Energy consumption readings

Hi Inovelli community!

I was looking at energy consumption in SmartThings app and noticed for my switches the data seems way off. My recessed lighting on a red series dimmer is almost never on but reporting its using 14.9 kWh and my fan+light switch is reporting 44.6 kWh. My entire house usually uses 30 kWh a day. Power meter is accurate for all my switches although fan+light switch does show 1w when off.

Am I reading this wrong or has anyone else had this issue with inovelli switches / smartthings?

3 different switches and all have neutral wires
red series dimmer - firmware 1.47:


red on/off - firmware 1.2:

fan+light switch - firmware 1.36

I have seen what I believe are inaccurate readings as well. Just posted about them yesterday.

Basically, I am seeing load being reported on switches that have nothing connected to the load line. I am on 1.57 if that helps.

Just making sure weā€™re talking about the same thing: kWh is not the current load, but the cumulative load since the switch last had its energy consumption reset. When did you last reset the energy consumption on the switch? Has it ever been reset since you installed it? If youā€™re not sure, I would reset the energy consumption tracking on the switch and see if it accumulates at the expected rate afterwards.

FWIW, I do know that energy monitoring at low levels tends to be pretty inaccurate without specialized hardware. I wouldnā€™t be surprised if thereā€™s a lot of variance with say a switch thatā€™s only controlling an LED bulb or two vs. a switch thatā€™s controlling a fan or other larger load.

2 Likes

Yup, that was what I was talking about as well so your assumption is right.

I have never reset the counters but HA uses the last time reported and compares the deltas. I also see it going up in historical graphs as well as the switch reporting 1W used on the energy sensor (so both energy and power values go up). All of which is strange because there is no wire on output of the switch.

Understood at the low levels and the difficulty pinpointing that. Might be a good one to get dialed in though because most of switches control LEDs in this point in time so there are a lot of low loads.

I will try resetting the counters.

1 Like

Out of curiosity, I decided to take a look at my own dimmers and switches; I have at least several dimmers which are consuming a constant 2.5W (100Wh every 39-41 hours) according to their own energy reports, despite being ā€œoffā€ and the power report indicating 0W. Thatā€™s a lot of phantom power if true, which is really tempting me to buy an ammeter and actually look at the power draw to verify independently. These have 1-3 15W dimmable LED lights connected to their load, depending upon the dimmer in question (all firmware 1.48).

@EricM_Inovelli Do you know how much power the LZW31-SN is supposed to draw while the load is off? Do yā€™all have any testing data with recent firmwares that either confirms or refutes the accuracy of the energy consumption reports Iā€™m seeing from my dimmers?

1 Like

I have noticed that the Inovelliā€™s minimum power report seems to be 1.6w. At 115 volts, 1.6w is only 15mA which is probably what the switch itself consumes.

Also what I have noticed is if the switch ā€˜stateā€™ is ā€˜offā€™ then it reports Power usage as 0 even though the switch itself is consuming some power. But if the switch ā€˜stateā€™ is ā€˜onā€™ it will report a minimum of 1.6w even if there is nothing connected to the load terminal. I assume that is because the switch itself is consuming some power even with nothing connected to the load.

I have not watched the kwH Energy report close enough to see if its accumulating the 1.6w switch power draw all the time the switch is off. I suspect it probably is, because the switch does use power all the time.

That would make sense.

So, state is as follows:
Off = 0w
On = 1.6w minimum + actual draw

Usage = 1.6w / hour draw, or an update of 0.1 kw/h every 62 hours or so?

Here is my kwh usage for a switch with no load connected for the past 30 days.

It seems to tick up about every 36 hoursā€¦ maybe not something to ā€˜worryā€™ about. Maybe closer to a 2.5-2.7W draw.

My curiosity / concern is that that seems like a LOT for a switch, especially when multiplied by 20 or 40 switches. My home stereo receiver is able to keep the Wifi, Bluetooth, Ethernet, and HDMI chips active on just 0.5W, so my expectation prior to actually looking at the logs was that each switch would be using considerably less since itā€™s only managing 1 wireless radio and an ultra-low-power one at that (~1mW for ZWave).

Thereā€™s probably not a lot that can be done regardless, but Iā€™m hoping this thread will be able to answer/confirm if the numbers Iā€™m seeing are: expected, accurate, or both.

15mA isnā€™t much. And most of that is probably the LEDs. Remember, the Inovelli Dimmers have 7 LEDs and if each LED is only drawing 1.5mA (@10% Intensity) thatā€™s 10.5mA just to light all 7 LEDs leaving only 4.5mA or 0.5W for the electronics and zwave radio.

You could turn the LEDā€™s completely off when the switch is off (Parameter#15 = 0) and I suspect that would drop the idle power from 1.6W down closer to 0.5W

1 Like

Iā€™m seeing similar numbers - 2.3W.
I have an energy monitor on most of the circuits in the house (www.iotawatt.com - another small, open, helpful company) and one circuit has 3 dimmers on it - with neutral, 3 or 4 way using aux switches - with everything off, it drawn 7W.
Out of curiosity, I tried setting the ā€œLED strip Intensity (When OFF)ā€ to 0% (on all 3) - it didnā€™t make any difference; still 7W.

