Powering up with large capacitor in parallel

Hi, I have a board (Vbus is 160V) with few amplifiers on it that send short bursts to a load. The peak absorption is about 20A for 5ms (every 1s). Since I want to parallel some of them, I will have peaks of ~200A.
I want to perform some test by paralleling 100mF, but the problem is that I need to separately power the board, charge the capacitor and then close a switch that connects the capacitor in parallel.
I am pretty sure this is a common situation with large capacitor: do you know if there is some commercial chip that acts as a controller in such a situation? I would like this to be automatic.
Thank you

If it's for a test, why does the switch need to be automatic?

For such a high current I would recommend a capacitor right across the power for each amp.
Otherwise a 200A surge is going to require a very healthy wire connection.

Most controllers have a soft start function and this ought
to be somewhat adjustable (because users' bypass cap
bank value varies, as does the upstream supply's ability
to deliver current). You want C*dV/dt to be less than the
controller's current limit setpoint and below the upstream
source's limitations. You can pick timing component values
to make this so.

Having a switch connecting the capacitors to the DC bus is not the best way to deal with inrush. If your supply has a current limit function, then that should limit inrush by itself. Otherwise, you can put a current limiting resistor in series with the supply, then bypass the resistor with a switch (MOSFETs) after the bus is fully charged.

mtwieg said:

Having a switch connecting the capacitors to the DC bus is not the best way to deal with inrush. If your supply has a current limit function, then that should limit inrush by itself. Otherwise, you can put a current limiting resistor in series with the supply, then bypass the resistor with a switch (MOSFETs) after the bus is fully charged.

Click to expand...

That is the way I would also recommend.

The peak absorption is about 20A for 5ms (every 1s)

Click to expand...

So you are having some time constant: can you specify R and C separately?

Once you parallel the caps, you will have lower R and larger C and perhaps the time constant will not change much.

At constant voltage, you need not worry much. Best to use high power SCRs but MOSFETs are not that expensive any more...

c_mitra said:

Best to use high power SCRs but MOSFETs are not that expensive any more...

Click to expand...

SCRs shouldn't be used when the load is pulsed, since you'd have to re-trigger the SCR at the start of every load pulse.

mtwieg said:

Having a switch connecting the capacitors to the DC bus is not the best way to deal with inrush. If your supply has a current limit function, then that should limit inrush by itself. Otherwise, you can put a current limiting resistor in series with the supply, then bypass the resistor with a switch (MOSFETs) after the bus is fully charged.

Click to expand...

If I do so, then the voltage will raise slowly due to the fact that most of the current will go into the large caps. I already did this, and the consequence is that the power amps do not like it and start to oscillate

Maybe you need to "gate" the power amps until power
conditions are met, as part of the startup dance. This
is also a common function (UVLO - undervoltage
lockout) in power management but might be absent
from your radio (?) lineup - although phased array
element controller chips I've done, had this (both
Vdd and Vgg in-spec) requirements for the drain
switch functions.

Power supply supervisory chips are one approach, a
SSI analog comparator (w/ supplies suitably scaled,
and maybe some simple open-drain wired-AND logic)
could be another (what that logic would command done,
that suppresses PA action until OK, is for you to figure out)