THS4531DGKEVM: THS4531DGKEVM

Philipp Hack

Part Number: THS4531DGKEVM
Other Parts Discussed in Thread: THS4531, THS4561, OPA1637

Dear TI Support Team,

I work as an audio engineer for the Audio company Bang & Olufsen. I am evaluating the EVK of THS4531 at the moment. We plan to use this fully differential amp as a mic preamp that drives a differential ADC Input. The ADC's internal PGA gain is too noisy for our needs, especially at higher gains.

According to datasheet, the THS4531 provides a noise density of 10nV/sqrt(Hz) which corresponds to -117dBV (20Hz to 20kHz) at a gain of 1 (using 2k for Rf and Rg). However when I measure differential in to differential out with an Audio Precision, I can only measure -110dBV (A-weighted) at the THS4531 output. I use 50cm long balanced XLR cables and a source impedance of 40Ohm. The supply is 3.2V taken from a battery, single supply. I removed everything expect for the IC and Rf&Rg from the input and output path. So the noise-floor is 7dB higher as the spec in the datasheet. Do you have an idea what is going wrong here ?

When I change Rg to 470Ohm this yields a gain of 12dB. Here I measure -103.5dBV A-weighted at the output, which corresponds to -115.5dB input referred noise. For our purpose we would need an output noise of -110dB @ 12dB gain. Would this be possible with this amp?

Best Regards,

Philipp

over 1 year ago

0 Michael Steffes over 1 year ago

Guru 47465 points

Hello Phillip,

No way to tell without your schematic - but,

The THS4531 is the highest noise FDA in the product line. You might want to look at the THS4561,

There are few articles I did in AudioXpress - some of them dealing with FDA noise (including popcorn) that you might find interesting, listed here,

Summary artilce list for AX series.xlsx

0 Philipp Hack over 1 year ago in reply to Michael Steffes

Prodigy 10 points

Hi Michael,

thank you for your fast response. I will definitely look into some of these articles.

Unfortunately the power budget that is available is not very high. So 0.75mA of THS4561 is already too much since we will most likely need 4 of those.

I have attached a picture of my schematic. Red crossed components are NC.

0 Michael Steffes over 1 year ago in reply to Philipp Hack

Guru 47465 points

You should not be unaware of this discussion,

https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/505185/ths4531a-popcorn-noise?tisearch=e2e-sitesearch&keymatch=ths4531a%20popcorn#

0 kai klaas69 over 1 year ago in reply to Philipp Hack

Guru 140200 points

Hi Philipp,

don't forget to mount suited isolation resistors directly at each output of THS4531 to isolate the THS4531 from capacitive loads. Otherwise instability and increase of noise can occur.

Kai

0 Philipp Hack over 1 year ago in reply to kai klaas69

Prodigy 10 points

Hi Kai,

thank you for your response. I have also tried with 40Ohm resistors in series at the output to reduce capacitive loading as you mentioned. Unfortunately this did not help with the noise-floor issue.

+1 kai klaas69 over 1 year ago in reply to Philipp Hack

Guru 140200 points

Hi Philipp,

you should keep in mind, that the input voltage noise density of THS4531 appears to be multiplied by the noise gain = 2 at the output pf THS4531. And you will observe the thermal noise of 8k at the output.

Kai

0 Philipp Hack over 1 year ago in reply to kai klaas69

Prodigy 10 points

HI Kai,

you are absolutely right :), thank you. I was not aware that the signal and noise gain are different for inverting op-amp circuits. This would explain the missing 6-7dB, so I seems that my noise floor measurement at the output was correct and just my calculation of the input referred noise was missing this factor of 2.

Yes you mean thermal noise of 8k resistor if I use Rf=8k and Rg=2k to achieve 12dB gain ? I tried this and got -102.3dBV at the output. I also tried 470R and 2k (also approx 12dB gain) and got -103.5dBV at the output but the current consumption went up to 2mA. Is there anything else I could do to lower the output noise-floor by some dB ?

Best Regards,

Philipp

0 Michael Steffes over 1 year ago in reply to Philipp Hack

Guru 47465 points

You could reduce the 2k resistors to 1kohm, but

The lower voltage noise opitions are the somewhat higher supply current OPA1637 and THS4561. Both offer a shutdown feature you could possibly use to save system power when the channel is not active. Those are both on a different IC process than the THS4531 - so perhaps (but not for sure) could be more free of popcorn noise issues,

0 Philipp Hack over 1 year ago in reply to Michael Steffes

Prodigy 10 points

Yes thank you. I will look into these. Thanks for the great support !

Best Regards,
Philipp

Amplifiers

Amplifiers forum

THS4531DGKEVM: THS4531DGKEVM