Not to be pedantic, but are you sure there is nothing else on that same circuit? Its very rare to have a single circuit breaker feeding just one light. :nerd_face:

In addition to that, I believe the Inovelli 3-way/4-way implementation is also going to use some additional power in order to sense the position of the 3-way/4-way slave switch. :thinking:

With that said, Iā€™m just speculating all this and it would be nice to get some official specs from Inovelli regarding what the expected ā€˜idle power drawā€™ should be. The spec sheet shows what the maximum wattage is, but not the idle or off-state wattage. :straight_ruler:

Voltage matters, though; 20.8mA @ 120V (2.5W) is the max power output on a standard USB 2.0 port. Itā€™s enough to power the RGB backlights on an entire keyboard (104+ LEDs) at full brightness and the integrated arm processor, or a 7ā€ LCD display.

15mA @ 120V isnā€™t a lot of power, but you can still do a lot with it. LEDs like these draw in the ballpark of 0.044W; The processor should be haltā€™d. Does the Zwave 500 chip really draw that much power when running as a repeater? Is it the power leakage across the load?

This I think is the most important part. If these switches are increasing my power bill by $9.50/month (Because 2.5W per switch adds up), Iā€™d like to make sure thereā€™s not something thatā€™s been missed or some configuration parameter or firmware change that could be made to reduce this. $9.50/mo isnā€™t going to break the bank, but thatā€™s funds I could spend elsewhere. Like more Inovelli products. :wink:

1 Like

The breaker has those 3 lighting circuits (with 4-5 LED bulbs each), plus a few sockets - but the sockets were all unplugged (not just appliances turned off).
And yes, the accuracy at very low wattages is also slightly suspect. But the numbers seem rather similar to @BuilderTroy.
I make that about $0.22/month (per switch) - 2.5(W) * 24(hours) * 30(days) / 1000 (to get kwh)= 1.8kwh/month. My rate is a bit less than $0.12/kwh = $0.216/m. So 45 switches would get @darthandroid up to around $10.00/m.

2 Likes

Off by one; Very good!

1 Like

To me it is not only the $10/m but the principal.
One of the benefits of home automation should be increased energy efficiency, which includes the ability to identify vampire loads and eliminate them. One of the tools used to identify the loads should not itself be SIGNIFICANTLY adding to the loads.

In the example given by Sandy B above, a house with 45 switches would use an additional 81 kwh/ month. My home has only 22 switches, so Iā€™ll use 40 kwh/m for my example.
Last month, my house used 567 kwh total including running the ac. If I had 22 Inovelli switches, that would increase by 7%.
In the fall and spring shoulder season, my home uses about 200 kwh/ month. A house full of Inovellis switches using 40 kwh/month would increase that by 20%, which is SIGNIFICANT.

Iā€™d really like to hear from Inovelli about this. In the meantime, this is sufficient reason for me to put a pause on buying any more switches, at least until we know more.

2 Likes

Lets make sure weā€™re all on the same page here. I donā€™t believe the switches themselves are actually consuming that much power. The problem, IMO, is the value it reports which is not necessarily a super-accurate number ā€¦ especially at lower power levels. It doesnā€™t have milliamp accuracy.

And that right there basically supports my previous statement that the reporting is wrong at very low power levels and the switches are not consuming 40kWH per every month for every switch. Common sense says that smart switches would not be consuming 20% of your total electric bill.

That is the problem. Low level power reporting is not accurate without specialized hardware. The power sensing hardware in smart switches is intended for sensing higher loads in the 50w-500w range, NOT the low level 5mW-50mW range that the internal switch electronics use.

For the sake of this discussion, I hooked up a LZW31 Black Dimmer (donā€™t have a spare Red at the moment) on my test bench with a kill-a-watt type of device which is not particularly a super accurate device, but its capable of measuring as low as 0.1w and its reporting 0.3w for just the switch itself (no load, no bypass) and that comes to 0.22kWH per month. Where I live ($0.11/kWH) thatā€™s only 2.4 CENTS per month - no where near the $9-$10 per month that some of you estimated.

I believe there is something in the hardware/firmware that imposes a lower-limit on the power reporting. For my Inovellis that number seems to be 1.6W. Thatā€™s the lowest value it reports for me under any conditions, but Iā€™m confident the actual power consumption of the switch itself is less than that. Iā€™m not saying 0.3W is the most accurate number, but I do believe its a lot closer than the 1.6w-2.3w that the switch is reporting with no load.

2 Likes

0.22kWH/m per switch would put 22 switches at 4.84kwH / month, which is MUCH more reasonable.
Thank you for the clarification.
It would still be nice to hear from Inovelli on how much the switches are supposed to draw.

I put a red dimmer on a kill-a-watt type meter and got 0W and 0.00A.

I would like to test with a load and see how much current the device leaks into the load.

I agree completely. @EricM_Inovelli or @Eric_Inovelli can you provide us with a semi-official number of what we should expect for the idle power draw of the dimmer itself? (dimmer powered but load off)

EDIT: to be clear, I understand the power sensing/reporting in the dimmer is calibrated for higher wattage and probably not capable of reporting below 1W. This question is about the actual power consumption of the Inovelli switch/dimmer itself with no load. It should be well below 1W.

Looks like that device is not capable of measuring tenths of a watt (0.1W) or milliamps (0.001A) which is the resolution needed to accurately report the power usage of the switch itself. It also illustrates why switch reporting at very low power levels is neither accurate nor intended.

That would be interesting. Keep in mind that its probably going to be different depending on how you have it configured with regards to using a wired neutral or no neutral (parameter 21), Load-only or 3-way Toggle (parameter 22), and possibly also depending on the actual load you connect (incandescent, LED, CFL, etc.)

1 Like

@mamber Probably not, there are no specs on what it can measure. But, it does show the dimmer itself doesnā€™t draw >2W.

1 Like