# watt-hour meter

Discussion in 'Electrical Engineering' started by Sawney Beane, Nov 21, 2005.

1. ### Sawney BeaneGuest

For years I have used the watt-hour meter at the service entrance to
measure the rate of power consumption. I'd switch off all circuits
except a known load, such as incandescent bulbs or a water heater, then
time the on the meter disk to determine the constant in the relationship
between speed and watts.

Turning everything off requires resetting digital devices. If a year
later I wanted to check wattage and couldn't remember the constant, I'd
have to go through the procedure again. I couldn't be certain that my
test load was using exactly the watts I thought, and a constant found
for one watt-hour meter wouldn't necessarily apply to another.

Today I read that on most meters, Kh is the watt-hours per revolution.
(I wonder what Kh stands for.) Aha! You read the Kh, multiply by 3600,
and divide that constant by the seconds per revolution to get watts.

My Kh is marked 7.2. The time for 1/10 revolution was 41 seconds. That
came out to 63 Watts, right?

The house was running a computer in sleep mode, hooked up to an external
drive, an external modem, and a boom box. I switched that all off,
leaving only the power supply for the computer clock. There was a night
light. I switched it off. Otherwise, there was standby power for
furnace controller.

I timed the disk. Still 41 seconds. How could turning off my computer
equipment and the night light not have reduced power consumption? I
went out and timed it again. Still 41.

The third time, it had slowed way down. I don't know of anything in the
house that would have switched off since my last check.

Over the years I've seen that with other watt-hour meters. When a load
is turned on, there can be a long pause before the disk suddenly speeds
up. When the load is reduced nearly to zero, the disk may continue
spinning, then suddenly come nearly to a stop. What could account for
these delays in electromechanical watt-hour meters?

2. ### Charles PerryGuest

It is not really happening. The response of a watt-hour meter to changing
load is instantaneous (OK, there is a finite delay, but you won't measure it
with the naked eye).

Most watt-hour meter test equipment tests the meter by applying a known load
and measuring the time for the disk to rotate.....one time! If the meter
paused long enough for you to see it, it would fail the test.

The disk does not pause. You are seeing some other loads cycling. You used
to be able to buy a desk lamp with a watt-hour meter and an incandescent
light bulb. Turn on the bulb, meter spins immediately, turn off the bulb,
meter stops. No pause.

Charles Perry P.E.

3. ### daestromGuest

Like Charles said, your typical watt-hour meter doesn't have such a delay.
The inertia of the disk is so small that for all intents and purposes, they

Have you looked for some other cycling load? Perhaps a refrigerator?

daestrom

4. ### Sawney BeaneGuest

I was sure it wasn't the refrigerator. Yesterday I had time to
investigate further. I discovered a possibility: the air conditioner.
It has a heater for the outside unit. I wonder: is there any harm in
leaving it off for the winter? How long should it be allowed to warm up
before I run the compressor next summer?

With only three breakers on, I got 135 Watts. Oh-oh, I'd left a
fluorescent on in the bedroom, estimated at 100 Watts. I turned it off
and got 35 Watts. I went through the house looking for circuits that
were energized. All that seemed to be on were voltages for chips in the
answering machine, the microwave, and the furnace. I unplugged the
answering machine and the microwave. The meter disk didn't move in
several minutes. I plugged them back in. The disk still didn't move in
several minutes.

I guess those three devices didn't provide enough torque to get the disk
moving. So far I haven't seen it move slower than 35 Watts, and I
haven't been able to account for more than an estimated 15 watts.

5. ### Don KellyGuest

Your furnace likely has a small transformer to step down to supply the
thermostat and control relay. You also have a small transformer for your
doorbell. These are both cheap units -lossy at no load- and if you have an
electronic thermostat, there will be some more losses. This will chew up
some of your 15 watts. Did you unplug any electric clocks, etc?
Try opening your main breaker - shutting everything off- and see what
happens. --

Don Kelly @shawcross.ca
----------------------------

6. ### daestromGuest

The crankcase heater is important to prevent the refrigerant from dissolving
into the oil. If it does, when you start the unit the oil will foam up and
damage the compressor. I turn off the breaker in the fall each year and
turn it on at least 24 hours before running the unit. (Crane manual says
six hours, but what the heck...)

daestrom

7. ### Sawney BeaneGuest

My only outlet-powered clocks are in the range, microwave, computer, and
answering machine. The range breaker was off. I didn't try the
doorbell to see if it was on a live circuit. The thermostat is
mechanical, but the board in the furnace is digital.

