Communication between MEMS mic and ADAU1701

Hello JCB,

The ADAU1701 eval board has a master clock source on the PCB. So you do not need to feed it a master clock.

There is a switch on the eval board to select between and external master clock or the crystal on the eval board. S1 needs to be set to the right.

All the microphone will need is power, BCLK and LRCLK to get it to wake up and send data to the DSP.

The DSP needs to be properly setup. Attach your project if you want someone to look at it.

Dave T

Hello DaveThib,

Thank you very much for your reply, I am still not able to do what I want to do. My DSP does not even generate the signals on the header J7 to feed the mic. For now I only want to hear what the mic captures via the jack output. I tried a lot of thing but here are the screen of my actual project :

Thank you very much,

JCB

Hello JCB1,

I assume you are using the MINI Evaluation board. J7 would fit that description.

Well, the problem is this eval board was designed to be really simple so some things are difficult to do. The LRCLK and BCLK on J7 are the Serial Input LRCLK and BCLK ports. Those are always slave ports. The output ports that can be masters are on MP10 and MP11. Now, those are setup to drive LEDs on this eval board. So you need to do some modifications.

Remove R33 and R14. This will disconnect it from the LEDs.

Then connect the DSP side of the resistor pads over to the serial input port J7 and also over to the MEMS mic.

This should make a difference.

Let me know if you have further questions.

Dave T

Hello DaveThib,

Thank you for your reply, everything is working fine now !

Have a nice day, best regards,

JCB

Dear Dave,

I also want to connect ADAU1701 with INMP441 like below:

My hardware configuration is same as JCB1, only the I2S word length is 24 bits.

My problem is I hear nothing from MP0_SDATA_IN0.

Please advise me.

Thanks for your time.

Hello,

The serial input port on this part is a little bit different in that the port is always a slave. This is why there is not register setting to set the serial input port to be a master.

So you must use the serial output port clocks to drive both the microphone and the serial input ports. Those are located on the MP pins 4 and 5. So you have to jumper the serial output port clocks over to the input port clock pins and then set the serial output port to be a master and you will see the clocks and it should all work.

Is this what you found?

Dave T

Dear Dave,

Thank you very much.

It is working now after re-connecting MP10 to MP4, MP11 to MP5. However, the Mic doesn't work out of box. I have to use the filter for 250 hz to 2500 hz with gain set to 5 or 10.

My target is to try the vocal effect with this Mic. But there is whistle when I use DAC 0 & 1 with 4 ohm speakers. Please advise me if you have any idea or should I ask this question in other thread?

Thanks for your time.

Hello,

Yes, Mics usually need some sort of processing to sound right.

When you say it is a whistle. Is it feedback or is it a steady tone or whine in the audio? These are very different situations. It is OK to stay in this thread.

Dave T

Dear Dave,

I attach a clip to describe this noise. I hope you can guess it.

This noise occurs when I turn the volume up, via potential meter. This noise also occurs when I speak something.

Please advise.

Thanks for your time.

Hello,

This is definitely feedback. As you noticed, the noise happens when you turn up the speaker volume. Your microphone is very close to the speakers; can you move them further away from each other? If not, we can look into other options.

Joshua

Hello tuongnguyen,

Josh is right, this is definitely feedback. The basic cause is that the microphone is too close to the speakers. There are some tricks to improve how loud it will go before it starts to feedback. There are some rather fancy ways to reduce and control feedback but it takes a much more powerful DSP to do that and I think there are other things you need to do with this DSP.

So here are a couple of things to make it a little better. This is far from perfect solutions but it may help you enough.

The first thing to do is to identify the frequency of the feedback and reduce the level at that frequency. The frequency of the feedback is dependent on the distance of the mics from the speakers and from reflections of the sound from the speakers on nearby things like table tops and walls. So if you change these then the filter frequency will need to be adjusted.

Then you can get it a little louder and then you run into the next frequency that wants to feedback. So you can add a filter section and filter that out. So you can get a little louder then you run into another frequency. So you will never get it so that it will get super loud unless you move the mic away.

Then the next part of the trick that can help a little is to vary the frequency of the audio slightly. If it keeps moving it does not stay at one exact frequency long enough to allow the feedback to build. This method can be good for voice signals. The more the frequency is varied the better it works but then you start to hear it and that is not good. So this is all about finding the balance between hearing the pitch shifts and reducing the feedback threshold.

I have attached a project that shows how to do this. I just arbitrarily chosen the filter frequency and also the frequency of the pitch modulation. Experiment with different settings and see how good you can get it. Keep in mind that what we see in the video is almost an impossible situation with the mics very close to the speakers.

Oh, there is one more technique. that is to get microphones that are not Omni directional. There are ways to shape the pickup pattern using the microphone mounting cavity and electrical methods. These are huge subjects and so I will not go into it here.

I have attached the project

Dave T

ADAU1701 Voltage Controlled Pitch Example.zip

Dear Dave and Joshua,

You are all right. Moving the speakers farther makes it better. This is feedback issue.

Thanks Dave for the example. I need sometimes to digest it.

Thank you very much.