As the old saying goes “even a blind man can see the difference when my front loading washing machine is on” as it reverses directions and drops voltage over and over.
Nothing else looks anything like it.
So why can’t Sense “see it” 
Because your washer produces a mess of different on and off transitions that all have different characteristics. Betting you couldn’t spot your washer’s “signature” if it was mixed in with AC and other waveforms.
Maybe you can circle the on and off transitions for just the washer, then ? I’m still trying to piece together which are from the washer and which are from other devices. Is the little sawtooth literally dozens of on and off cycles of your washer ? I can’t see them in enough detail but they don’t look like clear on and off transitions - maybe you can zoom on a couple of them where you baseline is the lowest so we can get some sense of time and power transition scale for just 1-2 ?
Try to zoom in closer on the saw teeth. If each up or down transition lasts more than a second, Sense’s main detection technique isn’t likely to work. And if you want to see the transitions that Sense’s main detection mechanism can see, watch the Power Meter in real time while whatever you want to detect is running. Sense will “tag” the transitions of interest with either the transition delta power, or with the detected name.
1 Like
To add to @kevin1 's great feedback in this thread, I suggest you take a look at the interview we did with Data Science a few months ago. We spoke about some of the challenges we are working on improving with some details and some great visual examples. I hope it helps to (partially) demystify the device detection process.
I have a HS110 on my front loader Miele washing machine. Here is a power profile over a typical load of laundry, though note it can look different for different program cycles.
This cycle started off with an 8 minute period of ~1000W when it was boosting the temperature of the water. During this 58 minute load, the drum motor would spin slowly back and forth every 20 seconds to agitate, where the drum motor would use 35W when spinning. This wash cycle had 4 spin cycles where the drum motor used 400-650W. The last spin cycle is the longest, and you can see the power usage step up to 5 different levels as the RPM is increased every 1.5 minutes until the max RPM which is sustained for 6 minutes. I imagine the drain pump activates during the spin cycle, though it is difficult to see its power signature as it is likely constantly on during the spin and it has much lower power. The water fill solenoids are likely too small in power consumption to be seen from this profile.
So while I agree that perhaps Sense ought to be able to recognize the agitation phase of a wash cycle, there is a lot else that is going on [and which consumes a lot more electricity] in the washing machine.
3 Likes
This is another important point - one washer or dryer can run a variety of different modes that impact when and how much electricity is used throughout different parts of the cycle, changing the signature significantly.
1 Like
Unfortunately, my dryer, like so many, is electric 240v and Sense still doesn’t have 240v monitoring devices for those of us who also have solar or need to monitor a handful of such circuits. This is BADLY needed for lots of our higher power draw devices that are beyond Sense’s signature detection capability.
Kasa serves this nitche for 120v devices quite nicely, and makes Sense far more useful that it otherwise would be.
@andy, I recently put together a 240v power monitoring device that integrates with Sense. I plan to do a write up on it soon. It uses ESPSense loaded onto a ESP32 controlling a Circuit Setup 6 channel energy monitor. I had to install Home Assistant on my PC first (within a virtual machine I had to install beforehand) to config it, but HA nor the PC are required after getting it setup and running.
None of the steps to get this setup, configured, and working were too difficult, though there were a lot of steps involved. (I have one more step I need to do…move the device out of my main electrical panel and into a plastic box nearby…the wifi signal strength is at -85dBm otherwise which is a bit too low to reliable communicate on my network.)
Fortunately I don’t need this device for my clothes washer or dryer…they both use 120V and I have HS110’s on both already…but I am using this to monitor each of my 240V devices that Sense can not reliably detect:
- Car charging: never detected
- Oven: 50% detection rate
- Pool pump: never detected my main variable speed pump, never detected the salt water chlorinator, 50% detection of pool sweeper booster pump, never detected the 50W always on (my next project to hunt down the source and hopefully reduce it) for some device(s) power off the pool subpanel.
- AC1: 90% detection for my main single stage 5 ton AC compressor
- AC2: 50% detection for my single stage 2.5 ton AC compressor
- AC3: 50% detection for my other single stage 2.5 ton AC compressor
(I hope that >95% of the electricity these devices use is true 240V current (i.e. <5% 120V circuits running off of one of the hot legs to neutral), as I’m only monitoring the current with a CT clamp on one of the 2 hot legs of the split phase circuits; have it configured to double that current…so any 120V device I may be missing its power draw altogether, or I may doubling its power usage. The device is also only measuring the voltage for one of the hot legs to neutral; I assume the voltage between each hot leg and neutral is within 1% of each other, so that only 1 voltage reference is needed.)
BTW, there are other easier to use devices that are compatible with ESPSense, that you could use to monitor your 240V washing machine, e.g. a shelly 2.5. I have not tried using them, so I won’t comment on them, and leave that for someone else to eventually do a write up to facilitate others to be able to setup and integrate them with Sense.
3 Likes
This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.