Index Home About Blog
From: John De Armond
Subject: Re: I need a "watt meter" to measure consumption of my home 
Date: Wed, 26 Jul 2006 13:56:41 -0400
Message-ID: <>

The measured values (volts, amps) are accurate enough  but the
computed values (watts, PF, etc) can be way out in left field.

For example, my Pro shows my washing machine to have a PF of 0.20! I'm
not sure that's even possible.  My KAW(s) and my Fluke power analyzer
both show the PF to be 0.65 which is what I'd expect.

The Pro seems to match the other instruments down to a PF of about 0.6
or so.  Below that the indication is junk.

BTW, I have 4 KAWs.  I can string them together, one after another, to
a load and observe them all reading the same value to within a couple
of counts on the last digit.  Impressive.  Someone got the production
line trimming right.

I paid the extra money for the Watts Ups for the ability to ride
through power dips and to a lesser extent, the data logging ability.
They do indeed ride through power losses but the saved data is


On Wed, 26 Jul 2006 07:35:28 -0400, William P.N. Smith
<> wrote:

>"NeXTstep" <> wrote:
>>"Watts up? PRO" does not seems to be accurate
>What's inaccurate about the Watts Up Pro?

From: John De Armond
Subject: Re: House "Idle Speed"=300-400 Watts?
Date: Tue, 13 Mar 2007 22:09:27 -0400
Message-ID: <>

On Tue, 13 Mar 2007 19:57:56 -0400, "(PeteCresswell)" <x@y.Invalid>

>Being somewhat gadget-addicted as well as lazy, I got one of these too:
> which led me to start this thread.
>Seems to work as advertised - albeit the refresh rate is predictably slow bc it
>looks for the black spot on the meter disc, which doesn't come around all that
>Probably a good thing to check just before going away on a trip - same time as
>shutting off the water...
>Now that I know 300-400 watts might be a little high, I'll start spot-checking
>known devices with the Kill-A-Watt, keep a running total... and if it tops out
>at much less than 300-400 start looking for some hidden load - as in the
>doorbell xformer another poster mentioned.

Now that I have more information, a few more comments.  It probably
won't be a doorbell transformer.  Such a transformer will dissipate at
most a few (probably less than 5) watts.  Remind yourself that if the
3-400 watt load is real then a considerable amount of heat is being
dissipated somewhere.  Something dissipating even 20 or 30 watts will
be quite warm unless actively cooled with a fan.

Second, remind yourself that this gizmo uses the power meter as its
measuring element.  Though quite accurate, this low load is way down
on the bottom end of the meter's range and might not be terribly
accurate, particularly if the meter is fairly old.

Also remember that the meter wheel is not calibrated to be accurate
turn to turn.  The meter is calibrated to accumulated an accurate
amount of KWH at a wide variety of loads.  Especially at very low
speeds such as you're dealing with, the wheel speed can vary.  The
gear that rides the worm gear on the wheel shaft may have small
irregularities that vary the drag on the wheel from turn to turn.  It
all evens out over time but it will badly distort what this gizmo
reports.  This gizmo is putting the meter to a use that it was never
intended so don't be surprised if it isn't terribly accurate at the

There is some load or else the wheel would not be turning.  The
anti-creep feature (usually a hole drilled in the disc about half way
out from the center) prevents the disc from turning when there is no
load or an ever so slight (watt or two) load.  It's just that it may
not be nearly so large as the gadget is reporting.

I'll re-suggest that you open all the breakers, then close them in one
at a time and see what makes the meter turn.  When you find a branch
with a load, consider making up a Jesus cord to energize that branch
through the Kill-A-Watt.  A cord with a plug on one end and gator
clips on the other.  Make sure the breaker is open, then connect the
black lead to the breaker output terminal and plug the cord in through
the KAW.  That will energize the branch and allow you to see the
actual load.

My experience with checking the calibration of many (>10) KAWs against
a lab standard watt meter is that they are very accurate even at very
low loads.  It will accurately measure even just a watt or two.

