PCM2903C: PCM2903C

Other Parts Discussed in Thread: PCM2903, PCM2903C, PCM2900C

Tool/software:

I noticed that as soon as the PCM2903 connects to the Android USB host, the PCM2903C begins drawing high current as if it is streaming audio, approximately 49 mA. This is despite no Android apps being active. This persists, not dropping after a few seconds as I expected, and makes the PCM2903C unusable in our application due to the impact on the products battery life. I see isochronous USB data exchanges with the PCM2903C as long as the cellphone is active, stopping only when the cellphone goes into standby mode after a period of no user activity. Is this behavior avoidable?

  • Hi Gerald,

    The PCM2903C is a USB Class Compliant device and should behave like any other sound device from the Android OS point of view. I'm not an expert on Android OS but the solution would be to stop the OS from sending the exchanges to the PCM2903C. Once the PCM2903C stops seeing activity for 5ms it will go into its suspend state the the power consumption will drop to ~250uA.

    Best regards,
    Jeff McPherson

  • That is what I expected, but all apps are disabled as far as I can tell.  I'll check other models.

    So, before I close this, is it accurrate to say  "there is no way that the CODEC itself can cause the USB connection to remain active? "

  • Hi Gerald,

    Yes that's correct.

    Best regards,
    Jeff McPherson

  • Jeff,

    When you say "any activity" do you mean any USB activity at all, or do you mean audio packets. The datasheet doesn't make a distiction.

    I examined the USB bus and even when USB audio streaming ends there are still periodic very short packets, with a single packet sent every 1ms. This appears to keep the PCM2903C active indefinitely, though no audio is being streamed. As I mentioned in a presious support case (1426487) this peiodic activity effectively prevents the SSPND output from goinf TRUE (logic low.)

    I have confirmed that with a commercial USB headset that uses a RESLTEC CODEC, the periodic brief packets are still sent in the same manner.  I also tested usign a different USB host device (a Samsung TabA) , and confirmed the same USB activity was present, beginning about 10 seconds after audio streaming ends and continuiing indefinitely.

    So, if this behavior is standard USB behavior for Android devices, and if there is no workaround, this looks like it makes the PCM2900C very unattractive for portabel (i.e. battery powered) devices.

    Thoughts?

    Gerald

  • Hi Gerald,

    Those headsets might be including some other "debounce" type logic to maintain a lower power state even though the OS is sending short packets, probably to keep the system recognized by the OS.

    I agree that it presents a big issue for you, but unfortunately this is a case where the codec is going to do what it was designed to. My only other thought is that the codec was designed around USB1.1 and now that we're on USB3.0 maybe some design assumptions/expectations have changed. We haven't introduced any codecs past USB2.0.

    Apologies for the inconvenience,
    Jeff McPherson