TPA2011D1: Pop noise reduction at the EN On

Hi Tom,

So to clarify, the pop was improved, but not fully, when switching to a 0.47uF?  The threshold where to the pop occurs is reduced?  I will investigate.

-Jeff

Hi Jeff,

I changed the capacitor to 0.47uF and the noise became 0.68Vpp.

However, I think the noise of 0.68Vpp is still large.

Please tell me how to make the pop smaller.

Tom

Hi Liu,

Jeff is on timebank today and will reply to you on Monday.

Best regards,

Jeff McPherson

Hi Tom,

I see that you stated "Pop noise occurs when the EN is turned On from Off  at 2.0Vpp, cycle 15us".  Do you mean by the cycle 15uS, you toggle EN every 15uS on and off?  How long do you do this for?

-Jeff

Hi Jeff

Turn on EN only once after 4 seconds from AMP power on.
For 1uF, the period of POP noise is about 15us.
It continues for about 125us while attenuating from 2.0Vpp.
Do you have a good solution?

Tom

Hi Tom,

Can you try to go to differential output and see if the pop goes away?  Looks like the mis-match from the input inverting/non-inverting can cause this.  Can you send the schematic

Hi Jeff,

Since the previous stage is single-ended output, differential input is not possible.

I can't send you a schematic, but I followed the single-ended example in the datasheet.

Ci=1uF, Ri=27kΩ1%

POP noise is large, so I changed Ci to 0.47uF+-10% to suppress POP noise.

POP noise is now 0.68Vpp.

 Is it possible to further reduce pop noise with single-ended?

Tom

Tom,

Can you select the tightest tolerance for components Ci and Ri (from your example schematic) such that the inverting and non-inverting are matched?

-Jeff

Jeff,

The accuracy of the 0.47uF capacitor is +-10%.

0.47uF capacitors with +-5% accuracy are not commonly available.

Is there a way other than accuracy?

Tom

Hi Tom,

The device has a output offset is around 1mV but up to 5mV that can cause this.  Unfortunately, it appears from what I can gather is that the pop noise can be mitigated, but not fully removed.

Hi Jeff,

I know that noise cannot be completely removed, but I just want to further reduce the noise. 

Is there no way to further reduce noise?

Tom

Hi Tom,

Can you try 0.1% resistors for Ri?  Also, how are you terminating Ci (the single ended input) when doing this test?

-Jeff

Hi Jeff,

Unfortunately, Ri is 27 kΩ and the size is 0.3 mm × 0.6 mm, so there is no 0.1% product.

Does 0.1% reduce POP noise considerably?

The IN+ line side Ci is directly connected to AG.

The IN- line side Ci is 0.8mm long and is connected to the previous analog circuit.

Tom

Hi Liu,

First, I would check if the pop amplitude changes if you have the input connected to a) GND and b) the AG.

If you can't get a 0.1%, could you hand test some such that they are matched with 0.1%, - just to see if it does indeed help?

-Jeff

Hi Jeff,

I tried your suggestion.

Since Ci and AG cannot be separated, connect AG and GND to Ci,

There was no difference in POP noise.

I removed two 27k resistors from the board and measured the resistance.

It was 26950Ω and 26940 Ω. The difference was 0.04%.

When I tried attaching 27000Ω and 27010Ω as a trial,

Pop noise was the same.

Are there any other measures?

Tom

Hi Tom,

For an experiment to determine if we are on the right path, if you can, you could try to short the inputs such as in this attached schematic.  Please let me know if pop persists

Hi Jeff,

I tried your suggestion.

I removed Ci2 on the GND side and connected Ri2 and Ci1 with 2mm.

Even if the EN signal changed from L to H, the pop noise level did not change.

The length from AMP to Ci is 3mm.

The load is 100mm, 8Ω resistor.

The difference due to variations in the LC circuit for measurement is within 0.2Vpp.

Is the POP noise due to the difference in the charging current of Ci?

Tom

Hi Tom,

Our team is on holiday. Jeff will get back to you by the end of the day Monday. Please do not reply to this message unless we have not answered you in a timely manner.

Thank you for your patience,
Jeff McPherson

Hi Tom,

I found the EVM and am going to try to replicate your issue tomorrow.

-Jeff

Hello Tom,

I am seeing on my setup that the TPA2011D1 does have some Vrms peak from enable to disable - regardless if it's single ended or differential.  There is an advantage for no pop from disable to enable using diff input, but I do not see how to decrease the disable pop.