mmWave — power consumption?

I really love the idea of the mmWave switches but I am pretty concerned about standby (“vampire”) power. This is a thing especially if you install many switches. Electricity is super expensive here and I am literally saving every watt…

What will be the power consumption of a mmWave Blue Zigbee switch?

Please no “it will be low” or “too low to care” answers.

Understand actual consumption is not known yet because product doesn’t exist yet but what is the target/spec?

This is just my educated guess, based off of using and measuring some different mmwave sensors and the Inovelli 2-1 switches: between 2 and 4 watts per switch. My multimeters are not super accurate at the low end though.

I’m curious about the data you’ve measured behind your estimates. Doing a bit of searching shows that many people have measured their Blue dimmers at around 0.5-0.6watts standby and that mmwave specs tend to also be <0.5watts average draw, though can draw more power at times, apparently (?). I don’t know what Alibaba specs are really trustworthy or the exact details of the mmwave inovelli plans to use, of course.

Anyhow, those specs above would seem to hint at a blue series dimmer with mmwave tech added coming in at ~1watt average standby power draw, which feels more reasonable, but I may be off base here in my own assumptions.

1 Like

Again, my multimeters are not super accurate at the low end so my numbers could be off.

On the red 2-1 (Z-Wave) , I was getting around 1.5W at idle. I have several ESP32 wifi based mmwave sensors and I measured the USB draw through my USB meter for them at 1.25W. I have no idea how accurate the USB-C meter is nor how much is used by the ESP32 chip and how much by the mmwave radar, but I guessed that it could be as low as half.

So 1.5W + 0.75W at the low end (approx 2W) and 1.5W +1.25W at the high end (close to 3W). I threw in an extra watt at the high end just in case my measurements were under counting.


Yes, power measurement is tricky. Often people believe they can just measure the current with a clamp and multiply by 120. But that’s not right, both current and voltage have to be multiplied instantaneously and then integrated. Can only be done with a specific device or a scope.

Power factor of such devices tends to be poor… not surprised if VA is multiple times larger than W.

Anyway, 1W+ for such a device is huge. Mind a typical house can have 30 of them… that’s 30W 24x7. 362kWh/yr. 120$/year (!!) at my electricity costs.

There’s gotta be an official spec and /or target and I’m hoping @EricM_Inovelli @Eric_Inovelli can chime in.

I think 30W is 262 kWh/yr or about $86/yr, using your numbers.

Ops typo in the kWh but price is right:

30/1000 * 24 * 365 * 0.45=120$

1 Like

You’re paying .45 a kWh??? Yikes

Welcome to California. Yes, every watt counts (especially if up 24x7)

I measured the Red with a scope the best I could and got about 1/4W of power when off.

You can’t measure these with a multimeter. It requires a scope with power calculation capabilities. I don’t have a great way to measure the very low currents so even with the scope my read accuracy suffers.

Sorry bad writing on my part, I was referring to using the multimeter for measuring the DC watts used by the mmwave sensors I was testing.

For the switch itself, I measured it through my Emporia Vue panel mounted individual circuit power monitor. It’s calibrated against each of the power phases and the circuit I was getting the wattage from only has 3 red switches on it. I did my testing when all switches were off and got 4-5 W.

The Emporia itself is likely not that accurate compared to a decent scope though. It’s also possible that the power usage is likely different depending on how you have your light bar setup.

I wouldn’t trust that system to get the power metering right either.

I actually have the Vue 2 as well, flashed with ESPHOME.
Yes, It’s crazy inaccurate, especially at lower power levels.

I’ve actually purchased a load of 1% 5W resistors and plan to calibrate my CTs but didn’t get the chance to do this yet.
Ok curious already how the transfer function really looks like. For now I can say it’s extremely noise when no load (or light load) is connected.