Solution Finder Get Tech Support

Low Frequency Ripple at High Load

Posted by: T_A on

From attached shematic, I changed R1 to 40K in order to escape burst mode at lower load and get lower ripple. However, at low input voltage (120Vdc to 230Vdc or so) and at higher load (400mA,) I observe low frequency oscillations in the output ripple - at around 5KHz. This adds about 15mV in ripple (on top of regular 23mV or so), but it also looks like an unstable response.

I added R8-C11 phase boost to improve stability (this was not there in original deign), but this didn't help. No oscillations at this frequency observed on primary, secondary, or bias windings. No oscillations observed at optocoupler primary side anode either.

Any idea what might be the cause of these oscillations or how to fix it?

Thanks.

Files

Attachment Size
new_schem.jpg 104.27 KB

Comments

Submitted by PI-csbabu on 11/19/2020

Hi ,

Can you please share the waveforms at different time scales to differentiate between switching ripple and oscillation.

Submitted by T_A on 11/20/2020

Hello,

Thanks for reply. Yes - see attached. I think it is fairly easy to differentiate, especially when 40K current limit resistor is used.

When I use 25K current limit resistor, lower-frequency oscillations seem to be somewhat better, but "real" ripple gets higher. Additionally, with lower current limit resistor, I need more load to escape the burst-mode.

I am worried that I have an unstable system and lowering my current limit resistor only "masks" the real issue.

Submitted by PI-csbabu on 11/20/2020

Hi,

First the device you are using for a given output power not recommended. If you look at the X-pin resistor vs peak current limit graph (Fig. 37) and considering the tolerances, the device is very much under utilized. Please look into Tiny switch, if you have opportunity to change the design.
Present design, the output capacitor (C9) value is high, reduce it to half, so the increased ESR may help for stability. Next try adjusting the compensation, replace C12 with R-C combination. Start with 100nF, 4.7kohm combination and do some iterations.

Submitted by T_A on 11/20/2020

Yes - I am aware that TinySwitch is more suitable but unfortunately it cannot be changed at this point.

Already tried replacing C12 by RC combination. I will try some more iterations. In which direction should I change RC values? (Increase both, increase C only, decrease R, etc..)

I didn't have much luck with changing output capacitance. Ripple would get higher if I reduce it further without much effect on the stability.

Submitted by PI-csbabu on 11/23/2020

Hi,

Do you have any test equipment to measure the loop bandwidth. If you dont have, please verify the bandwidth through step load from 10% to 90% load. Based on the results, we can adjust the loop parameters.