Hello,

I am looking for some help with an over voltage protection circuit without using a crowbar clamping circuit or using anything too elaborate. I have attached a picture of the schematic for anyone who is willing to help.

I have a current mirror circuit where I am setting the reference current as 100μA. The DUT in this circuit is D1 and it must have 100μA through it. The device I am trying to protect is Q1, where VCE max is 24V. The transistor part number is CA3046 (specially picked for my application).

The nominal clamping voltage of the DUT (D1) is 390V. I am trying to protect the transistor (Q1) in case the DUT is somehow shorted. If it is shorted the high voltage will then appear across Q1.

I have only shown 1 DUT in my picture, but in my actual application there will be 25 DUTs using these transistors to mirror my reference current.

Any help is greatly appreciated.

If I did something that goes against this forum please forgive me as I am a new member (I did read the rules and regulations though).

Thanks again.

 

Place a 15-20V zener across the transistor (Q1), collector (zener cathode) to emitter. You will also need a resistor in series with the voltage source (V1) to limit the short circuit current.

 

Hello Crutschow ,

Thanks for your response.

I've tried placing a zener across the transistor with a resistor in series with the power supply. The problem is the current through the DUT is then governed by the value of resistor I place in series with the power supply instead of mirroring the reference current.

I've attached a picture of the simulation I tried. Any other suggestions you have is appreciated.

Thanks again for your help.

 

Well, it should be obvious that your power supply voltage must be less then the voltage drop across the DUT plus the zener voltage. If it's more than that then the current would, of course, be determined by the resistor, which appears to be the case in your simulation. You do understand Kirchhoff's Voltage Law(?).

Edit: Why did you pick a transistor with such a low voltage breakdown when you are working with such a high voltage? For a high accuracy constant current you could use a high voltage MOSFET controlled by an op amp in a constant-current configuration, similar to this.

 

If it is shorted the high voltage will then appear across Q1.

Yes, but if the current through Q1 is restricted to 100uA then the power dissipated in Q1 when it breaks down is only ~42mW. Is that likely to damage Q1?

Edit:
According to this article any damage would be due to heat generated rather than the breakdown mechanism itself.
The tricky bit would be maintaining the 100uA current through Q1 under breakdown conditions.

 

Yes, but if the current through Q1 is restricted to 100uA then the power dissipated in Q1 when it breaks down is only ~42mW. Is that likely to damage Q1?..........................

If Q1 breaks down then what would restrict the current?

 

Yes, but if the current through Q1 is restricted to 100uA then the power dissipated in Q1 when it breaks down is only ~42mW. Is that likely to damage Q1?

Edit:
According to this article any damage would be due to heat generated rather than the breakdown mechanism itself.

Contrary to your referenced article, even in a single transistor, I would worry that the parameters of the transistor would suffer from the breakdown. For instance breaking down the base-emitter junction makes the transistor permanently noisier.

I worry a lot more with the monolithic device (CA3046) thaat is being used since there are connections through the silicon to other transistors.

 

If Q1 breaks down then what would restrict the current?

Nothing much at present; which is why I said "if the current is restricted..." and "The tricky bit would be maintaining the 100uA current through Q1 under breakdown conditions."

 

For a 390V DUT and a 15V zener across the transistor, the maximum supply voltage should be no more than about 400V so that the zener doesn't conduct under normal operation. Then to minimize problems with a shorted device you could add a current-limit in series with the power supply. For 25 units to be tested simultaneously at 100μA, the current-limit should be slightly more than 2.5mA.

Below is a simulation of such a limiter. It uses a high voltage MOSFET which it turned off by Q1 when the current through R1 generates a base-emitter voltage at Q1 of about 0.6V. The load circuit is simulated by a resistor in parallel with the protection zener, D2.

 

Well, it should be obvious that your power supply voltage must be less then the voltage drop across the DUT plus the zener voltage. If it's more than that then the current would, of course, be determined by the resistor, which appears to be the case in your simulation. You do understand Kirchhoff's Voltage Law(?).

Edit: Why did you pick a transistor with such a low voltage breakdown when you are working with such a high voltage? For a high accuracy constant current you could use a high voltage MOSFET controlled by an op amp in a constant-current configuration, similar to this.

Thanks for the response crutschow. This was the original circuit I used to implement my current source at first and it worked well. I am currently sticking with this for now, but I was trying to see if there was a way a current mirror could do the same thing just as an experiment.

For a 390V DUT and a 15V zener across the transistor, the maximum supply voltage should be no more than about 400V so that the zener doesn't conduct under normal operation. Then to minimize problems with a shorted device you could add a current-limit in series with the power supply. For 25 units to be tested simultaneously at 100μA, the current-limit should be slightly more than 2.5mA.

Below is a simulation of such a limiter. It uses a high voltage MOSFET which it turned off by Q1 when the current through R1 generates a base-emitter voltage at Q1 of about 0.6V. The load circuit is simulated by a resistor in parallel with the protection zener, D2.

This circuit is good and I have a few different options available to me to protect the transistor in similar configurations.

I was wondering if there was a way to to keep the original current mirror circuit and somehow monitor the zener anode voltage of the DUT while still supplying a relatively accurate 100μA load current. Kind of like how a TVS is inactive for a circuit until it detects a high voltage (in this case a short circuit).

Thanks for all of the responses everyone. It has been very helpful!

 

....................................
I was wondering if there was a way to to keep the original current mirror circuit and somehow monitor the zener anode voltage of the DUT while still supplying a relatively accurate 100μA load current. Kind of like how a TVS is inactive for a circuit until it detects a high voltage (in this case a short circuit).

The current limit circuit I posted will protect your current-mirror circuit with the addition of the 15V zener across the current-mirror transistor. The 2.8mA current-limit current will flow through the zener in the event of a short circuit, without damaging anything.