12V Crowbar Circuit

[wilheldp]

I have a 12V circuit that I want to protect from overvoltage and overcurrent surges.  I found a 12V crowbar circuit diagram online that should be fairly reliable for the overvoltage protection.  The circuit diagram is attached, and I have a question about it.  The relay coil is connected to a positive 12V supply, then the negative is run through a transistor.  Is there any reason why that 12V supply couldn't be attached to the incoming positive rail from the battery?  I assume that the external 12V is provided in order to protect the relay coil from damage if there is a voltage spike, but since this a protective circuit, I would rather the relay be damaged than the attached load.  Where would you get a clean 12V supply if the rest of the circuit shuts down due to overvoltage?

I think it's possible to use an inductor for the overcurrent protection, but I really can't wrap my head around how it would work. 

[wilheldp]

Below is an illustration of the modification I'm talking about...

[KM4FER]

Take your two circuit examples and think about what would happen if the input voltage happened to jump to say 50V or 100V or something outrageous.  What would happen in each case?  How would the result be different?

You say you would prefer the relay to be damaged rather than the circuit connected to the output.  In your proposed modified circuit what would be the failure modes of the relay in the event of a large voltage increase?  How about the switching transistor, how might it fail?  In each case would the input still be connected to the output by the relay contacts?

How sensitive is your attached load to >12V.  Assuming the input jumped to 100V how long would the circuit take to disconnect the output load and is this time short enough to not destroy your load device?

Answers to these questions should answer your original question.

[Wilksey]

That is not a crowbar circuit it is merely a switch.

[Dave]

The crowbar circuit got its name because it it works similar to putting an actual crowbar between the output terminals. It provides a very low impedance path between the output terminals, so the voltage gets clamped to a low value, which can not do damage to the DUT. Sometimes the circuit sacrifices itself to protect the rest of the power supply and DUT (gets destroyed when absorbing the energy of the fault).

A silicon controlled rectifier (SCR a.k.a. thyristor) is normally used for this job, because once triggered, it doesn't stop conducting until the current through it drops to zero (read: until you remove the fault, which caused the voltage to exceed the limit in the first place).

Here are two proper crowbar circuits:



They perform the same function, but the bottom one gives you the ability to set a much more precise limit, at which the circuit gets triggered.

[codeboy2k]

That is not a crowbar circuit it is merely a switch.

+1.  that's not a crowbar, it's just an overvoltage switch, and depending on the load might be easier with a MOSFET.

I'm of the opinion that SCR crowbar circuits, while once useful in their day, are not really necessary anymore since we have readily available MOSFET technologies that can disconnect the load when there's an overvoltage, i.e. load switches. However, that said, MOSFETS can also fail shorted and thus won't disconnect, so the crowbar might still be the best protection if there is a concern that the MOSFET might fail short, essentially leaving you with no protection.

In that case, I'd use the crowbar paired with a fuse as in Dave's second image.  Fuses are great disconnects, they never fail if sized right.

[wilheldp]

Thanks for the replies, guys.  Like I said, I grabbed my first circuit diagram from another forum, and they called it a crowbar circuit.  I like that design since the relay forms a physical separation of the power supply from the load.

The load in this case will be a fairly beefy computer (not your typical car entertainment computer) with several monitors and peripherals.  The total load will be between 50 and 100 amps on a beefed up truck power system (120A sustained, 140A peak).  The goal is to provide some surge suppression for the total load, then have a distribution fuse box downstream of this suppression circuit, and a UPS feeding the computer.  Any suggestions on what the best over voltage and over current circuitry would be? 

[Wilksey]

The circuit in itself will work as a cut off switch,
The original diagram was used probably because the device being powered was not 12V, but anything above 12V would essentially flip the switch.

Over current, a fuse, a slow blow, perhaps, depends on configuration.
You will also want to make a note of inrush current (the "instant" current taken at system switch on) as if for example a lot of caps etc are being charged, it can take a hefty slap of power from the battery, sometimes fuses will blow too quickly, so you have to make sure you get the right one (slow blow).

Most of the automotive gear will work past 12V due to 14+V from the alternator.

In your circuit essentially the zener (D2) is the trigger for the transistor.

EDIT: Actually thinking about it, if your alternator is charging you battery at say, 13.8V your circuit would probably cut off depending on component thresholds, you would be better using a 14V zener or something, or if your equipment is really that tight on tolerances (automotive generally isn't) then a regulator.

[wilheldp]

Yeah, I was thinking that if I used the relay circuit, I'd design it to trip at 15 or 16 volts.  I don't think anything automotive would be damaged at that level, but anything higher usually indicates that something is wrong with the power supply.

Would a large inductor slow the inrush current from large caps?

[Wilksey]

An inductor is used to oppose sudden current change, amongst other things, if you want to charge your caps up slower you could use a resistor, but have you measured the inrush, it might not be an issue in any case?

[richard.cs]

The relay circuit will not protect you from overvoltage during the time it tsakes the relay to open. Whether that matters depends what you're protecting and from what, many electronic things, potentially including the PSU for your computer can be killed in much less than the few milliseconds it will take your relay to open.

[Wilksey]

Yes, every "switch" has a delay, and the time it takes for the relay coil to react might be too late.