DIY whole house electricity monitoring

[crna1]

Not quite. A CT can measure one direction and total. The total doesn't tell you which direction so you have to use total - polarized to get the other direction. Since I want to see produced/used data it's a bit more complex.

Actually, a CT does not lose information about current "direction". The whole current waveform including current direction, or more correctly phase, is read by the power metering chip ADC and used during power calculations.

The chip I use in my home-made meter, an ADE7753, will show negative/positive values depending on how the CT is mounted, or on the phase/"direction" of the current. So, a CT can be mounted so that when you generate power, you would see negative values, and when you consume you would see positives with a single CT.

Apparently, Brultech engineers decided to take the current absolute(unsigned) values in order to avoid confusion when mounting the CT since very few customers generate electricity

this homemade meter...think I might get plans for it, or maybe a prototype laying around gathering dust somewhere?

 

[vc1234]

this homemade meter...think I might get plans for it, or maybe a prototype laying around gathering dust somewhere?

I just followed the datasheet recommendations on page 15 ( http://www.analog.com/static/imported-file...ets/ADE7753.pdf ).

While assembling the meter part with the chip and about a dozen external elements on a home-made PCB should not be too hard for someone with basic electronics skills, interfacing to a PC may be a bit of a challenge. You can read actual power (as well as apparent power, current, voltage and frequency) from the chip registers via an SPI interface. You can do that either by implementing SPI signaling on a PC parallel port (if you still have one), or interfacing through a microcontroller (such as e.g. AVR family). In any case, some programming is required which, while simple, may be beyond a typical DIY'er skills.

I used a uC that I already had, so for me the project made sense because I could measure power with good precision and relatively cheaply:

ADE7753 $5
Various capacitors/resistors(see the datasheet), terminals ~$10-15
CT $6 ($16 split core, up to $40-50))

$26-40 per I/V channel total, maybe more due to the split-core CT cost for the large diameter power cable.

A conveniently programmable uC may cost you additional $20-30. I also used a $16 Ethernet/embedded TCP/IP miniboard from WizNET (http://www.saelig.com/miva/merchant.mvc?Screen=PROD&Product_Code=ETH042&Category_Code=BRD). That one is optional though because a uC usually has a serial port.

You will also need a voltage sensor, one per each channel or one for all the channels (if you are willing to sacrifice some precision). As a voltage sensor, you can use a 120->12/9/5 V wall transformer or a voltage divider. The former would be much safer as you'd achieve complete separation from the dangerous mains voltage, a DIY'er should not attempt to build a voltage divider for obvious reasons. The chip ADCs expect getting +/- 0.5 Vpeak, so the current/voltage channels should be suitably adjusted by choosing the CT load resistor and voltage-dividing the voltage transformer output.

The assembled gadget will need to be calibrated with a known load. It's a simple but tedious process which is fully described in the datasheet. I just used some resistive loads while measuring voltage/current with a multimeter.

 

[crna1]

this homemade meter...think I might get plans for it, or maybe a prototype laying around gathering dust somewhere?

I just followed the datasheet recommendations on page 15 ( http://www.analog.com/static/imported-file...ets/ADE7753.pdf ).

While assembling the meter part with the chip and about a dozen external elements on a home-made PCB should not be too hard for someone with basic electronics skills, interfacing to a PC may be a bit of a challenge. You can read actual power (as well as apparent power, current, voltage and frequency) from the chip registers via an SPI interface. You can do that either by implementing SPI signaling on a PC parallel port (if you still have one), or interfacing through a microcontroller (such as e.g. AVR family). In any case, some programming is required which, while simple, may be beyond a typical DIY'er skills.

I used a uC that I already had, so for me the project made sense because I could measure power with good precision and relatively cheaply:

ADE7753 $5
Various capacitors/resistors(see the datasheet), terminals ~$10-15
CT $6 ($16 split core, up to $40-50))

$26-40 per I/V channel total, maybe more due to the split-core CT cost for the large diameter power cable.

A conveniently programmable uC may cost you additional $20-30. I also used a $16 Ethernet/embedded TCP/IP miniboard from WizNET (http://www.saelig.com/miva/merchant.mvc?Screen=PROD&Product_Code=ETH042&Category_Code=BRD). That one is optional though because a uC usually has a serial port.

You will also need a voltage sensor, one per each channel or one for all the channels (if you are willing to sacrifice some precision). As a voltage sensor, you can use a 120->12/9/5 V wall transformer or a voltage divider. The former would be much safer as you'd achieve complete separation from the dangerous mains voltage, a DIY'er should not attempt to build a voltage divider for obvious reasons. The chip ADCs expect getting +/- 0.5 Vpeak, so the current/voltage channels should be suitably adjusted by choosing the CT load resistor and voltage-dividing the voltage transformer output.

The assembled gadget will need to be calibrated with a known load. It's a simple but tedious process which is fully described in the datasheet. I just used some resistive loads while measuring voltage/current with a multimeter.

oops..out of my league, back to finding something manufactured, but thanks for the info

 

[vc1234]

oops..out of my league, back to finding something manufactured, but thanks for the info

If you want a quick and dirty solution with minimum assembly involved, you can measure only current as mentioned earlier in the thread. You can get these sensors:

http://www.veris.com/file_uploads/h923_i0c.pdf

@ about $31 here:

http://www.plccenter.com/buy/VERIS/H923.

Then, you can connect the sensors to this (4 channel -10/+10V ADC module):

http://www.dataq.com/products/startkit/di194rs.htm ($25)

and connect the ADC module to your PC serial port.

The sensors convert 0-20A/100A/120A RMS selectable to 0-10 VDC which can then be read by a 194RS connected to your PC. There is a data acquisition app included and you can download an SDK to interface with your own app.

You can also save some but not much by buying a smaller/cheaper CT (the veris sensor is rather bulky but fits nicely over thick mains cables), a load resistor and building/using a bridge rectifier. That would give you current peaks (more or less) converted to DC voltage that can be fed to a 194RS.

The current sensing won't give you actual power but rather only an approximation to apparent power (VoltAmps rather than watts), neither will you be able to measure current "direction" since you do not measure voltage, but it's a good enough solutio in many cases...