The wheel could stop with the furnace, answering machine, microwave, and
perhaps doorbell energized, or it could indicate 35 Watts, apparently
with no additional load. I was unable to get it to turn between 0 and
35 Watts by disconnecting those small loads one by one.

If I knew more about shaded-pole motors, it might make sense to me.

8. ### ehsjrGuest

Sawney Beane wrote:
<snip>

Shouting ...
SHUT OFF THE MAIN BREAKER, AS DON KELLY SAID.

Then report your findings. What you *think* is connected,
and what is *actually* connected may differ. If the meter
does not stop with the main off, you have a serious
problem. The proper procdeure for shutting off the
main is to first shut off all the other breakers, then
the main. When restoring power, turn the main on first,
then all the other breakers, one at a time.

For branch circuit testing to find out which circuit is causing
the meter to run: (Shouting)
TO DISCONNECT LOADS THAT MAY BE MAKING YOUR METER RUN, SHUT OFF
THE BRANCH CIRCUIT BREAKERS.

Then report your findings. Unplugging some stuff and leaving
other stuff plugged in tells you precisely ... nothing. You end
up not knowing which "stuff" on which branch causes the meter to
run. Do it right and you'll get it right.
Ed

9. ### Sawney BeaneGuest

MAY I SHOUT, TOO?

I SHUT OFF THE MAIN SEVERAL DAYS AGO. NATURALLY, THE METER STOPPED.

AS I SAID, MY METER WILL ALSO STOP WHILE SUPPLYING POWER TO THE DC
SUPPLIES OF AT LEAST THREE HOUSEHOLD DEVICES.
AS I SAID, THE ONLY BRANCHES I RUN ARE THE ONES WITH THE THREE DEVICES
WHOSE COMBINED WATTAGE WON'T NECESSARILY TURN THE METER. SHOUTING IT
MAKES ME FEEL THAT I KNOW EVERYTHING THERE IS TO KNOW!
WATT WILL I GET RIGHT? WATT'S THE SMALLEST WHATAGE YOUR METER WILL REGISTER?

10. ### ehsjrGuest

Of course.
But you never told us that. So your shouting it now is unjustified.
But the information is welcomed.
No, you didn't. You said the meter did not turn after you
plugged them back in:
"All that seemed to be on were voltages for chips in the
answering machine, the microwave, and the furnace. I unplugged the
answering machine and the microwave. The meter disk didn't move in
several minutes. I plugged them back in. The disk still didn't move in
several minutes.

I guess those three devices didn't provide enough torque to get the disk
moving. So far I haven't seen it move slower than 35 Watts, and I
haven't been able to account for more than an estimated 15 watts."

So again, your shouting is not justified.

Since the distinction between unplugging something that had been
operating vs stopping the meter *by* unplugging escapes you:
A device can draw power periodically and store it, stay operational
off the stored energy for a period while not drawing any more, and
then draw it again when the stored energy level needs to be
and microwave I cannot say - but the possibility exists. Your
devices may have stored power while they were connected, and may
not have used it up in the several minutes they were disconnected
and the several minutes you monitored the meter after connecting them.
for help, but then resisted the answers.
Part of what puzzled you was why the meter was running when
you thought it shouldn't. (It seems that part may be solved
in your mind.) And part of what puzzles you is why the meter
doesn't move when you think it should.

Finally, I apologize for shouting - not because I shouted,
but because it apparently did not have the desired effect.
It was for emphasis, to try to get you to think about the
answers, rather than resisting them. All I see is that it
made you angry - not the intended effect.

Ed

11. ### PopGuest

....
: "All that seemed to be on were voltages for chips in the
: answering machine, the microwave, and the furnace. I unplugged
the
: answering machine and the microwave. The meter disk didn't
move in
: several minutes. I plugged them back in. The disk still
didn't move in
: several minutes.
:
: I guess those three devices didn't provide enough torque to get
the disk
: moving. So far I haven't seen it move slower than 35 Watts,
and I
: haven't been able to account for more than an estimated 15
watts."
....

