WEBENCH and SwitcherPro result validity

K. Sasaki14

Other Parts Discussed in Thread: TPS54620

About DCDC simulation tool,

Do you verify the validity of simulation result compared to actual board measurement result?
If you are doing it, please let me know the method and comparison point.

Best Regards,
Kohei Sasaki

over 9 years ago

0 Nikhil Gupta (Product Folder Models) over 9 years ago

TI__Mastermind 39210 points

Sasaki-san,

We do verify against bench measurement. Please let me know what aspects are of interest to you?

Thanks,

0 K. Sasaki14 over 9 years ago in reply to Nikhil Gupta (Product Folder Models)

Guru 10200 points

I posted this question because the actual Bode plot has large gap (phase margin was different about 30 degree) with SwitcherPro.
The device is TPS54620. Is the simulation model different between WEBENCH and Switcher Pro?
Then what kind of measurement item are done in bench measurement? Please let me know the typical item. And what kind of circuit are measured actually?

Best Regards,
Kohei Sasaki

0 Amod Vaze over 9 years ago in reply to K. Sasaki14

TI__Mastermind 26800 points

I recreated the EVM design on both WEBENCH and SwitcherPro with the same L, Cout( linearly derated to around 22.38uF from the original 47uF in the BOM), and the compensation components (Rcomp/Ccomp/Ccomp2 or R3/C6/C7). I got close enough values for crossover frequency and phase margin as in the measurements which are around 45kHz for crossover frequency and 50degrees for phase margin. There could be a slightly different value of Cout derating on the actual board leading to some differences. We usually calibrate the loop model with atleast the typical EVM board at different Vin's and Iout's and try to keep accuracy as close as possible to the bench measurements.

Regards,

Amod

0 K. Sasaki14 over 9 years ago in reply to Amod Vaze

Guru 10200 points

Amod-san,

Thank you for your support.
I understood about default EVM condition, however I measured 1Vout setting by changing EVM components value.
I will attach the measurement result , SwitcherPro result and WEBENCH result.
Please refer to the attached file.

Is the loop model calibration done at different Vout condition?

TPS54620 Bode Plot.pptx

Best Regards,
Kohei Sasaki

0 Amod Vaze over 9 years ago in reply to K. Sasaki14

TI__Mastermind 26800 points

Kohei-san,

Thanks for the report on the measured v/s model discrepancy. That helped with debugging the problem - it turns out that when porting over from SwitcherPro to WEBENCH, the power stage parameters of this device are not exactly same in the tools. This is a bug and needs a fix. The TPS54620 model in SwitcherPro seems to produce better results with respect to measured data even as you start deviating from the EVM conditions.

We will work on fixing this issue with the loop model in WEBENCH and will let you know once done. Please let us know if you have any more questions.

Regards,

Amod

0 K. Sasaki14 over 9 years ago in reply to Amod Vaze

Guru 10200 points

Amod-san,

Thank you for your answer. When can you fix about loop model?

Best Regards,
Kohei Sasaki

0 Amod Vaze over 9 years ago in reply to K. Sasaki14

TI__Mastermind 26800 points

Kohei-san,

I am aiming for the fix to be out in about 2 weeks or earlier. I will keep you updated.

Regards,

Amod

0 K. Sasaki14 over 9 years ago in reply to Amod Vaze

Guru 10200 points

Amod-san,

Thank you for your reply. I'm looking forward to your update.

Best Regards,
Kohei Sasaki

0 K. Sasaki14 over 9 years ago in reply to K. Sasaki14

Guru 10200 points

Amod-san,

Do you have any update? Has the loop model fix been done?

Best Regards,
Kohei Sasaki

0 Amod Vaze over 9 years ago in reply to K. Sasaki14

TI__Mastermind 26800 points

Kohei-san,

Sorry for the delayed response. The discrepancy has been fixed already. Please let me know if you have any questions regarding this.

Regards,
Amod

Simulation, hardware & system design tools

Simulation, hardware & system design tools forum

WEBENCH and SwitcherPro result validity