Circuit to protect current measurement shunt

[cameron98]

I was asked to implement a simple circuit to protect a current measurement shunt. The shunt has 30 watt and 300 joules capacity, roughly say about 25A.
We don't want it fried by very high current.

I am thinking using a spark gap and TVS across the current terminals, but I am not sure it is feasible for such low voltage drop across the shunt.
Anybody has similar experience?

 

[thylacine1975]

Depending upon the desired failure (/protection) mode and permissible voltage drop and circuit complexity, you could simply add a fuse...?

If more a elaborate approach is preferred, a MOSFET/relay could short (shunt the shunt upon overcurrent, or the current through the shunt could be inherently limited by a in-line current source/sink with a sink/source value = the max desired current flow.

 

[crutschow]

Spark gaps require a very high voltage before they conduct and TVS's also require a relatively high voltage. How about adding one or more large forward biased diodes in series across the shunt. A Si junction diode will conduct very little current below about 0.4V.

What's a "very high current"?

 

[cameron98]

Thanks for your reply
The protection circuit must be as simple as possible, preferably using passive components.
We recently have a shunt damaged due to about 2000A current passing through it.
We would like to limit the current at the level of less than 70A(the resistance of the shunt is 0.05 ohm).

A fuse in series is a good idea, I will discuss it with my colleagues.
For using a forward biased diode, at rated current 25A, the voltage across the shunt is 25A * 0.05ohm = 1.25V. So the diode will conduct at normal condition, this will bypass some current from the shunt which in turn lower the measurement accuracy.

 

[chuckey]

30 W @25 A ~ 1.2V max voltage. If you can handle the complexity, use an opamp as a comparator to compare the volt drop across the shunt with a suitable voltage, then use its output to fire a bistable which drives a relay that shorts the shunt out. If you do not use a bistable then the relay shorts out the shunt, the opamp sees zero volts then releases the relay. this then results in too many volts into the opamp, so the relay pulls in again. You have just invented a buzzer.
Another way is to put a series of diodes across the shunt, such as :- **broken link removed** , 320 mV drop at 40A. Or just put a thermostat on the shunt to trip the circuit if the shunt gets too hot.
Frank

 

[cameron98]

The fuse is not an option as we have quite high voltage involoved(about 25kV) unless we use a high voltage fuse cutout.

I will try a series of diodes parallel with the shunt.
What about a MOV? Is there any disadvantage of using it?

 

[Dan Mills]

Ok, at 2000A fault current you get (from I^2R a power dissipation in the shunt of 200KW, and will reach the shunts energy limit of 300J in 1.5ms.

How about a butch triac across the shunt with the gate tied to one of the main terminnals via a couple of diodes and a resistor? It should be easy to size things so that the triac fires when the shunt voltage exceeds a couple of volts shorting the shunt until the current falls to zero, and these parts have some fairly large surge ratings.

You will probably want a thermal sensor as well to catch smaller overloads lasting a long time rather then catastrrophic faults lasting a few cycles, this can tie into your telemetry systems to report overheating before (if it gets too bad) opening a shunt trip breaker somewhere upstream.

Regards, Dan.

 

[cameron98]

Dan

I like your idea. Using a triac might be the simplest solution I can imagine as the current through the shunt is AC.

If I want limit the current across the shunt to 100A, and I choose a triac having Igt=50mA, Vgt=1.3V.
The voltage across the shunt would be 100A*0.05ohm=5V. Then the voltage between the gate and main electrode will be 5V-1.3V-0.7V(should I use 0.3V ?) = 3V. The resistor to limit the gate trigger current will be 3V/50mA = 60ohm.
Do I miss anything?
And should I use 0.3V or 0.7V to represent the voltage drop on the gate PN joint?

Appreciate all you guys valuable opinions.

 

[FvM]

The protection circuit must be as simple as possible, preferably using passive components.
We recently have a shunt damaged due to about 2000A current passing through it.
We would like to limit the current at the level of less than 70A(the resistance of the shunt is 0.05 ohm).

Did you already consider a fuse?

Seriously speaking, I don't see a realistic chance to cut a 2000 A overload current of some duration with semiconductors. But the information given by the OP is incomplete, we neither know the expectable overload duration nor the driving voltage.

A staggered protection involving voltage limiting diodes and a mechanical short circuit switch or a fuse could be considred.

50 mOhm is an uncomfortably high shunt resistance for 25 A rating b.t.w.

P.S.: I overlooked the 25 kV specification. So voltage limiting diodes in combination with a short switch for longer overcurrent duration may be a way.

