to you): if the microphone preamp shares the same power supply as the headphone output amp, then the problem appears. If I move either one to a separate power supply (e.g one on battery, the other on bench PSU), the problem disappears entirely. (The mixer doesn't care who it shares its power supply with.)

This can be fixed, often just by connectiing the +- 9V power right near the headphone IC, so that the high currents around there don't have to flow near the circuit's inputs.

[Someone]

I'm still trying to wrap my head around impedance -- I think I understand how to manage input impedance, but output impedance is still kinda fuzzy, and what I totally don't get is how pots inserted in between to adjust levels affect impedance. It's unclear to me whether the value of the potentiometer counts as output impedance or input impedance, and thus how to choose a value. Maybe someone can explain this to me?

Here you are worried about the power supply impedance, consider that amplifier 1 is drawing a lot of current as it operates. How are amplifiers 2 and 3 affected? What are realistic values for the resistors shown?

Now where is it best to place capacitors to reduce cross talk?

[StillTrying]

If I move either one to a separate power supply (e.g one on battery, the other on bench PSU), the problem disappears entirely. (The mixer doesn't care who it shares its power supply with.)

Now, correct me if I'm wrong, but that's what I thought the decoupling caps are there to alleviate.

The audio signal levels at the mixer are 20 to 100 times larger than those at the mike amp, and it's gain can be much lower, so they and it, are much less effected by small variations in the +-9V supplies and any noise on GND.

I forgot to mention that the large decoupling caps should be at the high current end - near where the headphone, it's amp, and +-9V and battery centre GND are connected, so that their higher charging and discharging currents don't flow anywhere near the mike input/amp. Of course you'll need some decoupling near the mike input/amp end, but the main idea is to keep any large currents from flowing in GND near the mike input/amp - because the mike, and its amp's -ve input resistor have to be connected to GND there.

With this, using a bank of caps on each rail (220u, 10u, 1u, 100n, 1n) seems to completely solve the problem. It's a simple solution, I'm just somewhat baffled as to why it made no difference the first time around.

A properly decoupled RC filter type supply for the mike voltage will always help. You can increase the value of that 47u quite a bit if needed, 47u is near the minimum, 100u-150u would be fine, if you increase it to a too massive value you sometimes get that slow oscillation/motor boating effect.

I'm still trying to wrap my head around impedance -- I think I understand how to manage input impedance, but output impedance is still kinda fuzzy,

I'll try.
Impedance is just AC resistance, at DC impedance and resistance are really the same thing, it's really just an alert that the voltages are going to be AC, and some components will act differently depending on what the actual frequency of the AC is - mainly caps which are an open circuit at DC, but turn into a resistor at AC with a resistance value that changes depending on the actual frequency of the AC.

Input and output resistance/impedance together act just like a simple 2 resistor potential divider. If you put 2 10K resistors in series across a voltage, the voltage in the middle of them would be exactly half.
If a 10K output is driving a 10K input then voltage in the middle of them(which is the only voltage that the input sees) would be half - you'd lose half the voltage you're trying to pass on to the next stage if both in and out were the same value.

and what I totally don't get is how pots inserted in between to adjust levels affect impedance. It's unclear to me whether the value of the potentiometer counts as output impedance or input impedance, and thus how to choose a value. Maybe someone can explain this to me?

Well the resistance/impedance of the pot varies with the position of the wiper, and that's the problem!

So you make the pot resistance low, and the input's resistance/impedance high so that the variations in the pot's smaller impedance (with changes of the wiper's position) don't make too much difference in comparison to the sum total impedance of the 2 of them together. 
I'll draw an inverting amp and volume control in a minute, so you that can see the problem.

I've drawn it the way you originally had it with a 100K vol pot and 10K on the -ve input of the amp - which is an inverting amps input impedance.(I don't know what arrangement you've got now).

If you start with the Vol pot at near zero volume - at near the GND end, the op amps -ve input sees just its 10K resistor (GND is zero resistance) and it's easy to work out the gain it's just 100K/10K = 10X.

But, if the pot is moved to near it's middle (50K/50K) the -ve input now sees its 10K resistor plus the 50K of the pot. It's easy to miss, but the gain is not 10X now! It's gone down to 100K/(10K+50k) = 1.67X.
There are other +- effects, but in theory at least you may have noticed that the mixer's volume controls don't work properly between about the first 5% to 40%.

No picture this time. But if you get the impedances of the pot and the input of the inverting amp the right way round by reversing them, it fixes the disappearing gain problem.
10K pot, 100K -ve input resistor, and 1M feedback resistor(which is too high for the NE5532).

With the now 10K pot's wiper at near zero the gain is still 10X of course, 1M/100K = 10X.
But with the 10K pot at its middle 5K/5K, the -ve input now sees it's own 100K resistor, plus just another 5K from the pot, and the gain is now 1M/(100K+5K) = 9.5X which is quite a lot better than the 10X gain dropping down to 1.67X.

I'll post a new scheme pic when I get a round tuit so you can look for any useful bits.

[StillTrying]

I've had this for days while I was thinking. 
But still, using 2 buffers on the inputs seems the only way to create a good mixer using 100K pots.

[tooki]

Hey guys, just a quick reply so you don't think your posts were in vain. I've been working on it, built some up on protoboard, etc. Anyway I've got company coming for a week so I probably won't have time to work on it the next few days. But I really appreciate your help and I'll be back later!

[StillTrying]

Yes, let us know which bits you're trying.
You didn't say that the volume in headphones was low, but in the simulation the power in the headphone started clipping at only 45mW, so adding the helper amp (in the box) increased the power to 160mW before clipping.
The clipping is due to to the output of the NE5532 current limiting.