How to solve the following practical problem that I currently have in my "food bike" design in a practical manner:

Power supply (AC):

  • Generator: 800W

Loads (AC, purely resistive, each controlled with its own thermostat):

  • water heater: 750W
  • syrup heater: 517W


  1. The available power may never be exceeded. If so, a breaker trips, which must be avoided.
  2. The syrup heater load demand must always be satisfied.
  3. The water heater load demand is allowed to be interrupted in order to satisfy the syrup heater load.

Edit 1: Additional rules to narrow the solution space:

  1. Human intervention is not allowed to achieve the desired result because the food bike operator is too busy serving his customers.

Edit 2:

  1. I prefer to keep both heaters unmodified so their certification remains valid. Therefore both heaters (including their thermostats) are considered as "black boxes" with only their power cords as the "interfaces" to an external circuit.

Additional system details: This is a real-world problem that I try to solve in order to design a food bike for myself in order to have my own small business, in contrast to food-trucks, and in the long term to be powered with PV-charged LFP batteries. For now, I use a noisy and stinky generator. Each thermostat switches its heater on to keep it at a given temperature. Assume, for the sake of conversation, that this happens randomly at intervals of roughly a minute or so. The water heater is expected to demand power predominantly after the food bike operator washes his hands because that heater serves the hand-washing station. The budget to solve this specific problem is in the order of $100. Weight is an issue, so a lightweight solution is desired. Fictitiously, an "XOR" power strip with an configurable socket prioritization (the syrup heater socket would get priority) would solve the problem, by that I mean a pair of AC power sockets that mutually exclusive provide power without temporal overlap that would trip the generator's breaker. You could also see this as a "power strip with load-shedding".

Project's geographical location: California, USA (needed for appropriate certifications / regulations).

  • 2
    \$\begingroup\$ Turn off the water heater whenever the syrup heater is on...? \$\endgroup\$
    – user253751
    Mar 21, 2018 at 1:25
  • \$\begingroup\$ Exactly. But how to do this in an automated manner? \$\endgroup\$ Mar 21, 2018 at 1:31
  • 1
    \$\begingroup\$ Might need some more details here. Are you making a PCB for this? Or are you doing some sort of point-to-point solution in an enclosure? Or is this just some theoretical discussion you wanted to have? What's the cost constraint? There's thousands of ways to do what you are trying to. It's almost definitely going to involve a relay. \$\endgroup\$
    – DSWG
    Mar 21, 2018 at 1:54
  • \$\begingroup\$ @JohanvanRavenhorst What's going to be turning them on and off? \$\endgroup\$
    – user253751
    Mar 21, 2018 at 2:05
  • \$\begingroup\$ See the edits in the question. \$\endgroup\$ Mar 21, 2018 at 5:25

4 Answers 4


You can do what you require with the addition of one relay with a changeover contact and a little rewiring.


simulate this circuit – Schematic created using CircuitLab

Figure 1. (a) Existing circuit. (b) Syrup priority relay modification.

How it works:

  • RLY1 is an AC relay. Coil voltage should match the generator voltage. Contact rating should have a minimum voltage rating to match the generator and current rating of at least that of the load.
  • The SYRUP and WATER thermostats are normally closed and will open when temperature is exceeded.
  • When the SYRUP thermostat closes the relay will energise and power the syrup heater. The water circuit will be disconnected.
  • When the SYRUP reaches temperature its thermostat will open, the relay will drop back to its shown position and the water circuit can work as normal.


simulate this circuit

Figure 2. Using a thermostat with a changeover contact makes the modification simple.

If you could replace the thermostat with one with a changeover contact then the modification becomes very simple.


simulate this circuit

Figure 3. (a) Existing circuit. (b) Syrup priority current sensing relay module.

This circuit does not require modification of the syrup heater. It is similar to the socket monitor in Jeroen3's answer. It requires the use of a current sensing relay with an AC coil rating to match the current drawn by the syrup heater.

The problem with this circuit is that if the water heater is on then there is a transient high-current load when the syrup heater first switches on. Both heaters will be on for a fraction of a second until RLY1 moves. This may trip out your generator.


simulate this circuit

Figure 4. A current limiting lamp to prevent overload on switchover.

This circuit uses a lamp to limit current while the relay energises. When it does relay contact 1b bypasses the lamp and applies full power to the heater. Meanwhile the water heater is already disconnected so current is limited.

This solution relies on finding a relay that will energise on the low current but survive with the high current running through the coil.

