💾 Archived View for tilde.pink › ~ssb22 › ct.gmi captured on 2024-02-05 at 10:26:39. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2023-01-29)
-=-=-=-=-=-=-
From around 2007 to 2010 several UK electricity boards and other organisations provided CT-based “energy monitors” with various names that all ran essentially the same program from Current Cost (HK) Ltd on a PIC18F85J90 microcontroller.
A limitation that must be understood if you’re trying to measure small loads is this: any monitor that relies only on putting a “clamp” around a mains wire to measure its current, must *assume* that the alternating voltage is in-phase with the alternating current (i.e. power factor 1). This is correct for high-power resistive loads like heaters, but loads that are more “reactive” (inductive or capacitive) will cause a phase-shift between the voltage and the current (so peak current is *not* at the time of peak voltage), and CT-based monitors (which assume this current peak *is* at the time of peak voltage) might calculate an apparent power consumption many times higher than the real one.
For example, some (but not all) types of mains-connected smoke alarm use a simple capacitive voltage divider (two capacitors connected in series) to get a low-voltage power supply. As capacitors charge fastest when they’re empty, the current could lead the voltage by almost a quarter-cycle, causing the CT monitor to get it wrong by a factor of 10 or so.
(A less-serious inaccuracy is, the monitor might be 10% wrong if it can’t account for normal voltage fluctuations on the line. In theory a more advanced monitor could measure voltage fluctuations through its own power supply, but there’d be no *point* if it’s using a battery-powered transmitter with only enough power to transmit an average current every so often—without precise high-resolution time-correlated values of both voltage and current, it can’t correct the larger error caused by the bad power factor of some small loads.)
The real power consumption is required to be reasonably-accurately measured by the billing meter, so you’ll have to read *that* to be sure.
Nevertheless a CT-based monitor (if you can find one that still works) could be used if the actual meter is not conveniently accessible (e.g. if it’s in an outside cabinet and the weather is bad), and should give a good reading for high-load items like cooking and heating.
The “Current Cost” monitors were often supplied without instructions or with only vague ones, so here are some notes:
All material © Silas S. Brown unless otherwise stated. Ethernet was a trademark of Xerox Corporation but it was relinquished in 1982. Any other trademarks I mentioned without realising are trademarks of their respective holders.