I am looking to put together a circuit that measures power usage of mains devices (both 120V and/or 240V for here in the UK) but am struggling to find a current transducer that will give enough sensitivity at low currents (e.g. a few tens of mA for a phone charger) as well as full scale measurement of 13A. Has anyone here done anything similar, and if so could suggest a suitable device? Ideally I don't want to simply put a series resistor in line as I'm trying to keep this as efficient as possible!
Use a standard clamp-on ac current meter. For more sensitivity, wrap more turns thru the sensor. you can buy gizmos with a plug on one end, socket on the other and two or more openings to stick your clamp-on. Different numbers of turns == sensitivity multiplier.
Or spend $20 on a kill-a-watt...don't know if they like 240V. mike
A standard current transformer will work. It will be amplitude linear down to zero current, with maybe a couple of degrees of phase shift error at low currents.
Use a 12-bit or better ADC, with a couple of LSB's of random noise added to dither the steps; sample I and E close in time; multiply voltage:current sample pairs; average the result. A working dynamic range of, say, 20000:1 is practical if everything is done carefully.
If you sample E and I simultaneously, and multiply, and average the products, you get true power for any phase angle and, even better, for any waveforms.
If you want a wattmeter that handles 13A and 13 mA both with a percent or so accuracy, your ammeter part will have to be dual-scale. It's going to be a small series resistor (a copper wire, for instance) at 13A, and a much larger series resistor, in a current-transformer secondary, for the 13 mA.
The small sense resistor and current transformer primary are both in series with the load.
The trick is, your current-transformer will saturate (like, at 50 mA) so the series resistance on its secondary is no longer in-circuit when the power is high. Use transconductance multipliers to multiply the sense resistor voltage drop by the mains voltage, converting the product to frequency (volt/frequency converter), then use a microprocessor to count the pulses. There may have to be four volt/frequency converters, one for the (+) power/low current and one for (-) power/low current, third for the (+) power/high current and fourth for the (-) power/high current sections; that's because voltage/frequency conversion doesn't behave well if the input ever goes negative.
The microprocessor, each power cycle, must determine if the low-current section saturated, and use the high-current data for that time period, suitably scaled, if it did. If the low-current section didn't reach saturation current, its associated count is used instead, because it will be more accurate.
The 240V doesn't need similarly wide-range treatment, I trust. If there were significant DC current drawn, that would render the current-transformer circuit inaccurate, of course.
For extra credit, use multiple floating power supplies and op amp current mirrors, and generalize the scheme to five separate current ranges instead of just two...
You do not seem know or understand squat about current transformers. They are normally linear over 3 to 4 orders of magnitude, and can be really fast (sub-nanosecond pulses, wound low nanosecond 300 A ones myself). Specialized ones are useful to 6 orders of magnitude. You can buy ones that go up to 10,000 A Commercial-Off-The-Shelf (COTS). Try Ion Physics stuff or clamp on ammeters for examples. Electric service providers use them all the time for metering purposes. The rest of the measurement system is up to the designer.
Note, however, that the simple copper sense resistor also is linear over 4 orders of magnitude, and is much less expensive. My intent was to put a low-current sensor (the current transformer, of an inexpensive size and no great capacity) in series with a high current sensor so as to make a low-Z current sense array with two gain ranges.
Two resistors won't do it. A high-current transformer (my 800A example here weighs about 2 kg) is expensive. A low-current transformer, which drops out of the circuit when it saturates, seemed suitable.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.