Electronic circuit breaker

Thread Starter

boatsman

Joined Jan 17, 2008
186
I have a circuit for a model railway where the output from the mains transformer is 12 - 0 - 12 v. In the case of something causing a direct short across the lines, I would like the circuit to be shut down immediately before damage is caused to any of the components. Some years ago I saw a simple electronic circuit that was included in a model railway diagram that immediately cut out the power supply when a short occured across the tracks, but I can't find the circuit in question. Can someone help me with this?
 

Thread Starter

boatsman

Joined Jan 17, 2008
186
Toy throttle basic.jpg There' s about 12VDC going to the tracks. I know I could put in a cartridge fuse in line with the 220VAC input. The controller will be used by children and short circuiting across the tracks will be inevitable. I plan to use the circuit shown below so I want to add a quick acting circuit that a child can easily reset by pressing a button. Green and red leds acting as indicators would be an asset.
 

Thread Starter

boatsman

Joined Jan 17, 2008
186
Thank you all for your replies. I am looking for an electronic circuit I can incorporate into the model railway controller circuit diagram I posted above. I once saw such a circuit but I can't remember where. Any information would be appreciated.
 

crutschow

Joined Mar 14, 2008
25,103
Here's a simple add-on for your controller which, rather than switch off the supply, limits the current to a maximum determined by R3 :-
View attachment 97424
That should work but note that the power transistors now have to dissipate the current limit current times the input voltage power, which would be much higher than that normally dissipated during train operation.

Perhaps SCRs or SCR connected transistors could be used to give a circuit breaker type action, totally shutting off the output.
 

marcf

Joined Dec 29, 2014
259
This is how I did it before the internet was invented. I'm sure this circuit can be polished to modern standards.
Nice circuit. Should there not be a limiting resistor between the collector of the NPN and the base of the PNP? Looks to me like the base - emitter voltage on the PNP will be a little high.
 

#12

Joined Nov 30, 2010
18,210
Should there not be a limiting resistor between the collector of the NPN and the base of the PNP?
You are correct.

Looks to me like the base - emitter voltage on the PNP will be a little high.
It's worse than that! You should add a 10V zener diode from gate to source of the mosfet to protect the gate. and a high current diode across the (maximum) current detecting resistor to avoid any negative voltage to get at the base of the NPN transistor. The voltage at the base of the PNP is Vcc, and so is the emitter voltage, until a fault happens. About 10K to 47K from Vcc to the base of the PNP is merely an, "off" command. The NPN transistor defeats the "off" command when a fault is detected.

As I remember, there is also an opportunity to add a capacitor as a positive feedback path..somewhere...in order to ensure a hard saturation of the logic transistors a soon as conduction begins. Then again, an op-amp or a comparator could be substituted for the transistor section. That is part of what I would call, "polishing this up to modern standards". An op-amp or comparator would simplify the circuit IF you have a proper power supply voltage available to run the chip. It would allow a smaller current detecting resistor and allow an adjustability feature to the current limit. I used an SCR symbol in the drawing because I didn't have a triac symbol handy in my drawing kit. This circuit also needs an RC filter before the base of the NPN so it doesn't respond to inductive glitches in the, "Unsafe Load". Probably about a millisecond would work.

Only the basic principle is demonstrated by my crude drawing. It goes like this:

A triac allows AC current to flow unmolested until too much current passes. The detector part knows something is wrong within one cycle of the power line. The detector circuit loads a capacitor on the gate of the mosfet to cause a time delay before the circuit allows the AC current to resume so the detector can check for too much current, again. Using 1 meg and 1 uf gives about 1 second before the circuit tries to re-set.
 
Last edited:

GopherT

Joined Nov 23, 2012
8,012
That should work but note that the power transistors now have to dissipate the current limit current times the input voltage power, which would be much higher than that normally dissipated during train operation.

Perhaps SCRs or SCR connected transistors could be used to give a circuit breaker type action, totally shutting off the output.

Aren't the SCR circuits a type of crowbar circuit that is intended to blow the fuse while protecting the electronics - essentially risking (sacrificing) the power supply as the SCR shorts across the power rails. I think the OP is trying to protect his power supply from exactly a short circuit across his power (railroad) rails.

If there is another option, I would be interested.
 

crutschow

Joined Mar 14, 2008
25,103
Aren't the SCR circuits a type of crowbar circuit that is intended to blow the fuse while protecting the electronics - essentially risking (sacrificing) the power supply as the SCR shorts across the power rails. I think the OP is trying to protect his power supply from exactly a short circuit across his power (railroad) rails.

If there is another option, I would be interested.
My purpose suggesting the use of an SCR was to provide the latch function with the output being used to shut off the power, not act as a crowbar to blow a fuse.
 

Thread Starter

boatsman

Joined Jan 17, 2008
186
@Alec_t
Thank you for the circuit. R3 is 1K? Does this addition cut out the effect of a short circuit across the tracks or just limit the amount of current so as to prevent the transformer burning out?
 

Alec_t

Joined Sep 17, 2013
11,388
R3 is 1 Ohm as shown, not 1k. R3 ~= 0.6/desired current limit in Amps. However, as crutschow pointed out, "the power transistors now have to dissipate the current limit current times the input voltage power, which would be much higher than that normally dissipated during train operation.". Unless the current limit is low (I've no idea what current the train draws) then those transistors inside the controller will probably overheat. A better circuit would cut the power completely on detection of over-current. I don't see any simple way to do that without using a mechanical relay. A solid-state relay/switch (e.g. SSR, triac) able to deal with dual-polarity currents generally drops a significant voltage, which would undesirably reduce the voltage getting to the tracks.
 

crutschow

Joined Mar 14, 2008
25,103
This isn't guaranteed because I just did a bunch of math, but it should clear things up a bit. Of course, you can leave out the auto-reset feature if you want to.
I's not clear to me where your circuit goes in relation to the OP's circuit since the current to the track is plus or minus DC. :confused:
 
Top