This question is the logical conclusion of answers I received to this question, this other question is similar but doesn't address much of what I will ask and really goes nowhere.

How can I switch a low power circuit using a relatively higher powered one? I'm looking for something like a relay that does the inverse of what a relay is normally used for.

For context, I want to use a float switch as a sensor input to a RaspberryPI that will activate an ICE powered pump when water reaches a certain level. All the float switches I can find are designed to work with voltage and current way in excess of what the GPIO pins on a RaspberryPi can handle, and I'm concerned that the 3v3@16mA that a single GPIO pin is capable of mustering is not sufficient wetting current for the available switches.

I can't actually find the wetting current listed on the data sheet for any of these switches, which I think is an indication that the manufacturers never intended them to be used under such low power conditions.

My other option is to use an alternative power source and run power through the float switch to the control of a relay that itself switches the low power circuit on the Pi, thus signalling that the water level is high. However that would require finding a relay with a 12v coil and a switch with very low wetting current, one with gold contacts potentially? though again I can't find anything suitable online.

I'm aware a capacitor + resistor combo placed across the switch can be used to momentarily boost the current applied to it, however when the system is required, the switch itself will be submerged in water, so that option poses it's own challenges. Also the distance between the Pi and the floats will be reasonably large (roughly 1.5m) and I'm not sure how that will affect it's modest power offering.

I'm planning to use multiple floats anyway, but the option of just accounting for the failure of one float in code is sub-optimal as I don't know what their "failure mode" will look like; will they all fail almost simultaneously? will I get much useful work out of them at all? will their failure be intermittent? etc.

The systems designed to activate a large-ish 1600 litre/minute commercial trash water pump while simultaneously waking up me and my dad, at which point we'll sit and stare at the thing as it manages the water level of it's own accord. Obviously I'm reticent to just risk it as if it fails, we won't wake up and it won't start the pump and, if the water level gets too high, my parent's house will probably flood (again). All input is appreciated.

  • 1
    \$\begingroup\$ low power circuit and high power circuit could both be using the same voltage \$\endgroup\$
    – jsotola
    Commented Oct 27, 2023 at 19:35

1 Answer 1


The max current and voltage of the switch are not relevant here.

The switch will probably work for a long time at that power. Wetting current is the minimum current to keep the contact in good health. That doesn't mean that it will fail quickly if at all. Also if your switch is normally open then it might be less of a problem.

You could also run two switches for redundancy and check them every now and then. The odds of both failing at the same time are very low.

Your relay solution could work. Just pop the cap + resistor circuit on it rather than the switch. Probably not the best solution.

You could use the switch + higher power to drive a transistor that in turn drives your input. Even better, use an opto isolator.

  • \$\begingroup\$ thanks, the switches are so cheap I'd happily even use three upper and three lower floats and write code that implements a little switch democracy. Your idea of using a capacitor + resistor across the contacts of a relay that is actuated using more power is actually good. It would mean the relay could be kept close to the Pi and the low power circuit could be kept short, while the longer wiring for the switches could use higher power. I have been looking into both transistors and opto-isolators so I'll consider those too! \$\endgroup\$ Commented Oct 28, 2023 at 12:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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