TAS5780M: the ASIC issued a SPK_fault but I cannot measure any issue?

Sam,

ch1: fault

ch2: Vpdd current

ch4: Vpdd

zoom in

the voltage is still good enough. Also, from TAS5780M spec, the fault pin is relative with overcurrent on spk_out and over temperature.

I cannot tell the waveform could trigger the fault pin.

But from the TAS880021A spec, the fault pin is relative with overcurrent, overvoltage, under voltage and DC detect event.

not sure which spec is correct?

Can you tell the fault pin is triggered by which condition?

BR,

frank

Hi Frank,

An overcurrent event could be identified on the waveform. The max current is 14.8A according to the measurement. Any idea on what happens here?

Thanks!

Regards,

Sam

sam,

from spec, overcurrent is for SPK_out, right? not for VPDD.

it happens at normal music playing.

Customer found that the issue is gone when they adjust the compensation capacitor at FB pin of buck regulator for audio amp.

failed condition:

pass condition

any idea?

BR,

frank

Hi Frank,

OK, I see. Now I can differentiate the speaker current and PVDD current.

1. The PVDD supply voltage changes on the failed condition, while in passed condition the PVDD is more stable. Any comment on this?

2. Looks like it's related to the power supply to the audio amp. Is the 3.3V supply always good during the audio peak?

3. Are you using the "External Undervoltage-Error Protection" on datasheet Chapter 9.3.7.6?

4. BTW, the PVDD probe channel measurement results does not match with the trace on waveform, I guess there is a offset setting?

Regards,

Sam

Sam,

for 1, I believe that the change for compensation on power source IC makes better response on PVDD. But the drop on PVDD is only from 18V to ~12V. it's still far away from UVLO. not sure what trigger SPK_fault

for 2, yes

for 3, the voltage on SPK_MUTE is stable.

for 4, yes, it's just offset by scope.

any comments?

from TAS880021 datasheet, spk_fault is triggered by overcurrent, overtemperature, under voltage, over voltage and DC detect.

from datasheet we know the oce and ote clear time is 1.3s. 

Do you know what's clear time for other fault?

BR,

frank

Hi Frank,

I need to reach out to design team for the clear time.

Could you read the register 0x6B and 0x6D during the sound interruption? It shows which fault happens.

Thanks!

Regards,

Sam

hi Sam,

before and after spk_fault happens. the value for 0x6B and 0x6D is the same. We read the register every 100ms.

0x6B: 0xD3

0x6D: 0x00

by the way, we already change to page and book 0.

w 94 00 00

w 94 7f 00

r   95 6B 00

r   95 6D 00

Any idea?

BR,

frank

Hi Frank,

Just got a point, the device actually monitors the inductor current, not the speaker current. Similar with a switching power supply, the peak inductor current could be higher than the load current, especially when the output/PVDD is dynamically changing in class-d.

Could you monitor the inductor current and see if it hits the OC threshold?

Thanks!

Regards,

Sam

Sam,

the inductor is in series of speaker. so, where the current through inductor should be the same with the current of speaker, right?

Also, the current on speaker is only ~4A. it's still far away from 7.5A ocp spec.

BR,

frank

Hi Frank,

Class-d amp is similar as a switching power supply, there is ripple current on the inductor, which is related to PVDD, Fsw, Vout. The ripple current is filtered by the Cap. Speaker current is the average inductor current.

Also, could you share your switching frequency setting (pin SPK_GAIN/FREQ), and inductor part number?

Thanks!

Regards,

Sam

Sam,

Let me measure the switching current before the inductor and get back to you.

the switching frequency is 8 × fSYNC.

BR,

frank