Accurate Clamp Meter? I don't have a Smart Meter and want silly graphs

GeX

GeX

Soldato
Joined
17 Dec 2002
Posts
6,979
Location
Manchester
I have https://docs.openenergymonitor.org/emonpi/technical.html that is monitoring our incoming mains supply. It has
  • Current Transformer rated to 100A
  • AC voltage sesnsor
  • Optical Pulse counter that is attached to the meter, counting kwh

tkCe7wpl.jpg


That also feeds into Home Assistant. Throughout the house are UP111 smart plugs, that also feed into Home Assistant. This allows me to draw graphs like this



Before installing the current clamp on the meter tails, I calibrated it using a Fluke 179 to measure current and a Brymen meter for the voltage. I used a kettle as the load for doing this.

I used the same kit to calibrate the smart plugs, and they all read close to each other still.

The problem I'm seeing is that sometimes there is more monitored load than current transfer thinks is coming in;

RskdjVEl.jpg


I'm using the incoming data from that to total up how much power is consumed in the house per day and the reading from the current clamp consistently reads lower than the data I get from reading via the pulse counter on the meter. This difference is typically 4-5% per day.

To try and get some sense, I bought a clamp meter to run next to it and verify. The two sensors did not agree.

Now I'm not sure what to trust, so wired my Fluke 179 inline with the kettle and put the clamp meter on the live wire - to see what each meter said.

VP6Lq05l.jpg


Well that's not right, I'm more inclined to trust the in-line meter. I then moved the postition of the wire within the clamp and...

nCJDSecl.jpg


That's not helpful. It's quite a difference in just moving the conductor from the end of the jaw to near the meter body.

So now I don't know what to trust.

My Fluke clamp meter is old, and also reads up to 200A. (400A in the other setting). I don't need that scale, so started looking at 80A-100A meters instead. Prices vary massively and I don't want to buy something that's not going to be accurate, else what's the point.

Or am I chasing differences in how the two measuring devices work. The smart plugs can show me power factor and active / reactive power. What is the current transformer doing.
 
Last edited:
A fair question; https://forums.overclockers.co.uk/threads/gas-meter-earth-bonding-is-it-required.19001597/

Although I’d probably still be doing all this silliness, just with another data point
I suspected the latter, and should have remembered the former thread :D

So the simplest integration for smart meters is a Glow from Hildebrand. I don't know if it'd work without a smart meter but it's popular as the way to bring your meter data into your home automation setup. You could check if that works out of the box.

I would also get a smart meter TBH - I can imagine installers might not be all that careful beyond "Replace current meter if it's safe to do so". Our place had pensioners in before and everything was basically original in the house, aside from fitting smart meters. No equipotential bonding anywhere at all - gas or water.
 
yes clamps aren't accurate V .. but what's wrong with a pulse counter ?
these seem to have been redesigned/back in stock, which is what I'd considered.



I subsequently read a facebook comment on the non-zigbee shelly clamp product (which I think is premium/bounding versus tuya) that suggested +-10% , so yes not so good

Shelly EM is not factory calibrated device and on small loads it is quite inaccurate (loads bellow 1A or so). So in whole day cycle depending on loads yes it can accumulate to that error margin you have noticed.
etc.

hildebrand glow at £50 , which would give dynamic gas usage too looks increasingly attractive - get both your smart meter data every 30s from DCC
 
yes clamps aren't accurate V .. but what's wrong with a pulse counter ?
these seem to have been redesigned/back in stock, which is what I'd considered.

I have a pulse counter but I didn’t think they were much use for instantaneous use measurements as the time period they measure in can impact the readings.

I’ve not figured out the maths to get that from my pulse counter
 
I have a pulse counter but I didn’t think they were much use for instantaneous use measurements as the time period they measure in can impact the readings.
yes maybe you are right - thought they had 1000? pulses for 1kwh yes HA has to do some maths.
I had thought smart meter uses same data, it's 'real' time power (octopus i have) shows power with watt tolerance, possibly it updates after every pulse, so if last pulse was 10s ago it shows 360W,

anyway clamp meter has inaccuracy due to voltage variability, and you'd have have to decide on sampling rate and integration period, led lights and electronics are using switch mode power supplies at some high frequency,
so hard to judge what sample rate you'd choose.
 
Glow is what I'd be using. I don't imagine it supports non-smart meters.

My question was more on is https://uk.rs-online.com/web/p/clamp-meters/1469096 something like that going to help

or https://thecalibrationcentre.co.uk/product/di-log-dl6506-mini-80-amp-ac-dc-clamp-meter/ or something else. My default is to just buy Fluke stuff, but that costs an awful lot.
Depending on which Glow bit of kit you have, it'll either just tap into the database your smart meter would report readings to, or connect to your smart meter directly to get live usage. Eitherway, you'd need a smart meter
 
Depending on which Glow bit of kit you have, it'll either just tap into the database your smart meter would report readings to, or connect to your smart meter directly to get live usage. Eitherway, you'd need a smart meter
By "Glow is what I'd be using", I meant if I had a smart then I'd use Glow. I don't have either.
 
yes maybe you are right - thought they had 1000? pulses for 1kwh yes HA has to do some maths.
I had thought smart meter uses same data, it's 'real' time power (octopus i have) shows power with watt tolerance, possibly it updates after every pulse, so if last pulse was 10s ago it shows 360W,

anyway clamp meter has inaccuracy due to voltage variability, and you'd have have to decide on sampling rate and integration period, led lights and electronics are using switch mode power supplies at some high frequency,
so hard to judge what sample rate you'd choose.

I think this might be it, work out how to have it trigger on each pulse rather than at a set time period. Then use the time period to work out how much was consumed.
 
I looked further and the smart meter IHD's are listed as getting 10s updates, which I think is all that the Glow&co would then access.

this is an interesting paper on how to diy.
you need to detect voltage/current/power-factor and integrate the 'result' over the period you choose to output, seems they just get 1000 consecutive samples which for them then gives 120hz logging,
so, really real-time.

//T=micros();// 29 samples -> T = ~ 8.350ms for loop which is equavalant to 120Hz cycle m 0.287 seconds for 1000samplesfor(samples=0; samples<1000;samples++){ V_sensor = analogRead(A0); I_sensor = analogRead(A1);Serial.print(V_sensor);Serial.print(" ");Serial.println(I_sensor); float V = (V_sensor-V_offset); float I = (I_sensor-I_offset); sump = sump +V*I; I_rms = I_rms + I*I; V_rms = V_rms + V*V;}Serial.print(V_sensor);Serial.print(" ");Serial.println(I_sensor);T=micros()-T;t2= micros();td=t2-t1; // Period time in micro seconds
 
Back
Top Bottom