Jeez, guys: a nightlight and such isn't going to show any meter
movement for a LONG time; probably days, certainly many hours!
Think about it! Certainly you won't see anything in 'several
minutes' as mentioned here! Oh, and the wheel will most likely
turn - done it with vacation home with night lights; 3 of them;
left on for a winter.

12. ### daestromGuest

Well, a 7 watt nightlight on a meter with a Kh=7.2, should rotate the disk
once in just under an hour. But I think there is a minimum friction where
it just won't register. I think it was Charles that had a link to
watt-hour-meter calibration spec.

Utility has to answer to the state's public service commission if too many
of their meters register with *no* load, while if they don't register with a
tiny load, they don't stand to lose very much (most utilities have a
flat-rate for the first few kwh just for having service, so they get paid
for those kwh even if the meter doesn't register them). So it wouldn't
surprise me if there is a minimum load before the mechanical ones actually
start to turn.

Wonder if the new electronic style are better? Seems an electronic
integrator setup could probably be more accurate at tiny loads.

daestrom

13. ### PopGuest

:
: : > ...
: > : "All that seemed to be on were voltages for chips in the
: > : answering machine, the microwave, and the furnace. I
unplugged
: > the
: > : answering machine and the microwave. The meter disk didn't
: > move in
: > : several minutes. I plugged them back in. The disk still
: > didn't move in
: > : several minutes.
: > :
: > : I guess those three devices didn't provide enough torque to
get
: > the disk
: > : moving. So far I haven't seen it move slower than 35
Watts,
: > and I
: > : haven't been able to account for more than an estimated 15
: > watts."
: > ...
: >
: > Jeez, guys: a nightlight and such isn't going to show any
meter
: > movement for a LONG time; probably days, certainly many
hours!
: > Think about it! Certainly you won't see anything in 'several
: > minutes' as mentioned here! Oh, and the wheel will most
likely
: > turn - done it with vacation home with night lights; 3 of
them;
: > left on for a winter.
: >
:
: Well, a 7 watt nightlight on a meter with a Kh=7.2, should
rotate the disk
: once in just under an hour. But I think there is a minimum
friction where
: it just won't register. I think it was Charles that had a link
to
: watt-hour-meter calibration spec.

Yeah, there does have to be a minimum friction, I'd have to
agree. I guess I was reacting more to waiting a "few minutes" or
whatever the phrase was.
Also, everyone here says, and others seem to agree with, the
fact that 1 revoution = 1 kWH but that's not true here in upstate
NY. I don't recall the numbers anymore, but when I was using the
spinning disk to measure power, my utility told me where to find
the "multiplier" on the meter face to convert the number of
revs/minute to kWH. I don't recall the number any longer, but
two numbers are stuck in my head: 6.4 and 7.2. I do know the
disk didn't indicate that one rev was a kWH though; it was far
from it.
They recently stuck us with one of the modem call-home types
so I can't support my allegations anymore. I really dislike the
new digital meters: There is zero way to guestimate ANY power
usage unless it's a very high current or a long period of time.
The lowest digit is a kWH; you'd think they'd have been down to
tenths, at least, but I guess since they charge by even hours,
they only have to display the even hour.
Oh! A little sidelight here: When they went down the road
here switching out the meters for the new digitals, you should
have heard the complaining about the power company using THEIR
phone lines to call in their meter readings, and how could they
stop it, and what else was it telling them about their houses?
<G>. They just couldn't fathom that it didn't use the phone
lines for that.
:
: Utility has to answer to the state's public service commission
if too many
: of their meters register with *no* load, while if they don't
register with a
: tiny load, they don't stand to lose very much (most utilities
have a
: flat-rate for the first few kwh just for having service, so
they get paid
: for those kwh even if the meter doesn't register them). So it
wouldn't
: surprise me if there is a minimum load before the mechanical
ones actually
: start to turn.
:
: Wonder if the new electronic style are better? Seems an
electronic
: integrator setup could probably be more accurate at tiny loads.

