Hi Everyone: Been a while since my last post.

I have a simple question regarding charging a capacitor.

Basically I have a high voltage line in which I want to connect to a capacitor to charge it. Once the capacitor is charged, the high voltage line needs to be 'disconnected' from the capacitor. The capacitor then discharges into a load, and the process repeats.

My first idea was using a high voltage power mosfet. This becomes a problem in and of itself, because the mosfet requires a special gate driver to get the correct voltages. While I can get a driver that 'floats' with the high voltage switching line, resistive losses may become a problem when the device is switch on. Not to mention high voltage mosfets seem to have very high RDS on values.

My next idea is based on diode and thyristor theory. I have just become acquainted with thyristors as I never seemed to have to use them. They seem to be high power devices, perfect for this application; however, they have the nasty side effect of not being able to be turned off from the control gate signal. Basically, if you have a diode in series with a capacitor, current flows through the diode until the capacitor is charged. Then the current flow stops (if the cap doesn't have anything to discharge into). If you put a load on the cap, the cap will discharge into the load and the diode will allow the cap to be recharged.

Anyhow.. my idea is this. Instead of using a diode, I use a thyristor. The high voltage line comes from a switching power supply (say 400 volts), and if possible, I would like it to remain a DC voltage. When the thyristor gate is triggered, current flows through the thyristor to charge the cap. Once the cap is charged, the thyristor turns off, because there technically would be no current. In this way, the thyristor would do exactly what I wanted as per the problem statement. Now.. I know there is a voltage drop across the thyristor when it conducts, like a diode, therefore, I do not know if this method would work. A schematic of my idea is attached. Keep in mind, that I am just showing the components in the circuit. Any component model numbers shown are not what I would use in the end.

Basically, what happens is that the cap is charged via the thyristor when the control signal is applied. The cap charges all the way, and the thyristor shuts off disconnecting the HV supply from the system. Finally, the switch element (most likely a mosfet) discharges the cap into a load. The switch element would obviously be timed with the thyristor control signal to make this happen correctly.

So the question is: is this how it will work in reality, or will the thyristor not work as planned? Any ideas?

Thanks,
LostTime77

 

Google synchronous rectification.

 

Thank you for suggesting this to me.

Most of what I see here for schematics and what not use relatively low voltage mosfets to do the rectification. My application requires at least 400 volts. Granted, I seem to have found some mosfets that will do 250volts at 50mOhm RDS on, which is pretty good. A lot of these schematics have a mosfet switching into a transformer, where my system would not do this. I would like to switch at fairly fast speeds as the capacitor is a low value. Switching into a high inductance transformer would be an issue, not withstanding that this is not my intention either.

Also, keep in mind, unless my understanding of mosfets is flawed, which is quite possible since I do not have a ton of experience with HV, the mosfet would require a special gate driving circuit for HV. I would not be switching a source grounded N channel FET, which would be as easy driver since the gate voltage would be above ground (~5V logic). I would be switching high voltage into the drain of a mosfet. An N or P channel would require a floating gate voltage.

The entire idea with the SCR is that I would provide a short gate pulse of say 5V. Since the cathode is floating (not 400V at this point), the device would turn on. The 5V signal would then be removed. The SCR would shut itself off when the cap charged all the way (my hope at least). There would be no special gate driver required.

I will continue to look into synchronous rectification.

 

The SCR should work for charging the cap as long as you can wait long enough for the SCR current to drop to below it's holding current.

The SCR gate will have a high reverse voltage on it after the capacitor is charged. It can be protected by a high voltage diode in series with the SCR gate.

You could perhaps also use an SCR to discharge the cap into the load. For that you may need an isolated gate driver, such as a trigger transformer.

 

So basically, the SCR will stop conducting, after turned on, when the cap charging current goes below the holding current. Then I must be careful not to put a voltage on the gate while the cap is charged and the SCR is not conducting. This is because the device is reverse biased. Then as you said, if I put a diode in series in the gate, the diode will not conduct unless the gate to cathode is forward biased.

Thanks, that answers my question pretty clearly.