Better mic solution for power constraint application than PDM

RE: Better mic solution for power constraint application than PDM

Raoul — Thu, 09 Nov 2023 15:41:18 GMT

Hi,

I see, thanks. I don't have anything to add in that case. Good luck with testing!

Best regards,

Raoul

RE: Better mic solution for power constraint application than PDM

Hossein Kasaei — Thu, 09 Nov 2023 07:45:53 GMT

Hi Raoul,

[quote userid="117489" url="~/f/nordic-q-a/105386/better-mic-solution-for-power-constraint-application-than-pdm/454804"]Those numbers look about right. From the PS you can see that the PDM + easyDMA consumes around 1.4 mA. I believe the HFCLK consumes around 900 uA.[/quote]

Thanks for your confirmation.

[quote userid="117489" url="~/f/nordic-q-a/105386/better-mic-solution-for-power-constraint-application-than-pdm/454804"] Try the 3 methods on a DK, and measure their power consumption. I'm sorry to say that I don't have any better alternative.[/quote]

Don't mention it! I might test I2S to check the difference in power, but anyway I found that if I'm sending at least 100 bytes through BLE, the power goes high enough that even using different mic doesn't help.

[quote userid="117489" url="~/f/nordic-q-a/105386/better-mic-solution-for-power-constraint-application-than-pdm/454804"]Have you given any thought to the bit depth that you need? How low quality can you get away with?[/quote]

We rather to have at least 16 bit resolution and the ADC doesn't offer a good one, That's why I think of PDM and I2S.

RE: Better mic solution for power constraint application than PDM

Raoul — Thu, 09 Nov 2023 00:04:27 GMT

Hi Hossein,

[quote user=""]Currently MP34DT05 is the one we are using but this a PDM mic ,and as I understood Nordic chip like nRF52832 doesn't offer low power PDM peripheral and I'm getting the average current consumption of around 2.5mA[/quote]

Those numbers look about right. From the PS you can see that the PDM + easyDMA consumes around 1.4 mA. I believe the HFCLK consumes around 900 uA.

I inquired internally on the power consumption of our I2S peripheral, but it seems that we don't have data on this. A developer estimated that the consumption would be about the same as with PDM.

They also suggested that the best way for you to proceed is to do tests yourself. Try the 3 methods on a DK, and measure their power consumption. I'm sorry to say that I don't have any better alternative.

Thanks for mentioning your required sampling rate. Have you given any thought to the bit depth that you need? How low quality can you get away with?

Best regards,

Raoul