Interesting: I'll bet the new digitals don't have any trouble
seeing six or seven watts. I wonder ... oh well, not worth any
time or I'd go see what I could find out <g>. Reminds me of the
big scandals when the banks were caught rounding less than penny
billing fields to their favor. e.g. if you owed \$100.032, you
got charged \$100.04. And vice versa.

: daestrom
:
:

Cheers,

Pop

14. ### Sawney BeaneGuest

Do you remember when he posted it?
Did anyone in this thread say 1 revolution = 1 kWh? For decades, if
I wanted to use a watt-hour meter to measure watts, I had to time
the disk with a known load. At the start of this thread I said I'd
just learned where the watt-hours per revolution is printed on a
meter.
My disk turns 1000 times faster than yours: 1 revolution is 7.2 Wh.
My disk is divided into 100 marks. In 5 minutes, a 7.2W nightlight
should move the disk 8.3 marks.

15. ### Sawney BeaneGuest

I've found information on watt-hour meters. Speed is proportional
to power because magnets cause retarding torque proportional to
speed. Those magnets are adjusted to calibrate the meter at its

Then they adjust some sort of compensation to calibrate it at 10% of
its rated load. For me, that would be 720 W.

When that's done, it will read high at less than 10%. At 1% of its
rated load it may read 100% high. What do you know? All my
troubles have been below 72 W!

The disk has holes to stop it at very light loads. That way the
customer won't see the disk keep turning after he turns off his main
breaker, and the power company can keep claiming its meters are accurate.

That explains my two most recent obsevations: that a load of about
15 W reads 35, and that load won't start the disk moving.

So if I want to use the power-company meter to see how much power a
small device uses, I should first turn on several lights so the
total will be in the 700 W range.

16. ### PopGuest

: Pop wrote:
: >
message
: > : > :
: > :
: > : Well, a 7 watt nightlight on a meter with a Kh=7.2, should
: > rotate the disk
: > : once in just under an hour. But I think there is a minimum
: > friction where
: > : it just won't register. I think it was Charles that had a
: > to
: > : watt-hour-meter calibration spec.
:
: Do you remember when he posted it?
: >
: > Yeah, there does have to be a minimum friction, I'd have to
: > agree. I guess I was reacting more to waiting a "few minutes"
or
: > whatever the phrase was.
: > Also, everyone here says, and others seem to agree with,
the
: > fact that 1 revoution = 1 kWH
:
: Did anyone in this thread say 1 revolution = 1 kWh? For
: I wanted to use a watt-hour meter to measure watts, I had to
time
: the disk with a known load. At the start of this thread I said
I'd
: just learned where the watt-hours per revolution is printed on
a
: meter.
:
: > but that's not true here in upstate
: > NY. I don't recall the numbers anymore, but when I was using
the
: > spinning disk to measure power, my utility told me where to
find
: > the "multiplier" on the meter face to convert the number of
: > revs/minute to kWH. I don't recall the number any longer,
but
: > two numbers are stuck in my head: 6.4 and 7.2. I do know
the
: > disk didn't indicate that one rev was a kWH though; it was
far
: > from it.
:
: My disk turns 1000 times faster than yours: 1 revolution is 7.2
Wh.
: My disk is divided into 100 marks. In 5 minutes, a 7.2W
nightlight
: should move the disk 8.3 marks.

Interesting; and informative. Thanks for the education.

17. ### PopGuest

: daestrom wrote:
: >
: >
: > Utility has to answer to the state's public service
commission if too many
: > of their meters register with *no* load, while if they don't
register with a
: > tiny load, they don't stand to lose very much (most utilities
have a
: > flat-rate for the first few kwh just for having service, so
they get paid
: > for those kwh even if the meter doesn't register them). So
it wouldn't
: > surprise me if there is a minimum load before the mechanical
ones actually
: > start to turn.
: >
:
: I've found information on watt-hour meters. Speed is
proportional
: to power because magnets cause retarding torque proportional to
: speed. Those magnets are adjusted to calibrate the meter at
its
: rated load, typically 30 A.
:
: Then they adjust some sort of compensation to calibrate it at
10% of
: its rated load. For me, that would be 720 W.
:
: When that's done, it will read high at less than 10%. At 1% of
its
: rated load it may read 100% high. What do you know? All my
: troubles have been below 72 W!
:
: The disk has holes to stop it at very light loads. That way
the
: customer won't see the disk keep turning after he turns off his
main
: breaker, and the power company can keep claiming its meters are
accurate.
:
: That explains my two most recent obsevations: that a load of
: 15 W reads 35, and that load won't start the disk moving.
:
: So if I want to use the power-company meter to see how much
power a
: small device uses, I should first turn on several lights so the
: total will be in the 700 W range.

