TAS5630 Output Noise

Tomasz, please tell me about the audio source to TAS5630.  Can you send a schematic?

Also, can you try AC grounding TAS5630 inputs to see if the noise starts in TAS5630?  Disconnect input caps to TAS5630 and connect them between TAS5630 inputs and ground.  (If these are electrolytic caps please make sure to connect +terminals of the caps to TAS5630 inputs and -terminals to ground.)

I think there should be no significant difference in noise between AD and BD modes.

Regards,

Steve.

There is not a problem with the TAS5630 signal source... My configuration is:

Input Stage -> Dig. Attenuator -> Buf -> SigmaDSP -> Buf -> TAS5630

The whole audio path is balanced.

Having my TAS5630 disconnected from the SigmaDSP with shorted input I can measure around 250µV-255µV RMS A-weighted output noise. The measurement is performed with NTi XL2 Audio Analyzer. TI specified the TAS5630 typical value as 270µV RMS. So my application is fully in the TI spec.

- having connected with SigmaDSP (DSP output muted) I measure up to 260µV
- having DSP unmuted but the attenuator set fully down I measure around 265µV
- having the whole system connected together with attenuator set at 0dB and the open XLR input I measure up to 290µV A-weighted noise, what shows that the rest of the system works OK.

My question is, could it be possible to decrease the specified TAS5630 residual noise anyhow ??? I guess the noise is coming from the TAS5630 input stage, so maybe decreasing the overall chip voltage gain could I decrease the output noise level ... but how to decrease the voltage gain ??? Is there any trick to decrease the voltage gain of TAS5630 ??? any additional external circuit ???

If the noise is really coming from the TAS5630 input stage it should be no difference in noise between AD and BD modes...

Thanks for your support!
Tomasz

Tomasz,

Since you are achieving the noise performance spec'd on our d/s, there is nothing you can do to improve it...

The gain is fixed; not adjustable.

One idea might be to use TAS5631, then your modulator would control your noise floor.

-d2

Don,

Thanks for your replay. The amplifier module design is almost released, therefore I can not change the topology now to use TAS5631 with additional modulator chip. BTW - I measured something really surprising I do not understand with TAS5631 (TAS5631PHD2EVM) - could you be so kind and take a look on my issue here "TAS5631PHD2EVM output noise issue":

http://e2e.ti.com/support/amplifiers/audio_amplifiers/f/6/p/105405/371437.aspx#371437

I concerning replacing my TAS5630 with TAS5613A - it looks the TAS5613A has significantly lower output noise: 185µV vs. 270µV from TAS5630 in BTL mode. The TAS5613A has also 3dB lower voltage gain. Is the output noise coming mainly from the TAS input stage?

I checked several different PurePath configurations looking for the one with lower output noise. Could you tell me which one of PurePath amplifiers is the best audio quality / output noise level one?

I have one additional question concerning TAS5613A. The TAS5613A PVDD (recommended operating conditions) is defined as 38V max. value, however the absolute maximum ratings for PVDD is 53V. In my current design the main supply switches the rails (!CLIP controlled) between 25V and 42V. Do I definitely need to decrease the 42V rail of my power supply implementing TAS5613A?

Tomasz

Tomasz, TAS5611A in the Oasis family (the family for TAS5630 and TAS5613A) has slightly lower noise, 168uV typical, with lower power output but otherwise performance similar to performance of TAS5613A and TAS5630.  Maybe it can deliver enough power for your tweeter.  Of course it is pin-pin compatible with your PCB.

However, you must limit PVDD, to 34Vdc for TAS5611A and 38Vdc for TAS5613A.  Maybe it is possible to add a series element for one or the other of these chips to drop PVDD enough?  I know this will depend on your tweeter power requirement.

Regards,

Steve.

Hi Steve,

Thanks for your replay. I checked TAS5613A and TAS5611A specs and I've chosen the TAS5613A - I like the low overal voltage gain of the amplifier. However the whole PurePathHD is damn noisy ... I really like the chips, but with residual noise over 100µV they have 'no go' to use in active speaker as direct tweeter driver.

Back to my amp I ordered directly some samples of TAS5613A and replaced the TAS5630. Today I performed some measurements and the TAS5613A noise performance looks way better than TAS5630 - I measured around 165-170µV A-weighted RMS voltage noise on the output of my amplifier after changing the chips. Before I had like 255-260µV. It is already great step ahead! The amplifier has to drive high grade active speaker, like a studio near field monitor, therefore the tweeter must be really quiet. To have the tweeter really quiet, using this amp, I still have to use a passive tweeter crossover with at least -3dB band attenuation. However, how I said, 170µV is a really nice step ahead, because now, using classic tweeter crossover with band attenuation of -6bB (what is like normal for tweeter) the audible tweeter noise is really acceptable even for high grade product.

Cheers,
Tomasz

Steve,
Could you tell me what is the root cause of the residual output noise  for the TAS5630/15/13/11 PurePathHD amplifier family? It looks for me like a analog input stage is casing the noise and it is overall amplifier gain depend...is it a technology limitation? It would be really perfect to have a class-D easy applicable chip amplifier with unclipped output power around 150W and residual noise of 50-70µV...

I consider to use the fully digital amplifier in my next similar project but I do not really understand the noise issue using TAS5518 modulator. Different, however quite similar, applications using TAS5518 have sometimes completely different noise performance - from fantastic (TAS5518-5152K8EVM) to poor TAS5518+TAS5631, but I will try to discuss a little bit more about it in my other thread.

Tomasz

Tomasz, this is good news.  Thank you for letting me know.  I am very pleased that you were able to work this out.

I believe the source of output noise is the input stage in both TAS5630 and TAS5613.  Input stage noise performance was chosen to provide the 100dB signal-noise performance of the devices, really quite good.  However, I think your application uses a tweeter with unusually high efficiency, and that makes the difference.

Best regards,

Steve.

Steve,

Yes, somehow the design made one step ahead. I still didn't shared this results with the end-customer, who was really unsatisfied with the noise performance. I hope having the improvement we can go one step ahead with the project - start building the first pre-series.

You right, the 100dB signal to noise performance is good. My application is quite special and this is why I would probably need something like 115 dB or 120dB SNR to ba fully satisfied.

Tomasz