Apalis iMX8QM SGTL5000 Codec probe fail after rebooting with headset attached

Hi Support Team,

We are testing analog audio on Apalis iMX8QM and found that SGTL5000 codec met probe error after software rebooting via command with high-impedance headset attached on Ixora carrier X12. if no audio device on X12 or only small low-impedance ear set attached, all are okay. Detailed haredware set up and test process and log please find below, and kindly please help check and advise?

./ set up
Apalis iMX8QM 4GB WB IT V1.1B with Ixora carrier V1.2A, Ycoto linux Multimedia demo image latest quartly V5.3 installed.

Headset - Sony WH-1000XM3 and another local brand big headset (bluetooth off)
ear set - iPhone original small ear set

./ test process
a). attach Headset to Ixora X12 and power on apalis imx8, after booting all is fine

root@apalis-imx8:~# dmesg |grep sgtl
[    2.491235] sgtl5000 3-000a: sgtl5000 revision 0x11
[    3.292400] debugfs: Directory '59050000.sai' with parent 'apalis-imx8qm-sgtl5000' already presen
t!
[    3.302046] asoc-simple-card sound: sgtl5000 <-> 59050000.sai mapping ok
[    3.379712]   #1: apalis-imx8qm-sgtl5000

b). then rebooting via command
root@apalis-imx8:~# reboot

c). after rebooting, the sgtl5000 probe error, and please find full failure boot log in the attachment.

root@apalis-imx8:~# dmesg |grep sgtl
[    2.531016] sgtl5000 3-000a: Error reading chip id -5
[    2.536547] sgtl5000: probe of 3-000a failed with error -5

apalis_imx8_sgtl5000_probe_failure.log (44.6 KB)

./ addtional information
a). if switch to small ear set attached or no audio device connection, everything is okay for several times rebooting. Looks like the issue is related to the audio device load.
b). have also tested apalis imx8 with Ycoto Linux BSP V5.4 nightly 09-01 version, ycoto linux BSP quarterly V5.2/V5.1, all have the same issue.
c). for comparison, with the same HW set up except changing SOM from apalis imx8 to apalis imx6 with ycoto linux quarterly downstream or upstream V5.3 installed, everything is fine for several times rebooting.
d). the issue can be reproduced on multiple apalis iMX8 SOMs, so it is not a single case issue.

Thanks and Best Regards
Hai

hello,

which kind of 4 pin headphone connector standard are the headphones are following that you are using OMTP or CTIA. there are multiple pinouts. We are following the cita AV standard.

I just like to know that for our investigation.

Best Regards,

Matthias

Hi @matthias.tx

actually all the big headphone we had in hand like Sony/Bose ones are use 3-pin definition like below, it should compatible with both standard. And we also take further test that keep this 3-pin cable attached but not connected with headphone, there is no such issue, the issue happens only with headphone attached.

After some further testing, we got some progress for the root cause -

we use oscilloscope to check V3.3V_AUDIO voltage on apalis imx8 SOM test point

./ when no headphone attached or the second time rebooting(sgtl5000 initialization failing), then reboot, the voltage triggered as below, low to 720mV.

./ when sony headphone attached, then reboot, the voltage triggered as below ,only low to 1.4V, seems not enough to reset the SGTL5000 codec, make it fail to probe.

./ also compared when apple small earset attached, then reboot, the voltage triggered as below, low to 800mV, slightly higher than nothing attached, but i think still can make codec reset.

so finally, looks like this issue is not a bsp issue, but a hardware issue instead.

Hello hai,

we did find the root cause of this. are you already informed about this ?
Its the serial caps that can not discharge fast enough when high impedance headphones are used.

Best Regards

Matthias