BLDC/Inverter is effectively an open load at the dead time intervals?

[Flyback]

Hello,
We are controlling a BLDC with an inverter (voltage source inverter) which is for a 550W water pump..for a tulip field irrigation system in Holland.
The input voltage to the system is 18-32V, and the motor voltage at our required speed (8000rpm) is 34V.
Therefore, we need a buckboost converter upstream of the BLDC/Inverter.
We are using the LT8705 buckboost controller.
The Inverter switches at 400Hz with 400us dead time.
We are using the inverter in "dumb" mode, whereby it simply switches at maximum duty cycle all the time......the motor current is thus not controlled by the inverter, but instead , will be controlled by the buckboost converter, which is upstream of the inverter.
The buckboost converter has a feedback loop which is closed around the motor speed, so that the buckboost converter can control the motor speed. A Frequency_to_voltage_converter converts the motor speed to a voltage so the buckboost can use it in its speed feedback loop.

Our problem is that we have very, very little room for output capacitance at the output of the buckboost converter. Thus , during the dead time intervals, the Vout of the buckboost starts skyrocketing, and is in danger of overvoltaging the 50V ceramic capacitors there….therefore, we have to temporarily stop the buckboost from supplying current during the dead time intervals…..and then resume the buckboost operation at the end of each dead time interval.
(the dead times come round every 2.5ms)

When we temporarily shut off the buckboost during the 400us dead times, we want the Vc pin (compensation pin) of the LT8705 buckboost controller to “freeze” at its voltage level.
Do you believe the best way to do this is to inject a voltage above 100mV at the source sense resistor pin (CSP pin) for the 400us duration of the dead time?…..we can do this by sensing the vout overvoltage which occurs during the commutation dead times and then applying the said voltage at the CSP pin….obviously the vout will fall again at the end of the commutation dead time interval as the motor/inverter will draw current from the output capacitor bank.

LT8705 datasheet
https://cds.linear.com/docs/en/datasheet/8705fb.pdf

 

[cowboybob]

...Do you believe the best way to do this is to inject a voltage above 100mV at the source sense resistor pin (CSP pin) for the 400us duration of the dead time?…..

I must say, LT makes some incredibly versatile stuff.

Looks like it would work. 100mV is at the max upper range for that pin, so a slight reduction might be in order.

Obviously, the CSP "pause" sig timimg is critical. I'd scope the output for spikes and adjust the timing accordingly.

<EDIT>Just out of curiosity, would a small (10-20 ohm, high wattage) resistor in series with your caps achieve your goal? Slowing down the charge time constant might keep the voltage at a non-destructive level.

 

[Flyback]

...but that's the idea, i.e.. to go above 100mV so the LT3705 stops switching during the dead time...meaning we wont overvoltage the caps during the dead times

 

[cowboybob]

...but that's the idea, i.e.. to go above 100mV so the LT3705 stops switching during the dead time...meaning we wont overvoltage the caps during the dead times

As I said, it should work.

I mis-read the limit on the CSP pin.