An overcurrent capable current transducer would be the state-of-the-art solution. Is it AC or DC current?

 

[KlausST]

Hi,

with AC a triac is OK..

A MOSFET with shorted G-D is also possble. select by GS threshold voltage. For AC you need two. 2000A is a lot...
What duration / joule of overcurrent do you expect?


Good luck

Klaus

 

[cameron98]

I don't want it go as high as 2000A, 100A maybe is the upper limit for me to let the protection kick in.
The shunt is 24A rating, and the I^2*t is about 6000. So that if I use 100A, the duration would be <=0.6 second, any triac or MOSFET would handle it without problem.

 

[KlausST]

Hi,

I don't want it go as high as 2000A, 100A maybe is the upper limit for me to let the protection kick in.
The shunt is 24A rating, and the I^2*t is about 6000. So that if I use 100A, the duration would be <=0.6 second, any triac or MOSFET would handle it without problem.

True for the shunt.

But if you use a triac, then the triac has to handle the 2000A and the I^2*t. ... as long as you don´t use a fuse...


Klaus

 

[D.A.(Tony)Stewart]

Fast protection of any device must factor the source voltage & impedance and possible load impedance and probability of failure to determine cost of protection. e.g. frequent stalled motor If you are wanting to protect against accidental shorts, Household breakers are rated for 10kA for this purpose. But if you are worried the breaker or fuse may not be fast enough, then you could implement a crowbar short circuit using overcurrent detection to trigger a Triac to bypass the voltage and trip the breaker faster.

 

[cameron98]

It seems to me that there is no easy solution except using a high voltage fuse in series with the shunt and a diode or triac across the shunt.
I like the idea that using a thermal sensor to trigger a switch to bypass the kilo amp level surge current. My question is "is there any thermal senor that is fast enough to close the bypass switch?".

 

[D.A.(Tony)Stewart]

Thermal response will depend on heat mass being sensed , sensor mass and thermal resistance to sesnor. If the thermal time constant of the substrate to the shunt is slower than the thermal time constant to the sensor, protection can be effective.Contact to hotspot with a thermistor , bonded with instant epoxy may do the trick. I have used this method to protect hots spots in SMPS with variable speed fan control, regulated by the thermistor on the hotspot.

 

[cameron98]

My proposal:
Using a 15A house hold circuit breaker in series with the shunt and a triac across the shunt.
The triac will be triggered to conduct if the current exceeds the overload threshold, for example 50A. If the current continues to rise to an even higher level, the breaker will trip(here I chose 300A, because at 300A the breaker will trip quickly in about 1 cycle and the triac can withstand 300A current for a short period without problem.)
The shunt is usually connected at ground side, so there is no concern about high voltage across the breaker contacts when it is tripped.
To me, this is the simplest solution for this particular application. Please help with finding any problems that you can foresee.

And also please help me select a triac that:
• handles 50A current when it is conducted;
• conducts when the gate voltage exceeds 2.5V;
• And the voltage drop, when it is conducted, should be as low as possible.

Any help is appreciated and thanks for all your time.

 

[thylacine1975]

Unless I've missed something... opening a breaker in a circuit energised at 25 kV will result in a potential difference of 25 kV across the breaker *irrespective* of where it is in the circuit. The fact that it's in the ground return leg is immaterial. If you're going to interrupt the flow of current, you'll need a switch capable of withstanding the full applied potential (which I don't think a household breaker will

(that's a thin attempt a humour, btw... Yup, sounded funnier in my head Just to be clear - no, the household breaker *won't*!

 

[Dan Mills]

Yep, that's why you shunt trip the proper serious breaker upstream (All serious HV breakers have a means available to open them under the control of a much lower voltage circuit (This is the 'shunt' trip)).

Please tell me this is a uni project and not something that is actually going to see the medium voltage network in a real application?

Regards, Dan.

 

[D.A.(Tony)Stewart]

The choice of shunt seems to be improper if it needs protection.

Normally 50-100mV max is all you need with proper signal conditioners, CM filters and shield.

this is regardless if it is 1A or 10kA.

Examining your specs 30W@25A indicates the shunt is >1V.

Using existing conductor straps may prove to make excellent Kelvin bridge shunts with suitable taps and above instrumentation details.

Otherwise people tend to use CT torroids for transformers. Is there some DC requirement we do not know about?

Is this not possible?

 

[cameron98]

Thanks for all your replies. It seems HV fuse and triac is the only solution.