Power Meter accuracy

Household mechanical meters with max power of 24 KW will usually resolve about 15 watts on the low end. They have magnetic detents (a hole or slot in the disk) that stop them from rotating at low power, so that people don't get upset by seeing the disk turn at zero load.

There are commercial power metering chips from Analog Devices and maybe TI. Lately people just digitize current and voltage samples and do the math. It's not easy to do an electronic meter that approaches the performance of a rotating disk meter, much less the reliability, ruggedness, and cost.

I've worked with four companies who tried to do electronic meters. Three are defunct... Metricom, Synergistic, and the Niagra Mohawk fiasco. I hope it wasn't my fault.

John

Well, the Kill-a-Watt "meter" fails to register a (linear, unregulated) wall-wart powered light that takes about 4 watts.

It no doubt has a software cutoff, to keep zero drifts from indicating power when there's no load. It's hard to keep electronic power meters from having small offsets.

John

Most home meters are very inaccurate, really designed for measuring big power (100s of watts and KW) and really no use at anything below 10w due to noise and drift.

I have 2 mains plug power meters, 1 registers 4w with a 1w source but has at least 1 decimal place, and 12w on a 5w source. The other one has no decimal places but seems to be accurate to the display level and measures down to 1w.

I've similar problems with industrial equipment, but you can at least get a good datasheet specification with that. The datasheets I have for these home units are plainly incorrect for at least one of them.

You can get a good professional meter for about £400 specifically with low power directive in mind. Powertek ISW8001, there might be a cheaper version.

-- Tony

Whether the meter uses a current transformer or a shunt, it will be somewhat sensitive to external line-frequency magnetic fields. Try plugging in a transformer-type wall wart next to your power meters; it would be interesting to see if there's any effect. Just a square cm of pcb trace loop area can be a problm here, or any asymmetry in a CT core or winding.

John

On a sunny day (Mon, 10 May 2010 17:07:18 +0100) it happened Tony wrote in :

It is lonly logical. If they use a 10 bits ADC, then for a max of 2 kW the miniumum is already 2 W. Add any offset in the current sensing, or even one bit eror, and you are at a minium of 4 W etc etc.

As 4 W over say 24 x 365 = < 8 kWh, x 25 cent = < 2$, you would need 200 years to break even on that more expensive meter. Not me :-)

It's usual to dither the current signal before digitizing it. That washes out the ADC quantization error. There are more subtle problems on the low end, like ground loops, mag field pickup, ADC mux hangover, and bad code.

John

Tim

John

So they have an analog multiplier first?

Tim

Did you not hear the whoosh as it flew over their heads? :)

Grant.

message=20

How much variation in line voltage do you see before it impacts performance of lamps, motors, and most other loads; 10%?

Sorry, I was talking about power meters that measure power for the utilities companies.

I know TI AD and others have power meter reference designs.

Regs Joe

On a sunny day (Mon, 10 May 2010 17:04:25 -0500) it happened "Tim Williams" wrote in :

A 10 bit ADC has 1024 steps

Hello?

Presumably, they're converting voltage and current seperately.

Tim

Gee, I think I just embarassed John into not replying. I must be getting

Tim

"Classic" meters are nominally ~1% devices. Doesn't sound very impressive until you think about the wide operating range (environmental as well as electrical) that they have to operate in as well as their long term reliability

Modern "solid state" devices usually start at 1% and get better. E.g., there are 0.5% class devices, 0.2% class, etc.

Of course, the SS devices place some "burden" (a few W) on the line this has to be accounted for (either in the design or in the billing).

KWH meters are amusing devices. More effort is expended on all sorts of insane details (register ratio, watthour constant, etc.) that belong in a lexicon specific to just KWH meters. And, lots of "rule of thumb" stuff that often seems arbitrary to an "outsider" :-/

Keep in mind the types of folks who use/install these things. They aren't typically concerned with the same sorts of metrics that engineers would...

I have no idea what either of you are talking about.

Is the 2mW* thing some sort of joke? Please explain it.

John

How much does the line voltage change where you live?

John