I'm trying to solve this issue for my upcoming circuit, but haven't come to any robust solutions.
The circuit now is simply:
MCU (on 9V battery) switching a MOSFET (with a load at variable 40V-350V)
I want to put an optoisolator to isolate both circuits, but the problem is...how would i power the HV side (the transistor of the optoisolator), considering that i the HV source varies quite rapidly between 40-350V. Normally the mosfet gate shouldn't be above 20V and neither the optoisolator's collector-emmiter.
A simple divider probably won't work, since dividing 350V in the same ratio as 40V would yield barely 2V, which is below the mosfet gate treshold.
Further i have to keep the mechanism as simple and reliable as possible.
I'm thinking now...would clamping the mosfet gate line with a 5-15V or so zener solve the issue. I'd still need the resistor divider to drive down the HV to a manageable voltage fot the optoisolator and keept the currents down as to not burn the zener.
I'm thinking now...would clamping the mosfet gate line with a 5-15V or so zener solve the issue. I'd still need the resistor divider to drive down the HV to a manageable voltage fot the optoisolator and keept the currents down as to not burn the zener.
A Zener sounds like a good plan to me.
I haven't drawn-out the schematic, I'm just imagining this in my head.... Since the opto and the MOSFET gate are connected together, the Zener should protect both. And, you should only need one pull-up resistor which will determine/limit the current.
You do not have to power the MOSFET from the load side. There should be a dedicated power supply which provides the necessary gate voltage. Remember the important figure is the voltage between gate and source and it really does not matter what voltage is connected to the drain,so long as it is within the device limits.
That power supply may be derived from the high voltage but it is a dedicated voltage regulator circuit (possibly using a Zener) that needs to be designed and tested.
If your microcontroller is battery powered, why can it not float to the ground potential of the output voltage? It may not need any isolation at all.
I haven't drawn-out the schematic, I'm just imagining this in my head.... Since the opto and the MOSFET gate are connected together, the Zener should protect both. And, you should only need one pull-up resistor which will determine/limit the current.
Yes, something along these lines. Super simple to put together.
MorganS:
You do not have to power the MOSFET from the load side. There should be a dedicated power supply which provides the necessary gate voltage. Remember the important figure is the voltage between gate and source and it really does not matter what voltage is connected to the drain,so long as it is within the device limits.
That power supply may be derived from the high voltage but it is a dedicated voltage regulator circuit (possibly using a Zener) that needs to be designed and tested.
If your microcontroller is battery powered, why can it not float to the ground potential of the output voltage? It may not need any isolation at all.
It would be best, but i'm afraid it might not be an option.
Right now it does float yes, because it's on batteries. However being a N-Channel mosfet it has common ground with source. Eventually i'd need to connect it to the PC. IMO this might expose it to earth ground trough the PSU. If the mosfet would be ON or shorted i'd be putting the PC at risk if not also the circuit. Having a +50W source at 350V attached to my PC is not a pleasant thought.
Having two separate batteries is kinda impractical.
Seems that some kind of voltage regulator on the load side would be best considering the situation. At least as far as i can tell. I'm guessing linear regulators are out of the question?
350V is serious voltage. That's the kind of voltage that can jump over small tracks in a PCB and zap anything nearby. Designing a PCB for that voltage is possible but it's not something I would like to tackle on my own.
Is 350V the absolute worst-case inductive spike voltage or are we looking at equipment that has this nominal voltage, with the possibility for higher spikes? You need to design the input protection for the HV supply with the worst-case spikes. Remember it's not difficult for big rotating machinery to create spikes of twice the input voltage.
Good idea to keep it isolated from the PC. So why not put the optoisolators on that side of your Arduino? You need to charge the battery off USB?
Precisely so, yes. My current prototype circuit has absolutely no isolation implemented, but it's running on batteries so i don't have to worry about any HV burning the components. I connect it to the PC only when the HV is turned off. It's just a temporary solution.
I'm pretty confident (sadly have no oscilloscope at hand), that the 350V is peak voltage, so RMS is lower. You guessed correctly, a machine is the HV source. However i have a diode protecting the load side, haven't had any trouble in this regard.
I've just designed a simple voltage regulator with a very low current on the load side. Calculations show it should work.
However i'm wondering about making a spark gap, but then again the diode should take care of it. I've read 5 mills goes up to 1.5kV.
Look at how opto isolators are laid out on PCBs. There is an obvious separation between the two sides. Many even route slots in the PCB so that there is no way that dirt on the board can help the HV break the isolation.
1KV can jump about 1cm in air. If there is anything even slightly conductive then this distance can be greater.
1KV can jump about 1cm in air. If there is anything even slightly conductive then this distance can be greater.
Normally
air medium is widely use as an insulating medium in different electrical power equipments
and over head lines as its breakdown strength is 30kV/cm.
Interestingly enough, TO220 casings with the usual 2.54mm grid are sold up to 800V.
I'm not sure what the relevance of this is. The rating of a HV switching device is based on it's electrical characteristics and the distance between terminals (due to the arcing distance factor) .
To220 heatsinks, on the other hand are rated for the fact that they are made for TO220 devices, for which the pin spacing distance which is a fixed constant due to PCB hole dimensions for TO220 devices. We know from the 30kv/cm spec already given that the dialectric breakdown voltage for a distance of 2.54 mm is 0.254 * 30kV = 7.62kV, so the 800V spec must be due to the distance from the heatsink to any of the 3 pins of a TO220 device, rather than the distance between the pins , irregardless of whether a mica insulator is used. At work we sometimes use 1/4" Kapton tape wrapped around the pins of the TO220 device between the PCB (or perfboard) and the TO220 case, before mounting the heatsink. This prevents any chance of arcing from the heatsink to any of the pins. We have also used 1/8" flexible dielectric tubing cut to the right length and slid over the TO220 device pins prior to installing device in the board.
TO220 pin spacing is 40 mils which corresponds to a spark gap voltage of 1600V , giving a 2:1 ratio safety margin for an 800V rated part. It is very possible this is the recommended margin of safety and the 800V ratibg is indeed based on the pin spacing after all
In any case, a SSR is a good solution for your application.