Fuses and circuit breakers have been omitted for clarity.

  • \$\begingroup\$ These circuits seem valid solutions to the problem to me, thanks. However, it seems that they require modification of the syrup heater. I may eventually choose to do this, but I prefer to keep the heaters unmodified so their certification remains valid. Is it also possible to design a circuit where both heaters (including their thermostats) are considered as "black boxes" so only their power cords "interface" with an external control circuit? Possibly similar to what Jeroen3 proposes below? - PS: nice circuit schematic design platform! \$\endgroup\$ Mar 22, 2018 at 4:00
  • \$\begingroup\$ See Figure 3 and the description. \$\endgroup\$
    – Transistor
    Mar 22, 2018 at 7:11
  • \$\begingroup\$ Figure 3b: This design comes with the risk that the generator breaker trips, which is not desired. Do simple (off-the-shelf) circuits exist that would completely prevent this? For example by using a small energy buffer to power the syrup heater during the small time in which the water heater is being shut off? \$\endgroup\$ Mar 22, 2018 at 20:53
  • \$\begingroup\$ Energy buffering is difficult with alternating current. For $100? I don't think so. \$\endgroup\$
    – Transistor
    Mar 22, 2018 at 21:23
  • \$\begingroup\$ See update Figure 4. \$\endgroup\$
    – Transistor
    Mar 22, 2018 at 21:34

Here in Europe we have special extension cords to have two big loads on one breaker. Advertised for washers and dryers.

Energy distributor switch, Energieverdeler, Energieverteiler.
When the white socket is asking current, the black one disconnects.


Link to dutch shop.

You can also assemble this yourself with a industrical current monitor relay and a power relay.
Or a simple circuit with a LEM module and a comparator with delay that controls a relay. But $100 is very a tigh budget if you do not want to bodge something. (safety!)

  • \$\begingroup\$ Such a "power strip" is exactly what I am looking for. However, I am located in the United States and I have not yet been able to find such a device certified for the US market. I am willing to consider a higher budget if needed. I could build something from components. In principle, I only have "access" to the 120Vac power cords of each heater, not to other wiring such as those of each thermostat because those are in the machine that I would like to keep unmodified. I can solder a simple circuit, given a diagram and the required components, but I do not know how to design such a circuit. \$\endgroup\$ Mar 22, 2018 at 1:53

A device called a "current sensing relay" solves this problem within the financial budget. For example this one from CR Magnetics: it contains a current transformer and a SPDT (== Form C) electromechanical relay. The typical power consumption is typically 80mA @ 120Vac, so 10W of heat needs to escape. Edit: In my experiments, the heat production was negligible. Enclosure in a plastic box without venting holes did not result in a noticeable temperature increase for at least the first 15 minutes. However, more testing is needed to confirm this.

Note that this solution still has a small temporal overlap of all loads in the order of (half) the period of the 60Hz (USA) sine wave. It work if the total load does not exceed the generator's "maximum starting Watts" as it is called in my generator's manual. I actually tried this out and it worked as hoped: No tripped breaker!

  • \$\begingroup\$ Why not have the hot water heated first and stored in a super-insulated tank? A good tank can keep the temperature for many hours - then consider how you use it... \$\endgroup\$
    – Solar Mike
    Jul 5, 2018 at 20:09
  • \$\begingroup\$ That was my first idea. However, in California, that solution is not allowed on a mobile food facility by state law. This is no joke, the County health Department (they enforcing the state law) is very strict. I explicitly asked them. Here is the code: CRFC (California Retail Food Code) 114325.(a) On a mobile food facility that only utilizes the water for handwashing purposes, a minimum one-half gallon-capacity water heater or an instantaneous water heater capable of heating water to a minimum of 100 °F, interconnected with a potable water supply, shall be provided [...]. \$\endgroup\$ Jul 6, 2018 at 21:01

Use a dual pole switch. When the switch is one way the water heater is on, when its the other way the syrup heater is on. When the switch is in the center position neither is on.

Digikey has two switches that appear to meet your requirement.

CW Industries GTS448E101AHR

250V/12A, $4.42 each


Bulgin C1720H

250V/20A, $5.14 each


  • \$\begingroup\$ This would solve the problem as initially formulated, thank you for the fast reply. However, the operator could forget to turn the switch back from the water heater to the syrup heater, so the syrup hardens, delaying service to the customer. So the switching should be automatic. I added an edit to the question to reflect this. \$\endgroup\$ Mar 21, 2018 at 5:22

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.