Square wave noise filter for a Microphone preamp

[Zedman]

Please help me with this:

I made an RF audio TX device, it has a mic input digitizes the voice and sends it over 900 MHz to the RX device which than records it.

The problem is that the Mic is located close < 8cm to the Antenna and I get the 16ms (~60Hz) sending noise. It's approx a 1/3 ON and 2/3 OFF square wave.

It's for voice only and I need only the 300-4000Hz band.

I tried to build a 4th order bandpass filter, but it's useless for square waves.
I cannot place the Mic much further because it's a handheld device and the box is not as big.
As it's running on a dsPIC, should I use digital filtering?

Is it possible to eliminate this, how do they do it on GSM phones?
Any suggestions?

thanks,
Zedman

 

[FvM]

The problem is, that RF is demodulated in your circuit. It's not a problem of filtering in the audio band. The demodulation most likely occurs in the microphone (if it's an active, e.g. electret device) or in the amplifier input stage. Small RF blocking capacitors, ferrite beads or series resistors in the signal line can usually suppress the RF easily.

You hopefully verified, that the problem is actually caused by the antenna output and not simple supply voltage ripple or similar.

 

[saeed_pk]

yes FvM is right you have to block RF which is being injected in your circuit use beads it is an easy and effective solution.

 

[Zedman]

Thanks for the quick answers,

I did place a lowpass filter on the Vcc line for the mic, because of the ripples in the Vcc line caused by the 80mA sending current. And it's acceptably clean now.

It is definiately injected by the RF signal. If I remove the Antenna or place the Mic further the noise gets much lower.

What type of RF blocking capacitors / ferrite beads to use? Could you recommend some type / value?

It should be placed immediately after the Mic?

thanks,
Zedman

 

[FvM]

As blocking capacitors, small SMD chips, e.g. 47 - 100 pF should work. Some electrete microphones have them already built in. SMD ferrite beads in the 300 to 1000 ohm (Z@100 MHz) range should work. The impedance is usually already decreasing in the UHF band, so a 300 to 1000 ohm resistor is even better, if the low frequency impedance doesn't matter in your application.

 

[Zedman]

Okay,

thanks, will get caps and ferrite beads and resistors too, and we'll see which will solve this best.

thanks again,
Zedman

 

[Zedman]

Here are the results so far:

- capacitor 47pF / 100pF between the mic signal and gnd: a tiny bit lower noise
- resistor 100 ohm: nothing, 1000 ohm much better but still high (in series in the mic signal)
- ferrite bead: 100 ohm@100MHz like the resistor above (in series in the mic signal)

- ferrite bead: 1000ohm@100MHz MUCH lower noise, almost accaptable, but if it's amplified out to a loudspeaker it still can be heard and is annoying.

I noticed if I touch the test point marked "GND" on the attached pic the noise gets MUCH lower to a level what would be the goal if it's possible to reach.

Is there anything I could do more to make it lower?

thanks,
Zedman



Audio part schematic:

 

[biff44]

Is there some way to send the microphone signal back to the op amp as a balanced signal? Then if RF leaks onto the microphone signal traces, a simple 200 pF cap across the balanced like will short it out WITHOUT needing a good RF ground point to act on it. You could experiment with balanced capacitors in parallel, and additional capacitors in shunt to ground (ground meaning some AC ground point as close to the op amp as possible.)

 

[Zedman]

As I understand the Mic picks up the 900 MHz noise from the air so it's already on the output of it. If I feed it to an opamp and get it's inverted output to make a differential signal the noise will already be on the inverted signal in the opposite direction, thus no way to filter it out.
Or I am missing something.

 

[FvM]

It's not clear, if the RF signal is demodulated at the microphone, the OP or even both devices. But clearly, only these active devices are susceptible to interferences. A capacitor across the microphone pins can be tried too.

 

[Zedman]

What about this:

https://www.st.com/stonline/products/literature/ds/6650/emif01-10005.pdf

would this help?

Added after 2 hours 28 minutes:

Uhh, it doesn't really count, because I cannot get one in a timely manner...