TPS4810-Q1: SCP for 3-phase motor

Hi Stas, 

So TPS4810 does have low side sensing as an option. Including low side of the load, this seems like what you are describing being low side of the DC/DC converter? 

For PWM what is the frequency you are expecting? Essentially we expect that the INP pin is PWMd in order to turn FET voltage on/off between 0-28V. Whenever the FET is turned OFF due to INP going low, we have added fixed delay of max 60us (across temp) before the FET turns ON again (INP going high).This intentional delay ensures safe turn OFF of external FET so duty cycle will be limited for high frequency. As you go higher in frequency, max duty cycle will come down.

For Ex: if frequency is 3kHz (333us) then max achievable duty cycle is ~82% (333us-60us)/333us across temp.

This would be the PWM limitation. 

Thanks, 

Sarah

Hi!

Thanks for fast reply.

PWM expected 20kHz, but tps4810 is not generating this PWM, it is meant only for SCP on phase. 

Question 1: If I enable switch (EN,INP high) will it pass PWM 0-28V with one back MOSFET? With back to back mosfets? VS provided separately 28V.

Usually I disconnected high side DC with such solution  I am afraid PWM may cause some problem.

Question 2 : I want to use 3xcontrollers for each phase. SCP on one will trigger disconnect for others. I think to implement  INP all conneted together as pull-up with all FLT open drain putputs. So FLT on any tps4810 will pull down all INP. Will it work in some configuration? Main goal for SCP event to disconnect all phases +-10uS. Maybe some other tps48xx will fit better? 

Question 3 : Because tps4810 works as only scp, maybe i should go with vds sense? I see it is not precise, but maybe it will be less noisy due to much less distance of sensing? 

Stas

Hi Stas, 

Where exactly are you applying the PWM? 

I understand that you are providing the controller with 28V.

If the expectation is that you will have oscillations coming from the output 1 FET will pass this through. B2B FETs will block this with the external body diode. 

If you plan to use the fLT output as the INP input on the other controllers this could work if you are ok with the delay. 

I dont expect the high side sensing to be more noisy. You can design with an additional capacitor between ISCP and CS- if this is a concern. The VDS sensing is less accurate and the sense pin traces still must reach to the FET (not much trace length savings). 

Thanks, 

Sarah

Thanks for the help.

Let me try to clearify general task.

I have driver that produces PWM 28V 3 phase.

Now it directly connected to the motor.

I need to put SCP between this driver and the motor.

I see solution to use high side switch as "eFuse" but my concern is that high side switch is meant to be used to disconnect high side DC voltage, not PWM signal. From specification I don't see a problem to allow 0V throw high side switch, but I am incertain how it will work with PWM and jumping boost capacitor.

Hi Stas, 

Let me check with our systems engineer to be sure. 

Thanks, 

Sarah

Hi Stas, 

I checked internally and it looks like we think this could be possible using some ORING but would like it evaluated using the EVM. Would you be willing to tr this using an EVM? 

You can ZS order them if so. Otherwise I can try this myself but will not get to it until end of next week. 

Thanks, 

Sarah