Possible analog audio issue


#1

Hello,
I am experiencing an issue with headset audio input/output.

There is a high pitched whine on the audio lines that can be heard at the outputs and recorded at the inputs.

I’ve recorded a short audio file and analyzed it only to find that there are several harmonics - it peaks at about 4100Hz (-65dB), 8200Hz (-70dB), 12300Hz (-68dB), 16400 (-70dB).

A colleague who’s an electronics engineer, recommended using various power sources, and also some filtering and the result was about the same. It also seems to not have anything to do with any connected audio equipment since the noise is also present and apparently unchanged when one simply shorts the mic input to the ground (over a 10k resistor for just in case ;))

I’d really appreciate any advice on the matter.

Best regards,
George.


#2

Also Here’s some images that might come in handy…

InNoise-Frequency

InNoise-spectral

InNoise-waveform

… and a recording made on the board using tinycap

NoiseRecording


#3

Hi George

I am trying to reproduce your results, which audio input are you using?. I am obviously doing something different, can you help me out. The “wav” file I captured from the board to my PC for analysis is not nearly as noisy as yours. Can you give me the steps you did please? This is what I did:

C:\Users\xxxx>adb shell
root@msm8916_64:/ # tinymix 'MultiMedia1 Mixer TERT_MI2S_TX' 1
tinymix 'MultiMedia1 Mixer TERT_MI2S_TX' 1
root@msm8916_64:/ # tinymix 'DEC1 MUX' 'ADC2'
tinymix 'DEC1 MUX' 'ADC2'
root@msm8916_64:/ # tinymix 'ADC2 MUX' 'INP2'
tinymix 'ADC2 MUX' 'INP2'
root@msm8916_64:/ # tinycap /data/audio/capture_hs.wav -D 0 -d 0 -c 1 -r 48000 -b 16 &
ap /data/audio/capture_hs.wav -D 0 -d 0 -c 1 -r 48000 -b 16 &                 <
[1] 4907
root@msm8916_64:/ # Capturing sample: 1 ch, 48000 hz, 16 bit
kill -2 4907
kill -2 4907
root@msm8916_64:/ # Captured 1126400 frames
exit
exit
[1] + Done                 tinycap /data/audio/capture_hs.wav -D 0 -d 0 -c 1 -r 48000 -b 16

C:\Users\xxxx>adb pull /data/audio/capture_hs.wav
2929 KB/s (2252844 bytes in 0.751s)

When I am done I have a wav file on my PC, I haven’t put the file into the audio analyzer yet, but listening on the headset the noise level is far lower that your sample.

Full Disclosure: I am an employee of Qualcomm Canada, any opinions expressed in this or any other post may not reflect the opinions of my employer.


#4

Hi George

I ran my captured audio file through Adobe AU and created the same plots you have. The FFT shows everything below -98dB, with a very slight bump up to -93dB at 4.2kHz. The spectral display also shows a dim line at 4.2kHz, but nothing at the harmonics.

The board I am using is from “current production” so I believe it should be the same as yours. I am not sure where you are picking the noise up. I am (currently) unable to reproduce your problem.

http://i66.tinypic.com/4rth5x.jpgcapture_hs_FFT

Full Disclosure: I am an employee of Qualcomm Canada, any opinions expressed in this or any other post may not reflect the opinions of my employer.


#5

Hi ljking,
Thank you for a swift and detailed response.

Basically I used the same setup as you did:

my tinymix setup was default +:

'MultiMedia1 Mixer TERT_MI2S_TX' 1
'DEC1 MUX' 'ADC2'
'ADC2 MUX' 'INP2'

and I was using pins 4 (GND) and 6(CDC_MIC2_P) on the J7 analog expansion connector.

Very encouraging to see that the noise must somehow be external. I’ll try to run the audio test with the board shielded and will get back with the results.

Best regards,
George.