TAS5132: THD+N cannot match specification

Hi Ken

   Could you please share the test results at your side? We would like to check which area and how much difference from the datasheet plot. Usually our datasheet plot is not from EVM, we have special validation board to test these data, which board area is much larger, try to remove all the other factors outside of our device. 

Hello,

We want to check the THD+N vs Output power in spec sles190 at page 11, Figure 1.

Our test result after setting TAS5086 according to spec sllu097, section 3.5 Operational Sequence and Indicators flow is shown below:

The diagram is measured by AP, and EVM's PVDD is set to 18V provided by Power Supply.

Since it is totally different from the beginning. (suppose the beginning should less than 0.01%)

The EVM is connect to 8-ohm load.

The AMP reset LED turn off after use configure file inside CD, file name : TAS5132V2EVM_BTL.cfg

Is there any suggestion about environment setting ? thanks.

Hi Ken

   At the input channel A of AP's Filter, may I know why the both of the input port are used? The Red and Black wire seems to connect to EVM's output directly, what is the Bal input wire used for?

  And there's additional twisted wire connected from the load, may I know where it goes? 

   Your waveform seems has big problems. Larger than 1% THD, is already serious distortion happens. We could firstly check in idle state, what is the noise level. Then Play a fixed frequency, do FFT analysis of the output, check what is the distortion frequency. And when you test THD, did you enabled the filter in the AP? Like below.

Hi Shadow,

Shadow He said:

At the input channel A of AP's Filter, may I know why the both of the input port are used? The Red and Black wire seems to connect to EVM's output directly, what is the Bal input wire used for?

The Bal input wire only plug on machine, which is not used during the experiment.

Shadow He said:

And there's additional twisted wire connected from the load, may I know where it goes? 

The twist wire is dupont line for other experiment test use, in our experiment we use alligator clips connect both ends of the 8 ohm load

We test the mute FFT, the result is below:

Then we test -90dB/-60dB/-3dB FFT analyzation:

The above diagram is -90dB FFT result, seems like mute

The above diagram is -60dB result

The above diagram is -3dB result, behavior seem like 0dB

Shadow He said:

And when you test THD, did you enabled the filter in the AP? Like below.

we apply filter setting in AP during the FFT analysis.

Any suggestion for the TAS5132 environment setting? thanks.

Hi Ken

    From your waveform, seems the environment noise is too large, we need to get rid off them first before continue. From the datasheet the typical noise level from TAS5132 is between 50μV to 200μV, we need to keep the environment noise lower than this. In your waveform, at 60Hz there's -34dBV, is already 20mV. Also it's quite strange at 50KHz to 100KHz there's another peak, usually after you enabled the filter, the level will quickly drop when frequency higher than 20KHz.

   We suggest to check below things:

   1. Check the power supply, usually the low frequency noise may coming from the source. Use AP to test DC power directly, see if there's anything we could do.

   2. Remove all the unnecessary wire on the board, on the load, also on the AP. The wire could easily catch the environment noise.

   3. Enable and disable the Filter in the AP, compare the frequency sweep results, see if it really works at higher frequency. 

Hi Shadow,

Shadow He said:

 1. Check the power supply, usually the low frequency noise may coming from the source. Use AP to test DC power directly, see if there's anything we could do.

I have no idea about this issue, could you provide your test setting of power supply in AP?

Shadow He said:

 2. Remove all the unnecessary wire on the board, on the load, also on the AP. The wire could easily catch the environment noise.

The 8-ohm load's wire and measure tool's input is clean is the following test environment:

Shadow He said:

3. Enable and disable the Filter in the AP, compare the frequency sweep results, see if it really works at higher frequency. 

From our experiment, the setting for filter off is shown below:

set the left window setting to none or A-weighting for filter switching off/on testing.

the FFT result is shown below:

This one is Filter setting off result

And this one is Filter setting on result, seems nothing changed.

Can you provide your test environment setting script and AP's test script files ? thanks.

Hi Ken

Ken Ding said:

could you provide your test setting of power supply in AP?

Use the same FFT analysis is fine, use AP to test DC power supply directly, check the FFT results. Use oscilloscope to do the FFT is also fine.

If choose filter didn't work, how about change the Bandwidth directly? Maybe caused by the setting of your software, how about contact your AP supplier help you with this part? 

Hi Shadow,

Shadow He said:

Use the same FFT analysis is fine, use AP to test DC power supply directly, check the FFT results. Use oscilloscope to do the FFT is also fine.

We use oscilloscope to measure power supply output FFT:

The power supply setting is shown below:

Shadow He said:

If choose filter didn't work, how about change the Bandwidth directly? Maybe caused by the setting of your software, how about contact your AP supplier help you with this part? 

We test other class-D output result, seems OK for AP's measurement

Can you provide reference TAS5132 test environment setup and test setting script? thanks.

Hi Ken

I will measure the noise floor at EVM today and share to you