LM51231 Buck convertor convergence problems

Richard Martin

I designed a boost convertor using the LM51231. I used the spreadsheet tool to calculate the component values for 400W convertor, 12Vin, 50Vout.

I am trying to simulate a bust convertor using a LM51231, but I am having convergence problems with it. I have played with some of the simulator settings with no success.

Any suggestions would be greatly appreciated.

Rich

10 months ago

0 Bryan Radix 9 months ago

TI__Intellectual 2960 points

Hi Richard,

Thanks for reaching out. Could you please attach the filled out Quickstart Calculator?
Using this, we can start to pinpoint whether there's an issue with the circuit or with the simulation.

I would also have some other questions:

Are you using pSpice for TI or the regular pSpice?
Also: are you using the pSpice average or transient model?

Thank you and best regards,
Bryan

0 Richard Martin 9 months ago in reply to Bryan Radix

Prodigy 10 points

Hi Bryan,

Here is the filled out Quickstart Calculator. I am using pSpice for TI.

The circuit was built with 4 MOSFETs and is capable of handling 300W, but the MOSFETs fail when testing at 400W. So I am trying to analyze the circuit.

LM5123_Excel_Quickstart_Calculator_for_Boost_Controller_Design 400W.xlsx

Thanks

Rich

0 Richard Martin 9 months ago in reply to Richard Martin

Prodigy 10 points

Hi Bryan,

I should also add that I tried to simulate the switching of the MOSFETs using some variable power supplies and found that if the gate voltage of X1 is referenced to its source then the simulator has no problems. If it is referenced to ground then there are convergence problems with X3

Regards,

Rich

0 Haroon Wardak 9 months ago in reply to Richard Martin

TI__Genius 10385 points

Hi Richard,

The excel sheet you have attached seems to have some issues, did you change anything? it does not seem to be working correctly.

BR,

Haroon

0 Haroon Wardak 9 months ago in reply to Richard Martin

TI__Genius 10385 points

Hi Richard,

Could you also maybe for your MOSFETs use ideal models instead of NEXFETs and see if you have any luck with the convergence ?

Thank you

BR,

Haroon

0 Richard Martin 9 months ago in reply to Haroon Wardak

Prodigy 10 points

I use Libre Office, which might have changed some things.

Here is a new version edited with MS Office 2010.

Regards,

Rich

LM5123_Excel_Quickstart_Calculator_for_Boost_Controller_Design.xlsx

0 Richard Martin 9 months ago in reply to Haroon Wardak

Prodigy 10 points

Hi Haroon,

I tried using the MBreakN4 MOSFET model and it seems to have fixed the convergence issue.

Just waiting to see what the results of the sim will be now.

Thanks for the suggestion!

Regards,

Rich

0 Haroon Wardak 9 months ago in reply to Richard Martin

TI__Genius 10385 points

Hi Richard,

Do you want me to do a review of your design or is your question mainly on the convergence of Pspice for TI? Because for convergence I already suggested what you could try, if you need me to review your schematic, could I please have the full schematic?

BR,

Haroon

0 Richard Martin 9 months ago in reply to Haroon Wardak

Prodigy 10 points

Hi Haroon,

A full review would be great. Here is the schematic

Regards,

Rich

0 Haroon Wardak 9 months ago in reply to Richard Martin

TI__Genius 10385 points

Hi Richard,

I see, please allow me some time to get back to you on the review. I shall have something for you by end of next week.

Thank you for your understanding, my bandwidth is quite narrow at the moment.

BR,

Haroon

0 Haroon Wardak 9 months ago in reply to Haroon Wardak

TI__Genius 10385 points

Hi Richard,

Sorry for the delay, so here is what I have got:

- Everything from power stage looks good, just make sure your Inductor has high enough saturation current rating so that it does not get saturated, your peak inductor current will be around 44A, so make sure you have the right current ratings.

- your compensation circuit looks not so great, I would change the Rcomp to 25k ohms, Ccomp to 10nF and CHF to 150pF, for better loop response.

- The resistor on CSN pin, make sure it is 100 ohms and not shorted, I am not sure what is the purpose of that line above the R26 resistor, but it is important to have that resistor as 100 ohms there and CSP pin can have 0 ohm resistor there, you can follow the EVM schematics here.

Apart from that all looks good.

Hope this helps.

BR,

Haroon

Simulation, hardware & system design tools

Simulation, hardware & system design tools forum

LM51231 Buck convertor convergence problems