I’m trying to build a simple circuit that acts as a "voltage breaker" – meaning that if a voltage exceeds a certain threshold, analogous to a (current) circuit breaker it will open a circuit. There are a couple of things that seem to make this a bit challenging because of its application – which is to serve as a speed limiter on a Lionel model train for my young son. Little boys have a habit of turning the voltage up too high and running a train off the rails.

This implies a few constraints:

  • The supply to be constrained is AC, operating in the 8 – 16V range. I’d like to limit the voltage (i.e., train speed) to about 13-14V in a way that can be adjusted to account for different engines’ operating characteristics.
  • The current draw can be up to 5A on a regular basis. Lionel used universal motors until very recently and they can draw quite a bit of current when pulling a consist of cars. Other accessories also add to the load as they draw power from the track.
  • The circuit needs to be able to withstand occasional (but not rare) full shorts that last a few seconds, as it is not uncommon to cause a short for a variety of reasons (such as derailments that connect the rails). The variable output transformers used (I have a 1940’s Type V and a 1950’s Type ZW, for those who know of such things…) can output tens of amps when shorted and take a few seconds before triggering their internal thermal breakers which are usual in the 5-15A range.

Balancing all of these constraints, what I’ve been able to come up with is the below circuit which:

  • Creates an AC reference voltage that I don’t want my son to exceed. This is my LimitAC voltage source at the bottom portion of the circuit. The do-not-exceed voltage would be set at the transformer using one of the multiple controls.
  • Rectifies the AC in both the reference and track (this is my TrackLevelAC2 voltage source at the top – more on the “2” in a second) voltages, which I then…
  • …feed into a suitable op-amp comparator (likely a power amp, to provide the necessary current to actuate the relay’s coil). My thought is that rectifying both sources will make the comparison straight-forward.
  • When the track voltage exceeds the reference voltage, have the op-amp feed a separate 12V DC supply that triggers a NC relay to open the track circuit (ActualTrackAC1 and ActualTrackAC2). I’m simulating the relay with just a voltage controlled switch at this point.

enter image description here

A few questions/suppositions:

  • I’ve had to set up two identical voltage sources for the track to get the full AC waveform instead of the rectified AC waveform. In the real world, if ActualTrackAC1 and ActualTrackAC2 were one and the same, wouldn’t the track and the rectifier-to-op-amp portions both be able to operate independently? Related to this… EDIT: simple reference question.
  • …LTSpice doesn’t power ActualLoad (which I’ve set at 4 ohms, a reasonable approximation for an engine) as I’d expect. In the real world I would be drawing a couple of amps from the transformer, not a few microamps. What am I missing here? Isn’t an LTSpice voltage component an “ideal” voltage source and thus capable of feeding whatever current is required into the track circuit? EDIT: simple .model error now corrected.
  • Do I really need to rectify the sources to get the opamp to function correctly? If I don’t, it would seem that the comparison will work as intended on the positive half-wave but fail in the negative half-wave (since the reference will be higher, i.e., less-negative, than the track).

This all seems overly complicated given the conceptual simplicity of the goal, but given that I’m working with higher amperage AC and need to be able to withstand short circuits I seem unable to use solutions that are DC based, for low amperage, or are not very happy with full shorts (e.g., a simple Zener clipper – N.B. I’d settle for a clipper-type circuit instead of a breaker...) This cannot be an uncommon need, but since I’m just a casual hobbyist – am I missing something? Is there any way to improve this circuit?

  • 2
    \$\begingroup\$ In the days I was a (really) little kid, the speed controllers of my model train had this really big knob with an arrow like pointer, indicating the voltage (they were actually variacs, scary, hm?) or speed. My grandfather bolted a screw as the "maximum speed setting delimiter" so the arrow won't get further. \$\endgroup\$
    – PlasmaHH
    May 25, 2016 at 14:50
  • 1
    \$\begingroup\$ I assume the speed controller has some sort of dial, knob, handle. Could you open it up and simply add (move) a mechanical stop? \$\endgroup\$
    – Tyler
    May 25, 2016 at 14:50
  • 1
    \$\begingroup\$ Oh btw. you see the microamps on your load because your switch/relay is still in the open position. \$\endgroup\$
    – PlasmaHH
    May 25, 2016 at 14:55
  • \$\begingroup\$ @PlasmaHH and Tyler Ah, yes - the mechanical solution (in my case, a couple of pieces of tape built up on the face) is the one I used for awhile (and my father before me - all big boys were once little boys...) but it doesn't allow for the variation between engines. Some barely move at 12V; others are flying. Alas, a ZW has vertical throttle controls unlike the V which had knobs - so the tape thing doesn't work. \$\endgroup\$
    – JimMSDN
    May 25, 2016 at 15:03
  • 1
    \$\begingroup\$ Triacs come in all shapes and sizes. A household "lamp dimmer" is a triac with a variable "on" period, so some (most) triacs can switch high AC voltage and high current. An issue is though, that a volt or so may be lost across a triac, which may be important at your lower operating voltage. A triac can be simulated in LTspice, but you may have to supply it a SPICE definition file. \$\endgroup\$
    – rdtsc
    May 26, 2016 at 12:41

1 Answer 1


I'd be tempted to go into the digital domain, but if you don't want to do that, what I'd suggest is to divide down the measured voltage with a voltage divider, rectify it, low pass filter it, and then compare it with a DC reference. An AC reference will be out of phase!

You can then change the voltage divider to set the target voltage, rather than having to set the reference. LEDs can be used as budget references since the whole thing doesn't need to be too accurate.

This has the added advantage that you can feed the measured voltage to a DC panel meter and draw a red line on it at the trip threshold.

Consider also having a latching relay that needs to be manually cleared, or is cleared when the voltage goes below a much lower threshold. That way it doesn't start up at full voltage.

  • \$\begingroup\$ +1 on the latching relay - I hadn't considered that once the track voltage is brought back under the limit, the relay would then close and thus current would immediately hit the engine at full (limited) voltage instead of having some way to bring it back up gradually from 0. When you say "digital", are you talking about using some sort of fit-to-purpose chip that does voltage comparisons? \$\endgroup\$
    – JimMSDN
    May 25, 2016 at 16:18
  • \$\begingroup\$ Also, when you mention the reference being out of phase: I was a bit concerned about that as well. Are you talking about it being out of phase in what I've designed --or-- that being a concern in what you proposed? Now, LTSpice seems to show the comparison in my design working as hoped for - but perhaps there are too many "ideal-ities" at this point to show where the problem might crop up. \$\endgroup\$
    – JimMSDN
    May 25, 2016 at 16:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.