TPS23754EVM-383: TPS23754EVM-383:

Hi Nigel,

Thanks for reaching out.

To find out the phase margin and gain marigin, please refer to this application note:

Reference: Switch-mode power converter compensatin made easy.

From the plots you have attached, it has very high phase at high frequencies, it could be the results of high amplitude of the injected signal at high frequency. 

And at low frequency, the spikes on the plot could be from the low signal-to-noise ratio, you might want to increase the integration time or check the amplitude of the signal at low frequencies as well.

Best regard,

Zhining

My Phase measurements were erroneous. Here is a new one. Dont worry about the bottom(settling time) and top end.(signal/noise)
Phase margin looks to be 110 degrees at crossover frequency of 5kHz and gain margin looks to be 20dB at 40k.
Does this sound reasonable?

Hi Nigel,

Since we have limited information about your system and measurement setup, it is difficult to have a conclusion from your results.

Please check the application note attached.

Reference: Switch-mode power converter compensatin made easy.

Best regards,

Zhining

Thats not a very satisfactory reply is it? I have sent you an accurate gain plot an an accurate second phase plot. My design is identical to the EVM. 
Do you agree with my analysis of those plots as to the phase and gain margins? Do you have gain and phase margin figures for TPS23754EVM-383 as a comparison?

Hi Nigel,

According to the plots you attached, the phase margin and gain margin of your board is as you mentioned.

We don't have bode plot for TPS23754EVM-383. If you would like to check for the bold plot of some TPS23754 reference design, you can refer to the test report of PMP9563. 

Link: PMP9563 reference design | TI.com

Best regards,

Zhining

Thanks, that confirms that my interpretation is correct. When you read up about Phase margin it talks about -180 degrees reference, but here we are using 0 degrees. Presumaby because there is an inversion at 0 degrees which I cant measure.

Hi Nigel, 

It depends on how the loop gain is defined. The negative sign in the loop will introduce a -180 degree difference. 

Best,

Zhining