TLV320AIC3100: When the speaker is turned on, it will power off instantly

Hi,

Make sure the short circuit flag is for class-D page 1 Register 32 and not HP.

If you remove the power for class-D do you still see this flag?

Maybe the short is on the board, try different unit and check your layout and schematic.

Regards.

1. Make sure it's not the headset problem, because the headset works normally
2. We have been using aic3100 for a long time, and many products are using it. This is the first time we have encountered such a situation
3. It has been confirmed that there is no short circuit on the PCB
4. Try to fly the PCBA cable to another adapter board, that is, the sound output is not connected, but the CLASS-D is still POWER DOWN when it is started

I do not understand step 4, are you saying class-D power is off and you are still seeing the short circuit flag on class-D?

Have you tried swapping with different AIC3100 device on the same PCB?

Regards.

1. In step 4, I desoldered the aic3100 on the original pcb, extended the contacts to another vqfn-to-dip adapter board, and only connected the necessary pins. SPKP and SPKM are not connected to any components, so there is no possibility of short circuit. At this time, my software is set to start class-d, or it will be changed to power down instantly, and a short circuit flag will be read
2. Replaced some boards and parts, desoldered the aic3100 that worked normally on the old product, and soldered it to the new product. It is the board that is currently being discussed, but it is still not normal.

OK, understand step 4.

So you replaced the AIC3100 on the new board with a "working AIC3100" and you still see the short circuit flag - correct? That sounds like either there's a short in the new board or the reading of the register flag is faulty. Have you tried removing the DUT on the new board and check for any short? Check also the validity of your software in reading the registers maybe use a digital analyzer to confirm the transactions.

1. I removed the aic3100 on the board and tested it on the pcb, there is no short circuit
2. I measure SPKP and SPKM with an oscilloscope. When measuring with CLASS-D POWER ON, there is no waveform immediately after playing the square wave. Check CLASS-D POWER has been automatically changed to POWER DOEN

So let me summarize what you said above:

1. On your board, you see short circuit flag on class-D
2. You replaced the AIC3100 device with a known good AIC3100 on your board and you still see the short circuit flag
3. You removed the AIC3100 device from your board and there's no short on class-D SPKP/SPKM pins on the board

Are the above correct?

Have you tried to clear or reset the part and ensure the flag is clear before enabling the class-D driver again? You can first try power down the speaker power-stage reset by setting page 1 / register 32, bit D7 and if it does not clear then try using the master reset page 0 / register 1, bit D0. For master reset you will need to enable the audio path again as it will clear all register settings. Once the flag bit is cleared then try your path again.

What's your SPKVDD and AVDD voltage? Do you see any over current in SPKVDD? Maybe you have current limit in the SPKVDD, try isolating that as well.

1. correct
2. I will only read the flag (p0/r44_d7) once, because the power will be cut off instantly after class-d starts
3. Resetting all settings is also the same phenomenon as the second point.
4. There is also an attempt to start class-d immediately after the software is reset, which is also the same phenomenon as the second point.
5. SPKVDD unlimited current, the voltage value is 5.2v
6. Also tried setting class-d to mute
7. Also tried to set the Speaker power-up wait time (p1/r34) to the maximum

8. The AIC3100 is used in many of our products. This is our first time with a new product, and it has pretty much the same circuitry as the others.
please help, thanks

Instead of reading p0/r44_d7, read p1/r32_d0 instead.

It seems you are connecting the codec powers from the same source of SPKVDD.

Isolate your SPKVDD from your board's power, monitor the current and try again and check the register flag.

It does not seem to be the SPK output but likely the power.

What changed from your previous board, maybe issue with the new design.

1. Because class-d is powered off as soon as it is turned on, so read p1_r32_d0, only read to 0
2. In other words, if I set p1_r32=0x86, I will read 0x86 first when I read it immediately, and then it will become 0x06 when I read it again. At this time, read p0/r44_d7 again, and it will be 1 after reading once.
3. Test Separate SPKVDD and SPKVSS, and use 5v/5A power supply to supply power directly. It can be observed that the waveform of SPKP goes from top to bottom first. When the waveform falls, SPKVDD will be pulled to 3.45V for 50nS, then bounce 50nS to 7.35V, and then back to 5V
3. Similar to the old design, because the PCB shape is different, adjust the LAYOUT

On your step 2 above, you just set the register to power up the class-D driver and immediately it powers down, no other registers are configured - is that right? 

• Can you provide the scope capture?
• How do you layout SPKVDD and SPKVSS?
• Can you check the capacitors on your SPKVDD, maybe dry joint or bad cap that is causing that?
• Are you seeing the issue on all boards?

Start class-d, power down immediately
1. Try the settings of the old product
2. After trying soft reset, immediately set
3. Tried at least 5 PCS

4. Try SPKM and SPKP without connecting any external parts

Somewhere on the boards are causing the shorts, maybe isolate/check any cap that's connected to ground like C23/C26 and SPKVDD caps.