I am in the early stages of designing an adjustable high-voltage DC/DC converter (similar to the EMCO C80 module) and one of the goals is to be able to closely monitor & respond to the output current (e.g. if we sense arcing in the application we want to reduce the output voltage). As a result, I am wondering if instead of using capacitors to filter/buffer the output I might be able to use inductors for this (help maintain stable current).
Is this feasible? Should I be concerned about the withstanding voltage? Despite being used in a high-voltage supply, there shouldn't be much voltage across the inductor. What are some things I should be keeping in mind?
Update:
This supply is used to generate a corona, which will result in some small current (~50uA?). If the current suddenly rises to a higher level that indicates a fault (i.e. arcing or short). My current design uses a full-wave Greinacher/Cockcroft–Walton multiplier. There is a capacitor at the HV output to reduce the ripple, but I am wondering if this possibly also reduces its ability to limit current (capacitor could supply a burst) or if I should care. Are there simpler ways to limit output current than monitoring it (e.g. via shunt + op-amp) and using that as feedback?
Best Answer
If you don't have an output capacitor, then you will see the switching step voltage at the output. So I would assume the output capacitor is needed.
Now let's see what happens when you add a simple series inductor to the output. When arcing occurs, there will be a sudden drop of load resistance. The intended purpose of the output inductor is to limit/retard the current increase. Consequently the output voltage drops significantly. So now you have a big voltage across the inductor and the voltage ramps up the current through the inductor. Then the arcing stop and the load resistance returns to normal. But the inductor has an excess of current, so the output voltage rises above the nominal voltage. With an increase of voltage and the remnant of an ionized path, it seems highly likely an arc would occur again. So once the circuit arcs, the circuit may become an oscillating arc generating machine.
I hope the following diagram can give you some ideas. Start with a current sense resistor (R sense) at V-/Gnd. The Control block can be simple or complex, for example you may use a high pass filter to detect arcing.
simulate this circuit – Schematic created using CircuitLab
Additional notes after your comment:
Actually, looking back at this, I want to modify some of the comments. The converter itself is unlikely to be able to hold up the voltage during an arc and it will take time to charge back up again, so an inductor combined with a diode circuit in parallel and reversed to clamp the flyback voltage and discharge the excess current can be useful for your purpose. Both the inductor and the diode would need to withstand the full output voltage.
In my diagram, the current limit block must be able to withstand full output voltage. At 8KV, I am not certain how to implement that. Perhaps a cascade of many FET can give a reliable solution, but it is not going to be simple.