I’ll try tomorrow on another socket just to see if the problem is with the specific socket
There is a voltage reset channel, add it, turn it on/off, then restart the socket (kill power for 10 secs).
This sorted all mine out
Don’t think I broke it in latest versions. I had a few all messed up on the power and this basically resets the power metering chip so it reinitialises (found it by pure chance when developing 
)
Seems I have issues with all sockets. Power metering is not accurate. Another example is my rack, I have a ups that is using 305W (I can see that in the web gui) and the socket shows usage of 200W.
I’ll send them an email
I tried what Dave suggested but still the same
Un installed, re downloaded a better looking size this time, Binding loading correctly
The response from Lightwaverf
"Thanks for your query.
The devices are to be used as an indicator of power usage.
The energy monitoring function is for indicative purposes only, when the measured load contains capacitive or inductive components (most switch mode power supplies), it does not have a high accuracy."
You should be able to get/work out the power factor for inductive/capacitive loads and have some rules going. Think this is probably the case with mine too but never actually checked.
The hair dryer should be accurate though as this is resistive…
I try and get some time to check mine over the weekend
@pentiumdb does seem to be out (see below - Kettle) which makes it pretty pointless. I’ll do some more testing against my clamp meter and energy monitor over the weekend. Pee’d me off tbh as was one of the reasons I chose LW.
I’ll calculate against the voltage reading over the weekend too as it may be that reads incorrect and therefore power is incorrect
Same here.
It seems there is an offset depending on the load, I noticed a difference from 20% up to even 100%(hairdryer example)
I’ve just looked through all of their spec sheets / sales literature. There is nothing in there to specify what you have quoted from them above. It clearly states ‘monitor energy usage’. Once I’ve got enough data I’ll see if I can find any pattern etc. If I can correct it through the binding I will, but I’m certainly going to bring it up with them as to me this is not acceptable.
My kettle is a resistive load so from what they are saying this should be accurate (I understand capacitive/inductive won’t be) - And I would expect this to be accurate to at most 5%
Have you measured the kettle load, I bet it alters at it warms up, a bit, also the voltage can vary place to place
I use an rfcom monitored current coil that I pull the readings into openhab, I’ve not checked accuracy
( I am an electrician by trade )
I’m a fridge guy so yeah electrics are my thing too 
Still… 3500 / 240 (voltage normally there or there abouts) = 14.58 amp… thats not right.
and even if the voltage is lower, amperage is going to increase.
#InvestigationRequired
Still not as good as my 11kW reading 
Tried again with the hairdyer on several sockets. My mains measurement always went up approx 2000W. The reading on several LW sockets was different. I got a range of 1030W to 1250W for the exact same load
shocking 
@Fixer - you have the energy monitor don’t you? Maybe you could test the difference on this for a known load
The wattage is not the constant in the calculation the resistance is
Even the resistance is not constant as this will change with temperature and humidity.
Going to do some testing shortly…
So have run a few tests with various resistive load devices and the accuracy against my test meter is as follows:
@3000W LW 17.5% higher power
@1500W LW 17.5% higher power
@900W LW 19.2% higher power
@500W LW 22% higher power
Going to hunt for some low power devices over the weekend see if I can get a full spectrum, I’ll be able to work this into the binding then. Seems the top end of the scale flattens off so need something to test between 900 and 1500w to find the drop off point.
If anyone can also do any testing if they have the gear then that would help to confirm.
Will also then work on the capacitive / inductive but I’m a little restricted on what I have available on this front.
Nothing complex needed tho… 
