RF coupling on the microphone input

Hello Friedrich,

I have worked installing many recording consoles in high RF places and it can be a challenge. This will all come down to how grounding is done in the unit. The shield should have its own path to the place where ground enters the device. Keeping shared paths to a minimum is part of the strategy. There are other PCB layout or wiring issues to pay attention to. 

I lecture on this for an entire class every semester so I cannot go into this on a short post. Plus, there are so many variables here that you have to look at the situation. I have seen issues with the AC power wiring in a facility that was then made worse by not properly grounding the shields on a console. In the case of a microphone that is usually phantom powered it gets ever more difficult. Still, the answer is similar. Make a short direct path to ground for the shields and the chassis box. Then where power comes into the unit then tie the audio ground to this earth ground right where it comes into the box thereby minimizing the shared ground paths. Then the PCB layout needs to be careful to not allow these chassis ground paths to have much capacitance to other ground planes or traces. This will allow the RF to bleed over to the other ground path. 

Are you using USB for power or connecting to a computer? That is always tough because the ground on a USB is always done very poorly with respect to RF. They often have an inductor in series with the ground!!! 

This can really be a tough problem and you have to look at the entire system sometimes. 

Dave T

Hello Dave,

thank you very much for your answer and all the information. In the meantime I found some infos about a suitable filter and I tested it with LTspice. I wll be on site again in the next days to test the filter and to do some proper grounding. Afterwords I will post here.

Thanks

Friedrich

Hello Friedrich,

OK, sounds good. A filter could be good for the last little bit of suppression but I maintain fixing the main source of the problem is usually best to tackle first then perhaps a little bit of filtering helps to reduce it a little more. the thing about testing with a modeling software is that the PCB layout and how power is handled coming into  the unit is all not included with the model. I have seen some attempts to do that which is good but still does not exactly fit what is going on in the unit. You may implement the filter and it makes it a little better but still does not solve the problem. It might just end up being a patch on the problem. I hope the result is good enough for your purpose. 

Dave T

Hello Dave,

today I was at the customers site and I was successful!

I did three things:

First, I made proper grounding as you suggested. At the XLR socket I connected the GND pin 1 with the case and then with the shortest possible wire directly to the pcb's GND. Now most of the noise was gone and it was already quite good.

In the schematic, the microphone is the dynamic mictrophone with an impedance of 600Ohms. R1 is the input resistance of the MAX4062 and C3, C4 are the input decoupling capacitors.

In the second step I added a filter between the XLR socket and the pcb. The filter consists of 4 inductors with 100uH and a ceramic capacitor with 1nF like in the schematic but without C2 and all four inductors are 100uH. This reduced the noise even more.

Then, as the third step I did some trial and error and finally added C2 with 4.7nF. This silenced the noise even more and I am really happy with that. Now the noise is practically gone!

For a next build I suggest to change L3 and L4 to 220uH each as the frequency curve in LTspice looks nicer than with 100uH.

This works perfectly for me and I put this here for future reference.

Many thanks to you, Dave! Your info really helped a lot!

Friedrich