PDM microphone glitch at beginning of the recording

[frnt]

Hello,

I'm using the IMP34DT05 microphone using the I2S interface. I'm sampling at 22kHz but I have a problem when I look at the recording saved on micro SD. In the first part of the audio, I see a glitch and then an exponential decay of that spike (see attached image).

The strange thing is that not all the recordings have such a glitch at 0.046 s, which is exactly after the first 512 Bytes (see second screenshot).

The PDM data are converted to PCM using the PDMFilter function. Any suggestion on how to get rid of this? and the reason why is happening?

 

[frnt]

Dear @SPozz.1​ ,

I have checked the power supply with the scope and I have no more than 50 mVpp when the power supply is on on the STEVAL-MIC003V1 microphone VDD connettore pin.

However, when after starting the acquisition with DMA_Receive, the noise increases a lot I have regular spikes that reaches about 350/360mVpp in some cases. The regular peaks have a period of about 13kHz...

I have also checked the PDM clock (also at the connector pin) it is about 1.4MHz but it has strong overshoot and undershoot of about 0.5/0.6V

 In the following the signal recorded again in anechoic chamber.

[SPozz.1]

Please confirm that the scope probe alligator clip was connected to gnd very close to the point you had placed the probe tip. Large probe ground loops may lead in false detection of noise and over/undershoot behaviours (ringing), The probe should be switched to attenuation of ten (x10) to reduce the signal load.

If confirmed, the ringing of the clock may be too high, try to configure the pin speed to a lower one (change GPIO_SPEED_FREQ_VERY_HIGH to GPIO_SPEED_FREQ_HIGH) and/or, if not already present, cut the clock trace as close as possible near the MCU/driver and put a 33 Ohm resistor in series. I'm not convinced that this solves the noise problem but it can be tried.

Add 10 to 22uF in parallel to the microphone, and if possible cut the VDD trace and put a 10 Ohm resistor in series to decouple the noisy part of the circuit from the mic power supply.

[frnt]

Hello @SPozz.1​ ,

yes, I have used the alligator clip close to GND. I were right, the GPIO for some weird reason was set to high-speed and changing it to low reduced almost completely ringing on the clock PDM signal.

On the contrary, the pikes on the VDD signal are still present even if their amplitude has been reduced. Now I have about 200/250Vpp. I have also checked the DATA signal and it is noisy.

The microphone I'm using is the one from the STEVAL-MIC003V1 board. The microphone is connected through a 10cm dupont cable to the board.

[frnt]

Dear @SPozz.1​ ,

I forgot to attach the recording in the anechoic chamber.