LM5119: How to approximate Error Amplifier Compensation network values for lm5119?

JS777

Part Number: LM5119

Hi,

I'd like to know if there are some reasonably quick ways to approximate the RC values for the error compensation network in the lm5119? If there are existing posts that answer this question, feel free to post a link.

From reading through some app notes, it looks like I can approximate the crossover frequency to be around 1/10~1/5 of the switching frequency? And then after that, I believe I can do a few small calculations and get my RC values.

Oh, one other question: would I need to switch to a type 3 compensation network if my load resistance is too low?

I don't have any experience with DC-DC design or control theory, so maybe there are some important things I'm missing? If anyone can summarize the best approach to a reasonable approximation that doesn't require solving 2nd order differential equations.

Thanks!

over 6 years ago

0 Qian Chen over 6 years ago

TI__Mastermind 24075 points

Hello,

The attached spreadsheet can be used to select compensation values.

Thanks

QianLM5119 calcs.xlsx

0 JS777 over 6 years ago in reply to Qian Chen

Prodigy 70 points

Thank you! That's very helpful!

Will I need to use a type 3 compensation network? My output has very low DCR and ESR.

Does TI have any helpful spreadsheets to approximate values in a type 3 compensation network for the LM5119?

0 Qian Chen over 6 years ago in reply to JS777

TI__Mastermind 24075 points

Hello,

LM5119 is emualted peak current mode controller. I think you can start with type 2 compensation. Type 2 compensation should work with low DCR inductor and low ESR output caps.

Thanks
Qian

0 JS777 over 6 years ago in reply to Qian Chen

Prodigy 70 points

Thanks Qian!

The Type 2 compensation network did help smooth out my output current (vout = 30V, Iout = 30A for the brief periods when I turned the system on). Usually the first 100-300ms have higher noise on the output, and then the system seems to settle in to a smooth, low noise waveform. It's not consistent, which makes me think I'm on the edge of a properly compensated system.

I'll keep playing with the RC values in my Type 2 network. I did find a TI app note that goes over the different compensations networks, and they show a Type 3 compensation with an op-amp. I think I might be able to match their example schematic to the LM5119, approximate the zeros and poles and then calculate the R's and C's in that circuit. If that improves the output noise, I'll post results.

0 Qian Chen over 6 years ago in reply to JS777

TI__Mastermind 24075 points

Thanks for the update.

Power management

Power management forum

LM5119: How to approximate Error Amplifier Compensation network values for lm5119?