CT sensors vs other more direct sensors

Submitted by Guest on Mon, 02/09/2013 - 10:36

Hi guys,

CT sensors seem to be the preferred way on this site to measure power consumption but I'm worried that these might not be accurate enough compared to other devices that measure power consumption in a more direct way like this DIY home energy monitor or the power consumption meter that I bought in a nearby store.

The author of the linked DIY home energy monitor mentions that the switch-on profile of his fridge can be seen so clearly that maybe even the expected lifetime could be derived from the graph.

So, to sum it up I'm interested the resolution of CT sensors vs other more direct sensors.

Thanks in advance!

Laci

Re: CT sensors vs other more direct sensors

Submitted by chris (not verified) on Mon, 02/09/2013 - 14:25.

I'm interested in the same thing. I'm looking to monitor the energy usage of three pool pumps and possibly more devices.

Cheers

Chris

Re: CT sensors vs other more direct sensors

Submitted by Robert Wall on Mon, 02/09/2013 - 15:47.

The principal concern with anything involving a direct connection to the mains is safety. As soon as you make a galvanic connection to a live conductor, the whole circuit must be considered as live and the appropriate safety precautions observed. Generally, this means that you have to have an optical or magnetic barrier and a separate power supply for the live end, and the design must maintain the required isolation and tracking distances; and you cannot make a direct connection to the live end even for programming or testing. I think you'll find the commercial unit is double insulated and has no means to extract the data (other than optically - read the numbers!) and Dave Berkeley's design follows the optical barrier and isolated power supply route.

The inaccuracies in c.t's are well documented and it's likely that other elements in the design impose much more severe limitations on the accuracy.

Re: CT sensors vs other more direct sensors

Submitted by mondalaci (not verified) on Thu, 05/09/2013 - 11:28.

Thanks for the explanation, Robert.

> The inaccuracies in c.t's are well documented

Well documented where? Could you please reference some relevant links?

> and it's likely that other elements in the design impose much more severe limitations on the accuracy.

In the design of CT or non-CT sensors?

Re: CT sensors vs other more direct sensors

Submitted by mondalaci (not verified) on Thu, 05/09/2013 - 11:35.

Ok, even though I've managed to find some info on CT accuracy on Wikipedia I'd be interested in a comparison.

Looks like this SparkFun CT sensor features a 2% accuracy. Seems pretty damn good to me considering that we're talking about non-contact measurement.

Re: CT sensors vs other more direct sensors

Submitted by Mattia Rossi on Thu, 05/09/2013 - 12:09.

http://openenergymonitor.org/emon/emontx/accuracy
http://openenergymonitor.org/emon/buildingblocks/report-yhdc-sct-013-000...

http://openenergymonitor.org/emon/buildingblocks/measurement-implication...

It's not really a matter of accuracy regarding the ct, but a matter of accuracy regarding arduino's adc converter at low voltages....

Re: CT sensors vs other more direct sensors

Submitted by Robert Wall on Thu, 05/09/2013 - 12:51.

Well documented where? Could you please reference some relevant links?

Any decent electrical engineering text book that covers the theory of a transformer should cover this.

It's not really a matter of accuracy regarding the ct,...

But the c.t. core losses are proportionally greater at low current, and saturation starts to cause non-linearity at high currents. Admittedly, ADC and processor noise problems mask the c.t's contribution to the error at low currents.

A c.t. is best in the middle of the range, so like anything else, it's desirable to size the c.t. correctly. If you are desperate for accuracy, invest in a revenue grade c.t.

Archived Forum