AC 230V consumption measurement

Hi,

I'm working on a power measurement on 230V (Europe) with wireless data transmission. The wireless link (2.4GHz) is already working. I'm now looking for any way to measure power consumption on a single phase 230V 50Hz line.

I think that I will use a MCP3909 chip to get active & apparent power but I don't now which current sensor I should use.

I'm thinking about a simple shunt resistor, a current transformer or some kind of hall effect sensor. Shunt is the easy way to do it but this brings a voltage drop on a supply line and an additionnal power consumption (heating).

I don't know anything about current transformer only that secondary should not be left open (high voltage).

Hall effect would be perfect however I'm not able to find a small (it will used on a 10x10cm PCB) with low current consumption (max 10mA) and have a good accuracy (1% error max).

Has someone already done something like this ?

Thanks for your help !

Delivereath

The MCP3909 has an input amplifier and will work down to 30 mV so unless the current is very large there won't be much heating.

I appreciate that isolation for safety could be a problem, but as you are transmitting the data by wireless, that shouldn't matter.

Current transformers tend to be larger than shunts.

But how to get a complete range from 0.1W to 2kW ?

To get 30mV on the shunt with 0.1W on the load (0.435mA), the shunt should be a 70ohms resistor... And 70ohms with 1A does 70V drop.

**broken link removed**

Meets all your specs except has a 1.5% error.

That seems to be exactly what I was looking for. I'm just a bit surprised that this small chip can several amps without any problem...

In fact, there's still a sensivity problem. This IC has an output of 185mV/A. With 1mA output would be at 185uV and I can't measure this with MCP3909...

Umhhhh, please define 'error'. 1% of 2kW is 20W.

So you're looking for five digits of precision with two digits of accuracy. Is this sensible?

Well you're right. I can't measure a range from 1mA to 10A with a full-range error of 1%...

As I'm doing a home current consumption monitoring system, what would be an acceptable power sensitivity ? 0.1W, 0.5W, 1W ?

0.1W means a current of 0.43mA which is probably impossible to be measured with a hall effect sensor. 1W represents 4.3mA on 230V line.

Datasheet of Allegro ACS712 hall effect current sensor indicates a sensivity of 185mV/A but a noise of 21mV. So I will probably not be able to measure anything below the noise level. The problem is that 21mV with 185mV/A sensivity reprensents 114mA. Is it really the minimum I would be able to measure ? Or I'm wrong ?

Actually, the 185mV/A is for the 5A model. I think you need the 20A model which has even less sensitivity.

I think anything less than about 20W is not even worth measuring.

Buy a "plug in power meter" made in china, (i.e. **broken link removed**), take it apart and see how they do it?

The 20A model has a sensitivity of 100mV/A and a peak-to-peak noise of 11mv in a 2kHz bandwidth. You should be able to reduce that noise significantly by analog filtering the output since you only need about a 100Hz bandwidth to measure the powerline frequency current.

It's highly unlikely that the power consumption in a home ever drops below 10-20W due to all the various electronic vampires, chargers, and clocks plugged in. Even a modern refrigerator has electronics that are on all the time, even when the compressor is not running. Anything with a remote control generally uses at least 1W even when off.

But in the case I'm monitoring any mobile charger or small devices a range from 20W to 2kW won't be enough. As this kind of chargers can have a stand-by power of 1W. I think that 1W is the lowest value I will need to measure but this means that current will be 4mA.

I'm thinking about a multi-sensors solution like a hall effect sensor for high currents (>500mA) and a shunt resistor for small currents (<500mA) but I don't know exactly how to separate the two sensors and use only the one that is needed.

Build two meters?

I think that I will try to use 2 MOSFETs to have a 230V switch. One line with a 10 ohms resistor and the other line with a hall effect sensor. I should be able to switch between both sensors depending one current (<100mA for shunt resistor, >100mA for hall effect sensor).

I've seen some 400V MOSFETs with 0.2ohm resistance when switched on with a 10amps current.

Do you think that this could be a solution ?

Thanks