TPS65400 on PYNQ-Z1 board

[lowuze]

Hello!

I bought PINQ-Z1 and I liked your solution, to use as a power supply chip TPS65400. I decided to use this chip in my solution. Could you explain how you calculated the values of the elements of the compensation scheme for TPS65400 (C167, C168, C172, C173, C179, C180, C184, C185, R252, R255, R262, R265).

Thank you!

[jpeyron]

Hi @lowuze,

Welcome to the forums. This question is beyond my experience and knowledge. Hopefully one of our more experience forum members will be able to help you with your question. 

cheers,

Jon

[gcsipkes]

Hi @lowuze, a little late with an answer but if you still haven't found your solution you might want to look into the TPS65400 datasheet offered by TI. You can find it here

http://www.ti.com/lit/ds/symlink/tps65400.pdf

Beginning with page 23 you'll find the small signal equivalent model of the regulator and the equations used to determine component values in the feedback and compensation networks.

 

[lowuze]

Hi gcsipkes!

The problem is that if you calculate the values of the elements according to the formulas from the datasheet (with electrical parameters of PYNQ-Z1 board), these values will not match the values of the elements that are installed on the board.

Thank you!

[gcsipkes]

Hi,

The values obtained by using the datasheet formulas depend on a few variable parameters you are considering for your design. Some of these must be set/programmed before building the supply. The switching frequency, the internal reference voltage, the lowest load resistance (highest load current) and the filtering capacitor ESR will all influence the calculations directly.

The final values may also suffer adjustments during validation, depending on the measured real phase margin and dynamic performance.

Regards,   

Gabor

[lowuze]

Hi gcsipkes!

Yes, i know that, but, for example, If we calculate the compensation circuit for second channel: if using type II, then R255 = (2pi * Fc * Vout * C177)/(Gm * Gmps * VREF), We know, that Fsw about 1MHz, because R257 = 150K, then Fc = 10^5. I know, that VREF = 0.8 V (I read the data from the TPS65400 using I2C). Ok, then R255 must be about 48K (for type III - about 12K). I understand that the value of the R257 = 100K is taken with a margin. And i understand that the values of C173 and C176 most likely was defined from bench test. But i don`t understand reasons for using value of C172 such 150pF. The value of this capacitor usually also taken with a margin (as in the compensation circuit for first channel, where the calculated value for C167 about 250pF), but for this capacitor C172 = Rload * C177 / R257, Rload = 3.3V / 1.6A, that about 2Ohms, then C172 = 450pF. I understand that only the designer of this scheme will be able to answer this question. Anyway, thanks for your reply!