How to verify emonTxShield measurements?

Hello everyone!

I have set up an energy consumption monitor using the EmonTx Arduino Shield SMT, the Arduino Leonardo board and the YHDC Current Transformer SCT-013-000.
I use the meter in order to measure the current only, since i don’t have purchased any AC-AC adapter.

The result is that i take measurements (different than noise) when i wish to measure the current that consumes my PC for example.

My question is how can i verify the measurements that my meter returns? Are they right and real?

One could say if the meaurements are reasonable, meaning that measurements do not exceed the maximum measurements that i could have regarding the maximum PSU power(350W) and the voltages from the supplier (230-240 V).

But how could be sure for the rightness of the measurements?

p.s. Sorry if i miss any related post.pls inform

You will not be able to measure real power with your emonTx Shield, because you do not have the a.c. adapter. You will only be able to have apparent power ( = voltage (230 V) × current).
Sadly, it is possible that, for a computer power supply, apparent power will be much larger than real power. The second problem is the current that you are measuring (~ 1.5 A) is very small compared to the maximum 100 A that the emonTx with a YHDC SCT-013-000 current transformer can measure, and it is likely that the measurement will not be very accurate.

You can check the accuracy very easily, but you must use a resistive load (e.g. a heater, kettle) and you need an accurate ammeter to measure the current. If you go to Resources > Building Blocks, there you can read how to calibrate your Shield.

Trying to clarify here what I believe to be a calibration technique. I read the building blocks, but I’m stumbling over some of the words…

My technique was to load the sketch for the EmonTX Shield (Shield_CT1234_Voltage_SerialOnly). Using a true RMS meter, I adjusted the voltage calibration so the output of the sketch was the same as what I measured.

Where I think I messed up was I used the multimeter clamp to measure the amperage of a load to calibrate the CT. But I was comparing the multimeter (clamp, not inline current measuring) to the Irms output of the sketch. Is that wrong? Should I load a sketch without voltage input and adjust to get the same values with that? That is, maybe load the Shield_CT1234 sketch and adjust the calibration value to get the correct output and then use that cal factor in my sketch?

My multimeter is a Klein CL800. Its cheap, so not as accurate, but still ‘claims’ to be true RMS.

Basically, all I am trying to do is get a calibration from a known value. I know what the meter reads, I just am not sure which value (sketch variable) to compare it to. And I’m trying to see if using a known value is more accurate than using the label on a resistive load, since it seems like that probably isnt as accurate as my meter.

And then I still have to work on the Power Factor (phase shift) adjustment.
The objective of all this is to get as close to the same values as my electric meter gets to measure usage (not used for monitoring a PV array, etc).

There’s no difference between inserting an ammeter (or multimeter on the current range) in series and using a clip-on ammeter. There’s nothing wrong with setting the sketch current calibration that way, it is exactly what Step 3 of the calibration procedure tells you to do.

You don’t say what you mean by “sketch variable”, and I certainly would trust a calibrated meter in preference to the label on an appliance.

Thanks Robert.
I guess I assumed that going in series was more accurate than a CT. Most meters I have seen do AC via clamp-on and DC via 'in-line" (series physical connect). What I cant tell is if my meter gives RMS values for current as well as voltage.

What I meant by the variable was that I had the sketch continuatlly printing the ‘Irms’ value returned by the EnergyMonitor class in EmonLib.h. I adjusted the CT cal factor until Irms was the same as the meter. I wasnt absolutely positive that those should be the same in a correct calibration.

I did those two but need to do two more things.
Voltage output by the sketch (Vrms) is correct under no load, but I need to check it under load.
Adjust the power factor using a purely resistive load.

I’d be absolutely amazed if it didn’t! There’s no reason not to do both as true rms if it does one - it’s only the front-end scaling that is different, the remainder of the signal processing and display is common.[quote=“hueydriver, post:5, topic:775”]
Voltage output by the sketch (Vrms) is correct under no load, but I need to check it under load.

But your meter and the Shield should still read the same (lower) voltage!

And I am guessing that calibration factor should be fairly linear! Just validating at a more realistic current value.

One more question, I guess.

When calibrating the power factor, do I need to ensure there is no other load on the circuit, or main, that I am using? I got some awfully weird PF cals (1.2) based on what I SHOULD have gotten, so wasnt sure if there was something else on the branch that was influencing the value.

Yes, you must have only a purely resistive load when you’re setting the phase error compensation. You cannot have a power factor of 1.2, because real power can never be greater than apparent power. If you have, it most probably means that you’re using a different value for voltage in the rms voltage calculation to the value you’re using in the real power calculation. Some sketches do that, and it’s wrong. It happens because the way the voltage wave is (apparently) time-shifted also affects its amplitude. This is explained in the piece in Resources > Building Block Resources about the phasecal algorithm. If your sketch does this, you should adjust for maximum, not a p.f. of 1.0, then check the voltage calibration again.

I was vague in my post and made a mistake. What I meant was that I was following a post I had found that said to calibrate the phase shift (I called it power factor calibration) in the code “ct1.voltage(0, 300.6, 1.7);” by using a pure resistive load.
I used a 100w bulb (pretty hard to find these days!) on an extension cord I have split to use with the clamp-on ammeter (as a line splitter). I adjusted the phase shift value until the PF was 1.0 and the real and apparent power values were the same. That resulted in the value of about 1.2 in that line of code.
My question was that the extension cord was plugged in to a power strip that also had my laptop charger, an inkjet printer, a monitor, and a couple AC-DC adapters. I was still sampling only from the extension cord, but there was a lot of switching power supplies there. Honestly tho, I dont see the difference though because no matter where I would plug in, anything on that main would seem to be at least affected.

Don’t you have an electric kettle or something larger than 100 W? - something closer to a typical load you’ll be running at? At least put multiple turns through your CT, because the CT’s errors are greatest (proportionately) at low currents - see the test reports.

As long as your resistive load is downstream of the place where you are measuring (and you only need your Shield CT for this), anything plugged in upstream is irrelevant.

I’d be cautious about following any post - the definitive instructions are in Resources (link above).

I’m going to try and find something bigger. Interestingly, as I looked, nearly everything I have uses an electronic control, which while very small, will change things a little. I did do 4 wraps of the cord through the CT, but was still only around 3.4A I think. I do have a hotplate that would be good, if I can keep it from cycling once it reaches set temperature…

And once again, I was mistaken but mostly just couldnt remember where I saw the calibration procedure but it was in fact the 4th step in the building blocks calibration page…

After making the phase angle calibration, I actually got a large increase in accuracy (at least over the 24 hours of so I watched). I have period pricing on my home which is why I am trying to keep track so closely. Peak rates are much, much higher so by monitoring I hope to determine the effectiveness of different steps I take.

I’ll open this on a different thread, if it helps, but I went ahead and got a second shield so I have more accurate measures of each main. The more I looked, I just figured it was worth $30 to get the increased accuracy because while the voltage difference between the two is relatively constant, I’m sure the power factor is way off. So far, it has been a worthwhile exercise…