DIY mains power consumption meter

It already runs a good deal faster than that...

Erm, that is pretty much exactly what it does. (it also low pass filters out the DC offset introduced to make bipolar waveform palatable to the ADC)

erm, again yes...

and the rest - it basically runs "flat out" during the sample window. I will time it later and see what it can manage on that platform.

What do you think I was using when I was testing it?

You may find this helpful:

But your mains 'test lead' is single-insulated.

Owain

Yeah, I like to live dangerously... :-)

Out of curiosity, I instrumented the EmonLib code to toggle the built in LED at the start of each iteration of the main sample loop. Here is what I got running on a nano with a Atmega 328p at 16MHz :

IMHO quite impressive for an 8 bit RISC CPU with no FPU...

Two ADC conversions, approx 16 double precision floating point operations, plus some integer maths, logic and branching

This is starting to feel a bit too much like my day job...

Ta!

That sounds like it was designed in Spain then. I've not forgotten the pvc tape used to insulate wires joined together under a bed to feed a reading light. Brian

Since you have not seen the photo, it is basically an extension lead about 2m long with a plug one end and a trailing double socket at the other. The mid section of 1.5mm^2 flex has its outer sheath removed. This enables single cores (or multiple turns of) to be clamped round for current test purposes. Not intended for normal use obviously!

A question which this interesting thread has reminded me of: I have one of the 'plug in 'piggyback' power meters, with volts and wattage measurements etc. Presumably using a similar current transformer sensing mechanism.

Any thoughts of the accuracy of such beasts?

Thanks, J^n

The one I had (since deceased) seemed to be ok. It could certainly cope with traditional (i.e. leading or lagging) current profiles. I am not sure how good it was on ones with lots of harmonic distortion on the current profile. While many probably use a similar design, it is quite possible that some are much better than others.

They typically don't use a current transformer, they use a small value series resistor and measure the voltage drop. More accurate because there's no worries about the nature of the inductive coupling of the current-carrying wire, and probably less susceptible to stray magnetic fields.

Theo

Yes, I was looking at Big Clive's tear-down of his 'Hopi' Power Meter.

I presume the output of a CT is to some degree dependent on the quality of the mating of the ferrite surfaces?

You can get nice little potted modules for that, eg:

I've not used one, but they would be my go-to solution in a project where I wanted a mains input and a USB port or wallwart wouldn't do.

Theo

Surely Clive must have looked at that ?

He seemed to like this one:

Yup, they want to be clean and preferably well aligned. (there is a small amount of positional misalignment allowed on the one I have. Having said that, fiddling with it mechanically does not seem to change the reading.