TPS552882: Unstable switching frequency

Hi Boris,

Thanks for reaching out on E2E. I have some questions need to check first, 

1. May I know what is the application? What is the input voltage range and max output current?

2. Is the waveforms captured at light load? Is Channel 1 the Vin waveform?

3. What about the waveforms at heavy load, do you have any waveforms?

4. It seems the frequency of Vout(CH4) is around 10kHz, and it is among the human audible frequency so we can hear the noise.

5. After reviewing the schematic, I found the inductor value is a bit small and it could cause unstable.

Regards,

Bryce

Hi Bryce,
1. This is a mobile device, and the power source comes from a PD adapter, we set the voltage to 9V which means that the input power is 9V, and the maximum output load is 5A. Since we have low power mode, we hope to set the TPS552882 to switch between different loads automatically. 

2. Yes, and channel 1 is Vin.

3.No

I changed the inductor to 2.2uH, the result is the same and there are some components of noise and frequency instability.
Is there any way to increase the whole switch frequency?

Waveform_2.2uH

Thanks.

Hi Boris,

1. Yes, this device has low power mode and the output voltage ripple frequency could drop in audible frequency which is not we want. Do you know the noise is from the inductor or capacitor? We normally recommend to change the inductor or capacitor material to improve the noise. And change to FPWM mode also help to remove the noise.

2. I think the input voltage ripple is a bit large, what is the input capacitance now?

3. We normally choose 4.7uH inductor at fsw=400kHz to obtain smaller inductor current ripple. The 2.2uH is still small.

Regards,

Bryce

Hi Bryce,

There are two things that need to be clarified and resolved, and I'm not sure if they are related.
1. Component noise.
These noises all come from capacitors and I'm trying to reduce the resistor value(Rfsw) which means increasing the switching frequency to avoid noise. do you think that is a good solution?

2. Unstable switching frequency.
As you can see from the waveform, the gray areas are all jittery and unstable, do you have any suggestions to improve that?
(CH1:Vin, CH2:SW1, CH3:SW2, CH4:Vout)

Waveform

Thanks,

Boris.

Hi Boris,

1. Yes, change the switching frequency would also help to make the voltage ripple out of the audible frequency which could help. You can have a try.

2. We observe the jitter in continuous conduction mode, since this is pulse frequency mode, that will of course causes a lot of jittery. You can make it run at heavy load and see the SW waveform, and please short the DITH pin if not used, the dithering function could cause jitter too.

Regards,

Bryce

HI Bryce,

Thank you for your support always. I think we can focus on the issue of audible noise first.

I change Rfsw value to 9.5kohm which means switching frequency change to 2Mhz but still heard some noise from capacitors, do you have another solution to change the duty cycle?please refer the below link for the waveform of light loading and heavy loading

Waveform_Light loading

Waveform_light loading_zoom in

Waveform_heavy loading

Best,

Boris.

Hi Boris,

1. Does the noise only happen in light load? Do you care about the efficiency at light load because this device can be set to FPWM and then the noise can be removed.

2. The waveform at heavy load is stable, it doesn't shows unstable issue.

3. You can have a try with the following suggestions: (1) change the capacitor value; (2) change the capacitor material (like MLCC with other material, aluminum electrolytic capacitors)and change to smaller package(0805, 0603...), also some capacitor vendors have their anti-noise capacitor; (3) add a dummy load at the output side if efficiency is not so cared.

Regards,

Bryce

Hi Bryce,
1. Yes, the noise only occurs at light load, we have low power mode so hope to use PFM mode to achieve better efficiency at light load.
2. You are right, this is one of our solutions but it may increase the cost so we hope to change the frequency to solve this issue.

Best,

Boris.

Hi Boris,

Understand, you can try the suggestions which I posted in last message, and try with NP0, C0G MLCC capacitors.

Regards,

Bryce

HI Bryce,
understood.
I have another question is the maximum output current is 5A, which means the inductor average current limit could be set to 10A. Does the current setting of our circuit need to be changed to 33k ohm?(original is 60.4K ohm)

Best,
Boris.

Hi Boris,

The ILIM pin resistor R2911 is used to set the average inductor current limit and the output sense resistor R2904 is used to set the output current limit, these two are different functions. Why do you need to set inductor current limit to 10A?

Regards,

Bryce