Interesting: Did you find that online and can you post a URL?

Pop

18. ### Sawney BeaneGuest

Here's where I found the explanation of inaccuracies:

of calibration (at 10% load) because the text turned to gibberish as
I scrolled.

Here's a page that says GE Watthour meters have the disk mounted by
magnetic suspension to avoid friction. (That wouldn't eliminate
worm-gear friction.)
http://www.ge.com/ph/indust_meter.htm

Here's a 72-page manual on checking and calibrating watthour meters:
http://www.usbr.gov/power/data/fist/fist3_10/vol3-10.pdf

Adjusting braking magnets affects speed by the same percentage at
all loads. When the magnets are adjusted to calibrate the meter at
full load,friction would cause the meter to run slow at light loads.

A short-circuited coil overcomes friction by increasing the effect
of the field from the potential coil on the disk. Its position can
be adjusted to adjust its effect. Its effect on disk speed is
inversely proportional to the load, so the coil has the effect of

If my meter was calibrated with a light load of 750 W, maybe the
torque adjustment to compensate for friction was equivalent to 20 W.
At 15 W, it would take only 0.4 W to add that same torque. So my

Disk creep can be up to 1 revolution per ten minutes and still meet
calibration standards. With a Kh of 7.2, that's 43.2 watts, or 1
kWh/day.

A disk has holes that serve as parking brakes. When a hole is
directly under the pole of the potential coil, torque will be
reduced. If the disk was moving very slowly, it may stop. With a
very light load, the disk should be timed for a complete revolution
because the holes will cause the speed to vary.

That helps explain the lag I've noticed over the years. I used to
live in a house where the meter and the breaker box were within
arm's reach of the back door, so I could see the meter very quickly
after switching breakers. If I switched everything off, the
light-load coil would keep the disk turning until a hole came into
position. If I switched on a 50-watt load, that might be just big
enough to get the disk turning, but it might take a minute or so to
move that first 1/4 of a mark, so that the hole was no longer
interfering with torque.

19. ### daestromGuest

Well, *I* didn't say it was 1kwh, and I see a couple of others that pointed
out that the meter's Kh number tells you how many watt-hours per rev of the
disk.
Yep, the number you want is shown as Kh=xx (mine used to be 7.2 before it
was replaced with a digital one).

Here in upstate NY in Nimo(oops, National Grid) country, they've been moving
to electronic meters. But the one I have has a small black 'dot' on the LCD
readout that pulses with each watt-hour, to show you it's 'turning'.

Look underneath the digital readout for a small LCD 'spot'. When you have
some loads running, it should blink on/off slowly, the speed tells you how
fast the older 'disk' would have turned.
Funny, our meters (national grid in the CNY division) don't use phone lines.
They are RF and the meter reader comes by once a month and simply drives
up/down the street. His unit polls each meter and gets kwh reading and
serial number while driving by at 10 mph. Still have a meter-reader, but
one can cover the territory that used to take about eight.

Are you *sure* yours uses a phone line? I know some commercial
installations in DTE country (lower Michigan) required a dedicated phone
line, but those were 'open choice' commercial customers, where the
non-utility supplier had to schedule power in 15 minute intervals and bill
accordingly.

