OPA4172: Audio signal bleeding into other channels (QUAD microphone preamp)

Andreas Metaxas

Part Number: OPA4172

Hey

I've created my first PCB design
using the design from http://www.ti.com/tool/TIPD181 as a basisfor the quad channel preamp OPA4172.
circuit wire up:
https://imgur.com/1e0fYMw

Pcb Design without ground plane showing:
https://imgur.com/DmMtIYb

Pcb with ground plane:
https://imgur.com/5kgW2RJ

My issue is that all my inputs (mic 1-4) are all showing up across all outputs. Eg; mic 1 will show at all outputs, 2 at all outputs etc etc.
I've tried making another circuit with no ground plane in the case that the PCB tracks were crossing with the outputs but it sounds awful and still has the same issue.

Thanks

over 5 years ago

0 Andreas Metaxas over 5 years ago

Prodigy 110 points

Edit: Audio feed is perfectly fine, just the signal is appearing at each output

0 kai klaas69 over 5 years ago in reply to Andreas Metaxas

Guru 140200 points

Hi Andreas,

how much layers does your board have?

What microphones do you use?

All the decoupling caps must be located much closer to the OPAmp. Otherwise the OPAmp can become unstable.

Kai

0 Andreas Metaxas over 5 years ago in reply to kai klaas69

Prodigy 110 points

Hey , this one has 4 layers, one being a dedicated ground layer.
the recording audio was suitable for what im doing, could the decoupling caps really cause such an issue ?

0 kai klaas69 over 5 years ago in reply to Andreas Metaxas

Guru 140200 points

Hi Andreas,

please don't feel offended, but the layout is highly disadvantageous. :-)

You should move the components much closer to the OPAmp. This is especially true for the decoupling caps and filtering caps. You should avoid long copper traces and minimize loop areas. And last but not least, you must strictly separate input and output lines.

If you use microphones with cabling, take shielded cables and connect the cable shield to signal ground.

Again, what microphones are you using?

Kai

0 kai klaas69 over 5 years ago in reply to kai klaas69

Guru 140200 points

Hi Andreas,

I have a bit analyzed the circuit of the www.ti.com/.../TIPD181.

It seems that this circuit does not allow any cabling at the input because the microphone line directly connects to the -input of OPAmp. Any cable capacitance would erode the phase margin and destabilize the OPAmp.

I think a much better approach would be to directly connect each microphone to its individual OPAmp circuit and to avoid any cabling at all. Just like it can be seen in the link. So, it seems to be better to build four individual amplifiers.

Kai

0 Andreas Metaxas over 5 years ago in reply to kai klaas69

Prodigy 110 points

Fair enough i was hoping that there would be another reason on this weird crossfeeding.
I've got 4 individual ones ,although through-hole variant so that should cover me till i order the soic ones in.

Thanks for taking the time to answer my query !

0 Andreas Metaxas over 5 years ago in reply to kai klaas69

Prodigy 110 points

Hey Sorry i didnt answer your prior question

Im using a cmi-5247tf-k
dont pay attention to that 1uf cap as thats was a mistake

0 Tim Claycomb over 5 years ago in reply to Andreas Metaxas

TI__Genius 15410 points

Hi Andreas,

Here are a few blogs I've written on PCB layout of an op amp to help in the next revision of your PCB.

The Basics: How to layout a PCB for an op amp

How to layout a PCB for high-performance, low-side current sensing designs

Kai is correct in that long cabling can effect the stability of the amplifier so I recommend either connecting the microphone to the PCB or changing the feedback capacitor to ensure the amplifier is stable with the cable capacitance included in your design.

Thank you,

Tim Claycomb

0 Andreas Metaxas over 5 years ago in reply to Tim Claycomb

Prodigy 110 points

appreciate it !
il check it out

0 Andreas Metaxas over 5 years ago in reply to Tim Claycomb

Prodigy 110 points

Sorry, i had just checked my image that i had attached at the start of the post and had noticed i attached the wrong image.

For some reason my feedback resistor should be 12p, along with r4 and c6 being 50 olm and 2.2uf respectfully.

Either way i appreciate the assistance, i resolved the issue using 4 separate boards until i can produce an ideal board

Regards
Andreas

0 Andreas Metaxas over 5 years ago in reply to Andreas Metaxas

Prodigy 110 points

should of read 12p*

0 kai klaas69 over 5 years ago in reply to Andreas Metaxas

Guru 140200 points

Hi Andreas,

you could use a modified circuit which is more immune against the cable capacitance of electret microphone.

What cable length do you expect?

Kai

0 Andreas Metaxas over 5 years ago in reply to kai klaas69

Prodigy 110 points

for the microphone end, about 10 cm, the analog end about 20cm

unfortunately i cant do it directly on the board as i need some distance for the project to work

0 Paul Grohe over 5 years ago in reply to Andreas Metaxas

TI__Mastermind 49906 points

Hello Andreas,

Is the feedback cap R2A-D really 2.2uF?? Compare to the reference design...it should be around 20pF.

If so, most likely that is your problem. There are many things going wrong if this is the case - mainly the various mic signals are sneaking around the amplifier and the "other" amps are back-driving the signal path.

Remove the 2.2uF C2's and replace with a few pF and it should work.

Also - increase the size of C1. The microphones have very low supply rejection (PSRR) - so any "modulation" of the supply will appear on the inputs slightly attenuated.

0 kai klaas69 over 5 years ago in reply to Andreas Metaxas

Guru 140200 points

Hi Andreas,

I would use the following scheme:

This scheme allows a somewhat longer microphone cabling. Also, the supply voltage of the electret microphone is filtered. This can considerably stabilize the amplifier, as Paul already mentioned.

andreas.TSC

Kai

Amplifiers

Amplifiers forum

OPA4172: Audio signal bleeding into other channels (QUAD microphone preamp)