TLV320ADC3101: TLV320ADC3101 Application Issue

Hi, Ian,

I would recommend several things to ensure that the stronger signal is not mixed with the weaker signal. You may try it on the EVM to eliminate the crosstalk effect.

- Go to the Audio Inputs Window on the Analog Settings Menu of the ADC3101 CS.
- Place all the unused inputs as NC (Page 1 / Registers 52 to 57). Then, on the EVM, ensure that the unused inputs are AC-coupled to ground.
- In the unused ADC, select the "Biased Weakly" option in the "ADC Unused Inputs CM" option (Page 1 / Registers 54 and 57).
- Mute the PGA and the ADC that are not in use (Page 1 / Registers 59 - 60 and Page 0 / Register 82).

Additionally, do you have the same results if you place the stronger signal in another differential input? Could you try it with IN2R&IN3R?

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,

Thanks for information, would you pls help check if it can get some information from .cfg?

I'll also check it in EVM again, thanks!!

8473.Hubble_ADC3101.cfg

Thanks, Ian.

Hi, Ian,

I recommend to follow my indications of my past post. Specially on the .cfg section where the analog inputs are selected:

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,
As your message, answer as below,
- Go to the Audio Inputs Window on the Analog Settings Menu of the ADC3101 CS.
- Place all the unused inputs as NC (Page 1 / Registers 52 to 57). Then, on the EVM, ensure that the unused inputs are AC-coupled to ground.
[Ian] Has confirmed all unused inputs have been set 11.
- In the unused ADC, select the "Biased Weakly" option in the "ADC Unused Inputs CM" option (Page 1 / Registers 54 and 57).
[Ian] Have tried it before and it didn’t work.
- Mute the PGA and the ADC that are not in use (Page 1 / Registers 59 - 60 and Page 0 / Register 82).
[Ian] Garmin need to use left ADC/PGA and right ADC/PGA so that I don’t think the suggestion is practicable. In addition, I tried to mute the PGA before but it didn’t work.

And, The registers you listed is what we set now.
But it still has the issue, do you have others ideas about it?
That's serious issue for Garmin, you can also try it in EVM to see if we can help solve it.
Thanks, Ian.

Hi, Ian,

Could you provide the configuration that customer and you used in the ADC3101 CS? I will test it in my side with the TLV320ADC3101EVM in order to accelerate this.

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,

Pls reference configuration at attach. Thanks for help!!

7870.Hubble_ADC3101.cfg

Thank you, Ian,

I will take a look at this and check it with the EVM.

Best regards,
Luis Fernando Rodríguez S.

Hi, Ian,

I tested your configuration on the EVM and I'm getting the left channel affected by the right channel signal. I used the same input lines that you are using (differential pairs IN1L&IN1R and IN2L&IN3L). The problem was eliminated by routing the unused inputs to the common voltage. I configured page 1 / register 54 and 57 / bit D6 as '1'.

Do you have the same results in your EVM?

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,

I modified the registers and did the test again with EVK, but didn’t see any improvement.

Would you pls double confirm?

Thanks, Ian.

Hi, Ian,

I tested again your configuration in my EVM and I didn't get the same results than you. The EVM is working correctly without any crosstalk effect.

So, I have few suggestions:

- Reduce the voltage level at the analog inputs. A high voltage level could generate a crosstalk event even if an input is not being used.
- Ensure that your EVM parameters are correct. WCLK must match with the sampling rate.

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,

That's really strange, could you help check if the setting of the attachment is the same as yours? 

Thanks Ian.Virb360_ADC3101.cfg

Hi, Ian,

These settings are made for a 22.05KHz sampling rate. Does this frequency match with the WCLK value? If not, I recommend to configure the sampling rate as the WCLK frequency.

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,

My customer using LRCLK=48KHz, BCLK=3.072MHz.

Would you pls help me to modify original setting to these condition for me? Thanks.

Sorry I don;t have EVM on have and applied.

Thanks, Ian.

Hi, Ian,

Please take a look at the attached document. It is the modified register configuration with the correct PLL values to generate 48KHz sampling rate from the BCLK input (3.072MHz).

Virb360_ADC3101 (2).cfg

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,

I did the test on EVM with the new setting you send but found that the result was the same w/ and w/o changing register 54 and register 57 on page 1.

In addition, the test result shows when one channel was -9dB and -3dB, the other one channel was -73dB and -42dB.

I don’t think the crosstalk has been eliminated. Could you help send us the file you recorded?

Thanks. Ian.

Hi, Ian,

Based on your observations, could you try with another signal source to determine if it could be a root cause? Do you get the same results with Audacity software?

Best regards,
Luis Fernando Rodríguez S.

Dear Luis,
I want to check a question about setting:
when I input signal into EVM "IN1+" and "IN1" by signal generator, which one is correct as you did experiment,
(1) Signal generator "+" connect to "IN1+"; Signal generator "-" connect to "IN1-" or
(2) Signal generator "+" "-" connect to "IN1+"; and "IN1-" connect to GND.
Thanks, Ian.

Hi, Ian,

The signal generator should be configured to generate a differential signal. So, signal generator '+' should be connected to "IN1+" and signal generator '-' should be connected to "IN1-". In that way, you will send a differential signal.

Best regards,
Luis Fernando Rodríguez S.