TPS56637: TPS56637 Audible noise

Hi, 

What is the mode setup? What is your Vin, Vout, L and loading? If possible, could you measure the SW frequency of your application condition?

If it works under pulse skip mode, the switching frequency has the possibility to fall within audible noise range.

Thanks,

Lishuang

Hi,

Eco mode is used, input voltage 24V, output voltage 4V(3.3uH) and 3.3V(2.2uH).

Currently difficult to measure the frequency.

And What is pulse skip mode?

Hi,

Pulse skip mode is another wording of Eco mode.

Eco mode is a DCM control mode for the converter. That means, if the load current is smaller than 0.5* inductor current ripple, then the switching frequency will be reduced to avoid the negative inductor current.

Eco mode will bring higher efficiency at light load, but it will also reduce the switching frequency to achieve that target. Then the audible noise will occur at light load.

You could make the mode pin floating and IC will work in FCCM mode. You can test if there will be audible noise then.

Thanks,

Andrew

It is difficult to float because the mode pin is connected to GND.
I hope I can tell you another way.

Hi,

What is the current loading? Can we add dummy loading for output? If so, we can increase the frequency and avoid the frequency to fall within audible noise range.

Thanks,

Lishuang

Currently, 2.17A per ic.

What is dummy loading?

How can I increase the frequency?

Hi,

Just increasing inductance value could let your IC enter CCM mode. But I don't understand why DCM mode will occur in your application.

DCM mode only occurs when the 0.5*peak to peak inductor current ripple>load current.

As you have said, input voltage is 24V, output voltage 4V(3.3uH) and 3.3V(2.2uH). It can be calculated that for the 4V Vout, the peak to peak inductor current value will be Vout/Vin*(Vin-Vout)/Lo/fsw=4/24*(24-4)/3.3/0.5=2.02A. Thus the half inductor current ripple will be 1.01A, and it is smaller than 2.17A loading.

For the 3.3V Vout, the peak to peak inductor current ripple will be 3.3/24*(24-3.3)/2.2/0.5=2.5875. And half of the inductor current ripple will be 1.29375A. That's also smaller than 2.17A loading.

So could you use our online webench tool to do a simulation in steady state with your BOM to see if there will be DCM mode?

Or maybe the loading in your application is variable and the DCM mode just occurs when the light load is applied? I think we need more information to help you figure out the issue.

Thanks,

Andrew

Hi,

First, let's check the results using the webench tool.
And we will increase the inductor value and check the audible noise.

Agree with you.