This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DEM-PCM2912AEVM: Crosstalk with headset on DEM-PCM2912AEVM

Part Number: DEM-PCM2912AEVM
Other Parts Discussed in Thread: PCM2912A

Hi,

I'm working with the DEM-PCM2912AEVM dev board on Linux. Using an Y-adapter, I want to use an CTIA headset with this dev board. The adapter has 2x 3,5 mm plugs (2 rings) on one end and one 3,5 mm jack (3 rings) on the other end. It joins the GND channels on Ring 2 as well as VIN and MBIAS on Sleeve.

Unfortunately, when using the headset, which works great on my phone, the mic port captures all the sound coming through the speakers. There is no acoustic interference, because you can barely hear the sound over the headphones and the headset mic isn't that sensitive. Is there a way to avoid this?

The headset I'm using is a ISY IIE 1101 blue.

Thanks in advance, Thorben

  • Hi, Thorben,

    Welcome to E2E, Thanks for your interest in our products!.

    The PCM2912A has a Side tone feature that allows the codec to playback the mic content through the headphone outputs, it is an useful feature for self-monitoring applications like phone calls. This feature should be disabled by the host system. In Windows, this can be done in the PCM2912A section of Playback Devices window, on the levels tab. Unfortunately I am not aware of how to disable this function in Linux.

    Best Regards,

      -Diego Meléndez López
       Audio Applications Engineer

  • Hi Diego,

    thank you for your answer.

    I know that hearing mic content through the headphones is a codec feature. In my case, it's the opposite. I capture headphone content with the headset mic during simultaneously playback and record. This is not an acoustical capture since i can also reproduce this with a pair of resistors connected to the speaker ports.

    Best regards,

    Thorben

  • Hi, Thorben,

    Thanks for the feedback. This may be related to the connection of the headset used, have you tied with a different headset?. Also, It would be useful if you can share a diagram of the plug connectors used. The EVM Mic connector is meant to be used with a 2-ring plug, where the sleeve is connected to GND, the ring is connected to the MICBIAS and the tip is connected to the mic input.

    Best Regards,

      -Diego Meléndez López
       Audio Applications Engineer

  • Hi Diego,

    I added a diagram to show you, how I've connected the my headset to the EVM. If this isn't the correct way to do it, how would I connect a 4-pin headset to the EVM or to a custom design?

    I also tried connecting a different headset to my setup with a similar result.

    1488_180212085941_001.pdf

    Best regards,

    Thorben

  • Hi, Thorben,

    Thanks for the feedback.The way you are connecting the headset seems fine. I wonder if the headphone content is leaking into the mic through the headset somehow, have you tied to record the mic signal without having the headphone output connected to the EVM?. I don't expect the leak to occur into he EVM board.

    Best Regards,

      -Diego Meléndez López
       Audio Applications Engineer

  • Hi Diego,

    like I said before, I tried different headsets with the same result. Also using a pair of resistors instead of the headset speakers showed this behavior, but not that drastic. I attached several audio files. ti_LR_demo.wav is my output file. While playing this through the headset speaker, I recorded the mic to ti_MIC_demo.wav. During the record, I unplugged the speaker jack and plugged it back into the EVM. You can head that at second 12, 22, 30 and 46. The last file ti_R_demo.wav is basically the MIC demo but with resistors instead of the headset speakers.

    Best regards,

    Thorben

  • Hi, Thorben,

    Thanks a lot for the additional information. I think this indicates the issue is within the headset cable assembly or headset type. I would like to try this in my side but unfortunately I won't have access to my EVM until next week as I am currently traveling. Please give ma a couple days to check this in my side and try to find a possible solution

    Best Regards,

      -Diego Meléndez López
       Audio Applications Engineer

  • Hi Diego,

    thanks for your afford. I'm looking forward to your test results.

    Best regards,
    Thorben
  • Hi Diego,

    have you been able to test this behavior yourself?

    Best regards,

    Thorben

  • Hi, Thorben,

    Sorry for the late response. I have made some tests with my EVM, and was unable to replicate the behavior described. I must add that I don't have an adapter like the one you are using, I tried with different headsets and connectors, but was unable to replicate the issue with the EVM. 

    Best Regards,

      -Diego Meléndez López
       Audio Applications Engineer

  • Hi Diego,

    that's unfortunate. We're using the IC in an actual design and encountered this problem with our design first. Luckily we were able to reproduce this behaviour with the EVM. However, we still have this issue and need to fix it. Do you have any example schematics for comparison of how to properly connect a CTIA headset to your IC?

    Best regards,
    Thorben

  • Hi, Thorben,

    I have made further testing, now creating my custom adapter and after fixing some issues I found a similar behavior. While doing my adapter, I noticed the headset used was not CTIA so the headphone output content was directly coupled to the mic input as the return path. After fixing this, I was unable to hear the headphone content while recording audio, however, I increased the gain on the headset up to a point where the signal got saturated, and that was when I was able to hear a small tone on the recording data. I disconnected the headset and replicated the test without anything connected to the EVM and found the same behavior. One important thing to note is that in order to hear the output tone on the recording data, I had to saturate and apply a very high gain to both audio output and mic gain. This seems to be an expected behavior of the part while playing a highly saturated signal through the analog output. This part is quite old and right now is very complicated to get detailed information about the internal behavior of the device to see if there is a way to explain this behavior and solve it.

    Best Regards,

      -Diego Meléndez López
       Audio Applications Engineer

  • Hi Diego,

    the jumper settings on the EVM during test were +20 dB and 500 mA. The input mixer setting was set to 33% to get reasonable audio recordings. Output volume was set to 50%. This is no extreme setting in my opinion since you can hear everything clearly in the recording and the output is just on the edge of being to loud for using in-ears once every minute.

    Could you please take a more detailed look at it? It would be great to resolve this issue somehow.

    Best regards,
    Thorben