Calibration 3-phase

Submitted by Bart on Tue, 22/12/2015 - 13:31

Hi everyone,

I have a EMONTX V3.4 and try to measure 3 phase by using the 3phase sketch. Till so far it works, only the calibration is still hard, because I'm not sure if I'm doing right.

Vcal calibration was easy, the same for Ical for ct 1-4. Only now I have to calibrate the phase shift to a power factor of 1. For ct 1 and 4 I fixed it by changing Phasecal, according to the guide. But there are strange results, to my opinion. (CT2 has a negative real power and negative power factor)

I have now attached ct1-4 on one single phase positive wire. The print screens of the results are attached to this topic.

My knowledge isn't that deep that I know what happens is correct, hopefully you can confirm.

Looking forward to an answer

Re: Calibration 3-phase

Submitted by Robert Wall on Tue, 22/12/2015 - 13:53.

If you put all 4 CTs on the same wire as the VT, then because the voltage is phase shifted in the software expecting the currents to be phase shifted too (but they're not), you should read:

Channel 1: Real power = apparent power, pf = 1.00
Channel 2: Real power = -0.5 × apparent power, pf = -0.5
Channel 3: Real power = -0.5 × apparent power, pf = -0.5
Channel 4: Like Channel 1 (or 2 or 3 if you've altered it to be on those phases, the default is Channel 1)

For the reasoning, look at the vector diagram below and think about the component of the yellow & blue phases that's projected on the real axis (the red phase). The projection is exactly half the length of the red arrow. Remember
cos(120°) = -0.5

3-phase vector diagram

So it looks as if your Channel 3 CT is the wrong way round on its wire, and you need to tweak phasecal a bit more.

Re: Calibration 3-phase

Submitted by Bart on Wed, 23/12/2015 - 10:53.

Your explanation is clear and makes sense to me, thanks!

Now I have used:

#define PHASE2 6 // Number of samples delay for L2
#define PHASE3 12 // Number of samples delay for L3

double Phasecal1 = 1.05; // Calibration constant for phase shift L1 - 1.00
double Phasecal2 = 1.38; // Calibration constant for phase shift L2
double Phasecal3 = 1.325; // Calibration constant for phase shift L3
double Phasecal4 = 1.23; // Calibration constant for phase shift CT 4

Why is it that the prescribed "starting values" deviate that much from my calibrated values?

Re: Calibration 3-phase

Submitted by Robert Wall on Wed, 23/12/2015 - 12:03.

The "prescribed starting values" worked for me! The phase shifts inside the transformers are heavily dependent on the actual current and voltage at which they operate, and especially for the third phase, there is a big time delay - 13.33 ms so a few percent difference between the speed at which your emonTx runs compared to mine could make a lot of difference. With phasecal > 1.0, you are actually extrapolating, so that is equivalent to moving to the next array element and using a phasecal less than 1, which is interpolating (and to my thinking, preferred).

Archived Forum