TPS65261-1: Temperature affects loop stability

Hi, Harry

Have you ever tested PM at 25C and -10C, -40C? Could you send the results to me? 

The attachment is the loop curve and schematic diagram under different temperatures

3V3D空载不同温度下环路曲线及1V8D低温SW开关波形.docxTPS65261-1PMU参考电路图(1).pdf

Hi Donghai

Zhao is looking into this and will feedback to you soon. 

BR

Ruby

Hi, Donghai 

The Bode plot looks normal.  

Under low temperature, the output capacitance and compensation circuit (including R and C) will affect the Zero and Pole distribution which will change the Bode plot test results. 

HI，Ma

-20℃，phase is not good ;

-40℃，SW jitter is bad；

If we want to improve the Bode plot at low temperature，in the datasheet ，Rc，Cc，Cb，Co，C1（Feedforward capacitor），How to adjust；

C1 is 270PF, Rc is 8.2K， Is it reasonable？ if C1 is too large，what will happen？According to your experience，Check whether the compensation parameters in our schematic are reasonable？ thanks.

 3V3D drops at low temperature, causing the machine to reset all the time, 25℃ is normal

Hi, 

C1 is the feedforward capacitor, it will enlarge the loop bandwidth, make loop disturbed by noise. 

So 27pf is reasonable. 

Hi, 

Could you check Vin, ENx voltage, and LXx voltage when issue happening? 

We changed C1 from 270pf to 27pf and have solved the problem.  Is 270p too big? Will it affect loop stability?

Although the problem has been solved, we are worried that at low temperature, the loop will be unstable and there will be other hidden dangers.

We also want to adjust the compensation parameters to improve the loop stability, but we don't know how to start this work

I have reviewed the compensation circuit, it is good. 

Besides, 

TPS65261 and TPS65263 are similar. 

You can use TPS65263 webench to check its stability.  

https://www.ti.com/product/TPS65263?keyMatch=TPS65263&tisearch=search-everything&usecase=GPN 

First, we use the method of changing C1 from 270p to 27P At the same time, later, we will test the low-temperature Bode diagram corresponding to 27P and the Bode diagram at different temperatures with the official demon board to see whether the results are the same to verify that our parameters and PCB are OK