400A service North America - emonpi and CTs

Jason_Yates · 2018-06-20T00:22:32.148Z

I have recently purchased an Emonpi and am trying to setup monitoring for a 400A service site. Considering the transformer is only 50KVA I opted to go with a 200A CT. Due to the wiring constraints and the aperture size I decided to go with the Wattcore WC4-200-MA100. After reading the North American guide and some of the other posts here I soldered two 22 ohm 1% resistors in on the daughter card in the provided thru holes. After soldering the connectors (white to tip, black to sleeve), the power reported is off by a large factor (one CT is reading 2 - 4 watts, the other is reading around 3200 watts. When I measure the current with a clip on meter the current on both each legs is around 50A so I should be reading around 24kW instead of 3000W. My questions are:

Is there anything I am missing regarding the configuration of the software (i.e do I need to change a multiplier?)
Any suggestions on what to try to help me determine where my problem resides ?

Thanks
Jason

Robert.Wall · 2018-06-20T08:36:52.559Z

First, the leg that measures 2 - 4 W. Have you got an inadvertent short circuit across your c.t., or is your c.t. open-circuit somewhere?

The other leg: have you changed the calibration coefficient, either in the “emon” part of the emonPi, or in emonhub.conf? Obviously, by changing the c.t. and the burden, the default calibration will be wrong.

A couple of hints: You can use a milliammeter as the burden for your c.t., you should measure 100 mA at a full load current of 200 A, or proportionately less at lower currents.
The burden value is calculated to give approx 1.1 V at full load, and the calibration factor in the sketch (not in emonhub.conf) is the current that would give 1 V across the burden.

Jason_Yates · 2018-06-20T14:12:46.053Z

thanks for the help. It is working now. ended up with a scaling factor of 4.28. Both legs are working as expected.

Robert.Wall · 2018-06-20T19:04:53.289Z

I would have expected a calibration factor of 4.0, rather than 4.28 (because our standard c.t for the default calibration is 100 A : 50 mA, yours is 200 A : 100 mA).

What was the problem with the low-reading channel?

Jason_Yates · 2018-06-20T20:58:28.016Z

I tried 40 and it was close. I measured the current on both legs and coverted to watts and it was a little off so I tweaked to to 4.28. After it collects for a bit I will compare it to the pulse count to see how far off it is. I will update this thread with my results.

Oddly enough I did not do anything to the other leg, I rebooted and unplugged / replugged the sensors and they both started working as expected.

Robert.Wall · 2018-06-20T21:05:20.477Z

Bear in mind that a sensor (not just a c.t.) won’t be recognised unless it is plugged in (and in the case of the v.t, powered) when the emonPi is powered up. Rebooting the Pi won’t restart the “emon” part, which is what is needed.

It’s quite possible therefore that the voltage calibration is off too. Also, the phase calibration is probably wrong, but that can only be adjusted in the sketch in the ATMega 328P, there’s nothing you can do about that in emonHub.

Jason_Yates · 2018-06-20T21:22:46.028Z

Last week, while I was waiting for my CTs arrive I measured the voltage and it was off a bit from the value reported via VRMS. I modified the scaling factor for that to 0.01083. Could this be why the CT is off when measured? SHould I undo the VRMS scaling ?

Robert.Wall · 2018-06-20T21:30:49.816Z

But of course (or maybe not!) that doesn’t affect the power value that arrives in emonHub, because the power calculation is done in the front end.
What happens is this:
The starting point for understanding is the emonPi is a stripped-down emonTx piggy-backed onto, and sending values to, a Raspberry Pi. The ‘emon’ front end measures voltage samples and current samples, multiplies them and averages and calls it real power, also calculates rms voltage, scales those numbers using its own calibration factors and sends those values to emonHub. EmonHUb can then use scales= to tweak the numbers - or in your case rather more than tweak.
So no, you should not undo the Vrms scale adjustment, but it didn’t affect the power reading.