TPS53915: Remote sensing to compensate for DC IR drop

Hi Herman,

The TPS53915 is not a full differential sense device.  The FB signal is with respect to the IC local ground which is merged inside of the device.  You can refer to the TPS549B22 device which support up to 25A with PMBUS and true differential sensing. However, it's not pin to pin compatible.

Thanks.

Amnat

Hi Amnat,

The TI article I sent you shows that this can be done with a single ended FB signal. We are at late stages of our design and changing to a different regulator is not an option. Can you take a closer look at the article and tell me if a 1uF bypass cap is sufficient? Also what is the circuit theory behind this choice, is it just the Thevenin equivalent R (R1||R2 of the FB resistor divider)*Cbypass?

Thanks,
Herman

Hi Herman,
I understand you now. Please, let me explore further on this. I will let you know ASAP.
Thank you.
Amnat

Hi Herman,
Would you please describe where you will place this bypass capacitor for the TPS53915 (between which pins)?
Thanks,
Amnat

Hi Amnat,

Please read the app note from TI which I sent you the link. This bypass cap will go between VFB and VO, I did an experiment to move the VFB resistive divider right under our ASIC and place a 0.1uF near the regulator. The output ripple was about the same but I compensated the DC IR drop, I just want to know the analysis to choose the bypass cap value as this may have impact on ripple and feedback loop stability.

Thanks,

Herman

Hi Herman,

Thank you for the clarification.  What is your output voltage?  Is there a resistor divider to the VFB?  You can also send me your schematic if you like.

Here are the 2 scenarios.

• With feedback resistor divider to VOUT.  Placing the "bypass" cap between VO and FB make this cap become feed forward cap.
• No feedback resistor divider to VOUT. Placing the "bypass" cap between VO and FB should not have an affect on the system since this cap is now shorted.  Based on the author's input, this bypass cap provides a low impedance path at the switching frequency than the sense resistor Rsns (apps note), which is ~10 to 50-ohm range.  1uF bypass cap has 0.53-ohm at 300kHz, which is much less than 10-ohm.

Thank you.

Amnat

Hi Amnat,

There is a resistor network as our target output voltage is around 0.9V. The resistor divider is 5.36kohm and 10kohm, Rsns is 5.36kohm // 10kohm (based on a Thevenin equivalent circuit) and yes the "bypass" cap is a feed forward cap with significant contribution to the circuit. For this scenario, what is the math behind choosing the optimal value for this cap?

Thanks,

Herman