Connecting Current Transformer to 0-20mA input

[DrEEvil]

Hi all,
I want to measure the power usage in watts that I am consuming in my home at a given point in time.
I want to use a 20A/20mA Current transformer(CT) connected to a IO module, an ADAM-6017.
This module has configurable analog inputs, and I will use 0-20mA. This reading will set a MODBUS register to
0-65535. I can then scale this in SCADA(Mango).

So the CT will output 0mA for 0 amps, through to 20mA for 20 amps.
Does anyone know if I have to rectify this signal first? The input spec does show + and -, but I am not sure.
The input has a built in burden resistor of 120ohms, which should produce a voltage of 2.4V AC on the input and 20mA at
full reading.

I have my GroWatt solar inverter feeding data into my Mango SCADA system via Digi Xbee Series2 (with a connectport X4)
and want to compare the power generated with the power consumed.

[Psi]

Yes, the CT output is AC so it needs to be rectified at some point and turned into RMS.

It will also go negative so you can't connect it directly to an input unless your input module can handle that.

If you're able to sample the waveform you should be able to rectify the reading in software. 

If you want to rectify in hardware have a look at precision rectifiers, they use an opamp to rectify without the Vdrop you'd get with diodes which destroys low voltage signals.

Your house current will peak at much more than 20A. The oven alone is probably ~30A
A 20A CT is more suited to measure the output from a powerpoint.

To measure the entire house a 75A or 100A CT should be fine.

[DrEEvil]

Yea, you're correct there, 10.9kW for the stove! About 45A, i guess with everything turned on.

I actually have two phases here, but still will need a bigger CT.

Any ideas on the rectification question? I think it probably needs to be rectified,
but may be dependent on the module's capability.

[Psi]

I had a quick look at the module but couldn't see anything definitive about reading negative voltages.

[IanB]

The ADAM-6017 is an instrumentation device, so it is going to expect smooth DC signals on its inputs. It samples the inputs 10x per second.

In order to use the current transformer you would need to convert the AC output from the transformer to an equivalent RMS DC signal to feed into the I/O module.

The most convenient way to measure power usage on an AC circuit is probably to use an off-the-shelf device like a Kill-A-Watt. Admittedly they may be geared towards single appliances and may be difficult to extract data from externally, but on the other hand they will do an accurate calculation of volts, amps, VA and watts and therefore will give you more accurate information than a homebrew solution.

[DrEEvil]

Thanks IanB,
I have a device like the kill-a-watt, but want to get the wattage in use into my Open Source SCADA system (Mango).
Because I only get 16 cents feed in for my excess solar, I want to be able to compare the solar generated with the power used,
and maybe have a little light com on when I am using more that generating. I could also have an indicator come on at a
power usage set point.

I can get an 'off the shelf' meter to go in the switchboard, that has RS485 MODBUS RTU capablilities, but just wanted to
reduce cost. These go for $350+ for three phase.

I have a heap of Advantech gear, and just wanted to use that.

I think if I full wave rectify the signal from the CT, and maybe add a filter cap, this might be enough.

What do you think?

[IanB]

If you pass the output of the CT through a resistor to produce a voltage, then rectify that voltage and filter it with a capacitor, you will get a DC voltage that is theoretically 1.414x the RMS voltage. This calculation will be upset by the diode voltage drop and other factors, so it won't be exactly 1.414. To improve that you would need to use electronics (op-amp buffers and precision rectifiers, for example). It can be done, but it won't be entirely simple, and you will get a system you need to calibrate to get the best accuracy out of.

[DrEEvil]

Thanks again,

I will use this simple method you suggest. I can calibrate the system somewhat in SCADA, using a test load with my multimeter.
If I find the data collected useful, I think I will go for a proper meter for accuracy.

The other option would be to use one of those Kilowatt Hour meters that have the pulse output. They are a LOT cheaper.
They output a pulse at a rate proportional to the Kilowatt hours, usually 1000 imp, or 1000 impulses
per kilowatt hour. But I don't want kilowatt hours, just watts.

Might be able to set and Advantech Digital input to read the counts, then divide by time to come up with watts.

[Lurch]

These are quite useful, and pretty reasonably priced. http://www.lem.com/hq/en/content/view/270/207/ I've used some of the self powered 4-20mA DC versions and fed them directly into various devices with good results. No external circuitry needed, they do what they say on the tin pretty much.

[cosmos]

Don't you have a meter showing the consumption of the house for billing?
That would be much more accurate than rectifying and hoping your loads are resistive or that your mains voltage has no harmonics.
If there is a meter you might want to have a look if it can be monitored (blinking LED or rotating dish).

[DrEEvil]

Thanks Lurch, I have requested some pricing from the Australian distributor. Ocean Controls here in Australia also have some like these:
http://oceancontrols.com.au/Current.html
Just want to keep cost to an absolute minimum.