TPS565208EVM-858: Audible noise of the TPS65208EVM

Hi,

TPS565208 is FCC Mode, from principle, should be no audio noise issue. Would you please upload Vout ripple waveform for analysis.

Also, we will do some test in our lab after China new year holiday.

Hi,

thank you for your reply.

I wonder does TI have a Lab in Taiwan which we could visit with our device? I think this could speed up the debugging on both sides. I am currently in Taiwan and have the device with me.

Best Regards,
Markus

Hi Markus,

TPS565208 is a FCCM device, and the switching frequency should be constant at around 500kHz at full range of load so that no 16kHz noises at the normal operation. Could you please check the output ripple waveforms to check if any low frequency ripple components? 

Also, could you please share me you schematic? What is the output and what is the output capacitance at output terminals?

Thanks.

Bruce

Hi,

I can try to measure the ripple but the oscilloscope which I have access to is very old and has some junction issues as well (since I'm in Taiwan at the moment, not in our Lab in Europe and I have the device with me).

We are using the TPS565208EVM (jumpwired the board) and changed R1 to 54.9K, so it's TI's reference design.

I wonder do you have any FAE in Taiwan/Tainan (entire Taiwan would be okay though) which we could visit? If not I'll try to get access to better equipment.

We have also observed the same issue with another TI item on a RF EVB using the TPS54335A, I could bring all the boards. The noise is not so loud but can be heard clearly and it just should not be there.

Best Regards,

Markus

Hi Markus,

Also please measure the SW node, to check the switching frequency is constant (around 500kHz) and no 16kHz components. Sorry that I am in Chinese new year holiday and have no idea which FAE in Taiwan you can visit. 

According to the recommendation table, I also recommend changing the inductor to 3.3uH/4.7uH.

Thanks and Best Regards,

Bruce

We're still waiting for the inductor, I will update the thread once we have a result.

Ok so we have found 4.7uH and implemented it in the TI EVB - as expected it did not solve the problem. The audible noise is still the same.

Seems like TI has some fundamental issues with many switching regulators and/or evaluation boards because most boards which we have tested basically work properly (return a smooth voltage) but also generate audible noise within the distance of 10cm they can be heard clearly and if there's no noise around also at a further distance. We have spent quite some time to test several EVNs and they're all the same regarding the noise.

Inside a housing this might not be an issue but it's not an option for our case.

The noise seems to be caused by the MLCC capacitors - not the inductor, nor after it.

Hi Markus,

Did you have the chance to capture the SW and output waveforms? And do you mean that you can hear noise even in some of TPS565208EVM?

In Eco-mode devices like TPS54335A, it is normal and understood that the switching frequency will decreases at light load and when the frequency decreases to 20Hz to 20kHz, noises may be heard then.

But TPS565208 is a FCCM devices, the frequency should be constant at ~500kHz at full range load. No 20Hz to 20kHz frequency noises so that the noises cannot be heard. 

As I recommended, please capture SW and output ripple waveforms.

Thanks.

>  And do you mean that you can hear noise even in some of TPS565208EVM?

yes, the noise clearly comes from the TPS565208EVM.

Both Components TPS54335A and TPS565208 are creating louder noise the more the load is.
The maximum load is 4.6W, does this still fall under ECO for the TPS54335A?

I have just received another oscilloscope I'll get the plot tomorrow morning.