Who's your utility, maybe you can take the model no. off the meter and
google it. That's what I did and found out all about it.

daestrom

20. ### PopGuest

:
: : >
message
: > : > :
: > : : > : > ...
: > : > : "All that seemed to be on were voltages for chips in
the
: > : > : answering machine, the microwave, and the furnace. I
: > unplugged
: > : > the
: > : > : answering machine and the microwave. The meter disk
didn't
: > : > move in
: > : > : several minutes. I plugged them back in. The disk
still
: > : > didn't move in
: > : > : several minutes.
: > : > :
: > : > : I guess those three devices didn't provide enough
torque to
: > get
: > : > the disk
: > : > : moving. So far I haven't seen it move slower than 35
: > Watts,
: > : > and I
: > : > : haven't been able to account for more than an estimated
15
: > : > watts."
: > : > ...
: > : >
: > : > Jeez, guys: a nightlight and such isn't going to show
any
: > meter
: > : > movement for a LONG time; probably days, certainly many
: > hours!
: > : > Think about it! Certainly you won't see anything in
'several
: > : > minutes' as mentioned here! Oh, and the wheel will most
: > likely
: > : > turn - done it with vacation home with night lights; 3 of
: > them;
: > : > left on for a winter.
: > : >
: > :
: > : Well, a 7 watt nightlight on a meter with a Kh=7.2, should
: > rotate the disk
: > : once in just under an hour. But I think there is a minimum
: > friction where
: > : it just won't register. I think it was Charles that had a
: > to
: > : watt-hour-meter calibration spec.
: >
: > Yeah, there does have to be a minimum friction, I'd have to
: > agree. I guess I was reacting more to waiting a "few minutes"
or
: > whatever the phrase was.
: > Also, everyone here says, and others seem to agree with,
the
: > fact that 1 revoution = 1 kWH but that's not true here in
upstate
: > NY.
:
: Well, *I* didn't say it was 1kwh, and I see a couple of others
that pointed
: out that the meter's Kh number tells you how many watt-hours
per rev of the
: disk.
:
: > I don't recall the numbers anymore, but when I was using the
: > spinning disk to measure power, my utility told me where to
find
: > the "multiplier" on the meter face to convert the number of
: > revs/minute to kWH. I don't recall the number any longer,
but
: > two numbers are stuck in my head: 6.4 and 7.2. I do know
the
: > disk didn't indicate that one rev was a kWH though; it was
far
: > from it.
:
: Yep, the number you want is shown as Kh=xx (mine used to be 7.2
before it
: was replaced with a digital one).
:
: > They recently stuck us with one of the modem call-home
types
: > so I can't support my allegations anymore. I really dislike
the
: > new digital meters: There is zero way to guestimate ANY
power
: > usage unless it's a very high current or a long period of
time.
:
:
: Here in upstate NY in Nimo(oops, National Grid) country,
they've been moving
: to electronic meters. But the one I have has a small black
'dot' on the LCD
: readout that pulses with each watt-hour, to show you it's
'turning'.
:
: Look underneath the digital readout for a small LCD 'spot'.
When you have
tells you how
: fast the older 'disk' would have turned.
:
: > The lowest digit is a kWH; you'd think they'd have been down
to
: > tenths, at least, but I guess since they charge by even
hours,
: > they only have to display the even hour.
: > Oh! A little sidelight here: When they went down the
: > here switching out the meters for the new digitals, you
should
: > have heard the complaining about the power company using
THEIR
: > phone lines to call in their meter readings, and how could
they
: > stop it, and what else was it telling them about their
houses?
: > <G>. They just couldn't fathom that it didn't use the phone
: > lines for that.
:
: Funny, our meters (national grid in the CNY division) don't use
phone lines.
: They are RF and the meter reader comes by once a month and
simply drives
: up/down the street. His unit polls each meter and gets kwh
: serial number while driving by at 10 mph. Still have a
: one can cover the territory that used to take about eight.
:
: Are you *sure* yours uses a phone line? I know some commercial
: installations in DTE country (lower Michigan) required a
dedicated phone
: line, but those were 'open choice' commercial customers, where
the
: non-utility supplier had to schedule power in 15 minute
intervals and bill
: accordingly.
:
: Who's your utility, maybe you can take the model no. off the
meter and
: google it. That's what I did and found out all about it.
:
: daestrom
:
:
No, ours doesn't use an actual phone line; sorry, didn't mean to
imply that. I had my tongue in my cheek there about how everyone
heard the "modem" word in an ad they left at each home, and
everyone with a computer decided it was going to use THEIR phone
lines. I was just commenting on the misconceptions of a couple
people talking to others. The ad never said it used phone lines
either, but clearly stated a "modem-like" device fed a small
transmitter and gave the FCC registration no., etc.. I threw it
away so I can't perfectly quote anything from it now.