Analog dynamic range, accuracy and number of bits

Yes. Inductive loads lag. Capacitors are leading loads.

Standards are great because you have so many to pick from.

I have always avoided such. People who do "highly regulated" stuff spend most of their time on silly paperwork and in boring meetings and not designing electronics.

You can get TOU from a disk type meter with pickup electronics. The disk is just the power transducer. You do lose the ability to measure voltage, current, outage glitches, PF, stuff like that.

be

Good for you.

Do I have to have a history of designing power metering and control gear to challenge you? What difference does the history make if I know more than you about this? After all, even if I present my history, you will not accept it any more than I accept yours.

You brought up the confusion, not me.

Yes, phase angle matters. Is that what you wanted to see? Note that phase angle is undefined in many cases. Your case is special and phase angle is important. Good for you.

I just do not quite agree with you. Take solar installations for one case with a peak output twice that of the same time load.

?-)

don't

We don't regulate creativity, but we do document it. And our documentation and design standards are by mutual consent.

That is an approximation, it's really multiplication of the errors that matters. At 10%, the product (1+0.10) * (1+0.10) = 1+ 0.21 means that you can only get to twenty-one percent with two ten percent errors, if you calculate on a worst-case basis.

That's part of the reason complex error analysis usually is done with other-than-worst-case rules, using Gaussian distributed noise models. The error distribution for a worst-case specification is often assumed to be uniform, where the extreme values are as likely as the mean. That distribution in I and in V does NOT yield the same uniform distribution in I*V, however. The I*V is closer to a Gaussian distribution than to a uniform one. And, the central limit theorem says it'll be that way in LOTS of other cases.

Probably, you're correct. So, bring this to the attention of your energy regulatory agency; they DO consider such things. The power company will get better use of their information by leaving the meter as-is (because the transmission line costs are better represented if you separate them from metering), but if they have to consider the metering power their own cost, they'll make better efficiencies there.

The utility can claim that they are bound by their charter to meter the electricity, and metering as well as bill-printing/mailing are business expenses, not profit. They might be vulnerable, though, to the claim that motor-type meters are more energy efficient than the electronic ones (about 3W quiescent).

In a revenue meter, each meter is calibrated for "registration", namely accuracy in metering kilowatt-hours. If it also reports voltage and current, they can be individually calibrated too, or just allowed to seesaw as energy is calibrated, if they don't have tight requirements of their own.

The hard part is getting X per cent accuracy over a huge range of load currents. It becomes a "per cent of reading" problem, not the usual "per cent of full scale" measurement. So, especially, zero offsets become very important. Electric meters need a zero power deadband, so that a homeowner can shut off all his load breakers and observe exactly zero power consumption. So something like the low 10 or 15 watts is usually not metered.

The sampling multiply-and-integrate algorithm, done right, gives much better dynamic range for measuring power and energy than the square-filter-root thing provides for current and power. So PF measurement is usually bad at low power, where it doesn't matter much anyhow.

Disks don't turn at all when there's no customer load. There's a slot in the disk that provides a hump/detent effect, so the disk stops at low loads.

In a rotating meter, the big potential coil is powered all the time, and there is I^2R loss in the current coil at high loads.

Bull. I'm not the leader on documentation standards, and I certainly don't make the final decisions. Why would I want to do that, if other people are better at it, and willing?

The problem with being kow-towed to is that then you're always being asked to make desisions, and then you're responsible for everything that happens. No thanks: I'd rather design circuits, and be responsible for them.

I do insist that any product be fully documented, and reproducible and revisable indefinitely into the future. That includes tools, too, like PCB layout software, compilers, build scripts, things like that. That's not unreasonable, although sometimes it's not easy.

I don't think I'm missing any technical point. Pissing contest points I'm not interested in.

Cheers

Phil Hobbs

case

Just trying to let you see that the sign of PF has standard use to indicate direction of power flow. JL's sign convention is what confuses others, no mater how many sparkies he works with use it.

?-)

You

the

more.

Thus deflecting my point into the basket? You only resolved one of at least two items. Jerk.

?-)

I posted a link to a GE manual that explains both sign conventions. Their instrument can be selectd to use either.

Your power meters are only useful in one city in the *world*? The state of Maryland mandates that power meters have to measure power delivered back to the line and that amount is rebated to the customer... at least the generated portion of the cost. The other two thirds are either fees for transmission/distribution or are taxes and other fees. I'm pretty sure there are other states like this.

You have already decided on a convention for positive power. Are you saying you need further info on what negative power is?

What value is PF when the load is delivering power to the source? I guess you can't say since your meters don't figure this in.

Still not trimming your posts I see.

Rick

Please don't confuse me with some of the others here. I'm not in a pissing contest. The point is that using phi conveys the same information as P/VA. Why prefer one over the other?

Rick