I completed my monitor based on your solution. Thanks again. Site is http://www.thomaswlewis.com/?p=548.

It has been a few months so the details are a little fuzzy. But the 1 volt offset was related to the circuit not functioning correctly when the output was near zero. Sorry I can't remember the exact details.

I constructed this project for one leg only so far. I used the AD627 and left the gain at 5. I am still messing around with the calibration, but was able to modify the html to accommodate only the single leg.

Could you please explain why you biased the circuit to run from 1.0V to 5.0V, rather than from 0.0V to 5.0V? Is there some linearity issue with the op-amps or the IOBridge near 0V? Or some other reason?

Very fun project, thanks very much for posting it!!

Tom Lewis

As well as pubishing the live data set up alarms and maybe simple email/sms messages to indicate where energy management savings can be made.

This could then lead to controlling lights, HVAC and other devices automatically.

Once great job!

Mark

I found an adjustment pot inside my clamps and cranked them both up to their most sensitive/amplified setting. One clamp is *way* off and I had to increase the gain on that channel of the LM324 to 5.9 to get it to match the other clamp, whose gain is set to about 4.6. They at least read pretty much the same right now. The next time I place an order with Digikey, I’ll be sure to stock up on some better op-amps.

Are you going to use the AD737 with your next iteration? I’m curious to see how that turns out. It looks like the perfect solution.

I’m also on the look-out for some smaller clamp-on CTs. The off-the-shelf units I bought via eBay are nearly identical to yours, but don’t fit into my crowded electric panel…

That was an error in the schematic. The capacitor was supposed to be facing the other way. Thanks for pointing that out.

It’s hard to know exactly what is going on with your circuit. I tried to use the LM324 and didn’t have much luck either. I’m able to see 0.5A changes on my system. I’m actually in the middle of redesigning the whole thing to incorporate power factor. You may want to take a look at Analog Devices, AD737. It does what my circuit attempts to do, but is much much better at it.

Ryan

I’m a complete novice when it comes to electronics, so please bear with me. I’ve got a pair clamps similar to yours. I picked up an LM324 op-amp from Radio Shack (I know, I know) and, using 3k3 and 12k resistors, got a gain of 4.67, which should be enough to get 75A to register as 3.5v on my Arduino’s ADC. The LM324 will only amplify the signal to Vcc – 1.5v, and I’m running at 5v. So far so good. Is the op-amp’s gain a straight multiplier? Eg, is 10mV going to get amplified to 46.7mV with a gain of 4.67?

I’m having difficulty understanding the capacitor in your circuit. It seems to me that it’s “backwards”, with the positive lead tied to ground, and the negative to the 47k resistor and ADC input. Can you explain how that works? I understand that the resistor is “leaking” the cap, but I don’t understand how it charges (except perhaps from the grounded side of the clamp, but that’s very small voltage…).

What sort of resolution are you able to get? I bought a new DMM at Home Depot yesterday with an AC current-sensing clamp, so that I’d know if I was even close. Using a clothing iron and the incandescent lamp at my workbench, I’m able to measure about 0.5A with just the lamp, 12.1A with the iron, and (as expected) 12.6A with both turned on. My circuit, however, is not as responsive, however; it registers about 11A (scaling the ADC value up) which is sort of acceptable, but it doesn’t “see” the 0.5A change from the lamp, either with or without the iron on. Is that a limitation in the sensitivity of my cheap clamp, or have I buggered something up (likely)?

Thanks for the great write-up!
Brian

http://www.kondra.com/circuit/circuit3.html

This one is way expensive because of the board he is using. But it seems like the IOBridge might work with this kind of setup as well.

One question I have is if you have seen the Black & Decker Power Monitor system. It's $99 and does everything except the web feed. Considering your materials cost, have you considered hacking into the B&D device to web enable the data? The B&D device is a re-branded BlueLine Powercost Monitor.

Since I'm not so good with the electronics as you are, I would much rather take the Black & Decker device and just hack into it (potentially with the iobridge) to feed the data to my web widget.

If you have a small microprocessor with several reasonably speedy A/D converters, you can multiply the voltage by the current as measured many times during each power cycle. The sum of those products is the energy used, and is correct for whatever power factor is currently in effect–in fact, you can calculate the true power factor using those measurements.

You might be satisfied by using just one voltage measurement, but heavy loads on one phase can cause voltage drop in the neutral supply wire and thus shift the midpoint, so that the other phase’s voltage gets higher. So it’s more accurate to measure voltage and current on both phases.

Power factor is caused by the current waveform either leading (capacitive load) or lagging (inductive load) or in sync with (resistive load) the voltage waveform. Modern electronic power supplies or controls can cause the waveforms to be extremely distorted, so ignoring the details of current versus voltage throughout the cycle can give very misleading results.

The power company’s meter physically measures the components of current relative to voltage, so it compensates for power factors quite accurately. If you had a pure inductive or capacitive load, there could be large currents flowing, but for half the cycle the Volts * Amps would be adding and for the other half they would be subtracting, so the net power would be zero (ideally). The power company hates such loads, however, because the large currents (that you don’t pay for) do cause resistive losses in the distribution system (that you don’t pay for!), so they insist that large users correct their power factor (usually by using banks of capacitors).

This is a really neat project–congratulations!

http://route.dnsalias.org:83/AP/PowerMonitor.htm

I’m still working on the formula and graphing.

One thing I did that I’m not sure about. I tied both clamps together in series. When testing, this appears to work instead of using two inputs and adding them together. Do you see any drawback to this method?

Todd

Wait a minute…you are measuring the current twice, so your current reading will be twice too big, but since the voltage is twice (240), the wattage still comes out. Cool!

Cool project. My electrical panel is surface mounted meaning the line comes into it form outside and is thus surface mounted as well. Can I simply place the clamp on the line coming from the meter?

The power meters the utility company installs inside/outside your house, either the old mechanical rotating disk type or the newer solid state, measure active power. They do not need to know what’s in your house but they can measure/adjust for the power factor by sensing voltage and amperage and the phase shift between them (assuming the sinusoidal wave form).

They do not use “the same cheap CTs” as some other poster mentioned.

But all this being said, the project is truly nice and affordable. I like especially the web access part.

Thanks.