This only works with 120 volt branches, of course.  For 240 volt
branches, the cheapest way to check that I can think of is to get a
very low range KWH meter such as this one:

Note that this meter's full load rating is 15 amps rather than the
more usual 200.  Also note that the Kh is only 2 rather than the more
usual 7.2 for 200 amp meters.  That means that a revolution of the
disc is only 2 watt-hours.

An even better meter for very low loads is the CT (current
transformer) meter.  This meter is designed for a maximum current of 5
amps and is designed to sit behind current transformers that step down
the main conductor current.  Typical are 200:5 and 400:5 CTs.  That
means that 200 amps or 400 amps through the main conductor produces 5
amps in the secondary.

A CT meter can be used without the CT for loads in the 5 to 10 amp
range (power meters are spec'd to operate within the specified
accuracy range for at least 200% continuous overload.)  A typical Kh
for a CT meter is 0.6 or 600 milli-watt-hours per revolution.  Hang
your gizmo on a CT meter and you'll have a really sensitive

Because any service larger than a typical residential 200 amp one is
CT metered, these meters are extremely common surplus.  Probably some
on sleazebay.  I have a whole bunch of these meters that I use for
energy surveys and I've never paid more than $30 for one.


From: John De Armond
Subject: Re: House "Idle Speed"=300-400 Watts?
Date: Thu, 15 Mar 2007 01:05:22 -0400
Message-ID: <>

On Wed, 14 Mar 2007 09:40:09 -0400, "(PeteCresswell)" <x@y.Invalid>

>The second spiel rang truer:  namely that this thing deals with the limitations
>of measuring disc rotation by having a minimum measured output of 400 watts.
>i.e. once the disc is rotation below a certain speed, it becomes futile to get
>an accurate measurement.

Yup.  Same problem as when I worked on the same sort of gadget back in
the 70s for that inventor I mentioned earlier.  In the
pre-microprocessor days I set up the hard-wired logic to count a clock
that was gated on and off by the disc black flag detector.  Some more
logic turned that period measurement into the KW display.  Probably
done the same way in software in that gadget.

I had to choose a clock frequency high enough that at the fastest
anticipated disc rotation rate, there would be at least a few clock
counts accumulated.  That means that at some low speed of rotation,
there will be sufficient counts that the counter overflows.  That's
the lower limit of measurement.  I used the counter "overflow" signal
to activate the "lamp test" feature of the display driver which made
it read "888888".  The instructions told the user that this display
indicated a load too low to measure.

Apparently these clowns decided to not be honest with the user and
instead "radio" in some bogus value.  Shame on them!  Given that it
contains a microprocessor, they could easily have programmed the thing
so that the user could set a minimum indication if they wished.

>He said that somebody measured a bunch of houses and decided that most houses
>draw about 400 watts even with everything nominally turned off.

That's a silly claim on their part.  That would be around 300 KWH
which at our still fairly cheap rate would be about $21 a month.  I
know people whose monthly bill runs under $50.  Yeah, small homes and
frugal usage but not that unusual.  If I recall correctly, the local
city utility says that the average bill is still around $60 a month.
Given that, if what this outfit claims is true then parasitic loads
would be from a third to a half of the total monthly load.  Rubbish!

>That makes it kind of useless for tracking down parasitic loads and/or checking
>for things being left on before leaving on a vacation.
>Sheesh!!!   They could have said something about that in the instruction
>manual... or, heaven forbid, on the web site to warn prospective buyers.

Appreciate you doing "research with your wallet".  I had almost
convinced myself to order one of these to see if it might have a use
on some of my portable watt-hour meters I use for energy surveys.
Obviously it doesn't.  You saved me some money.

>I did run around the house measuring various devices on the Kill-A-Watt.
>What I found was that none of my little power bricks consume a measurable amount
>when nothing is plugged into them.    OTOH, I've got a Sony tape/dvd player that
>sucks 6 watts when nominally turned off.   Ditto a Samsung player which pulls 3
>watts in the same state.    4 cordless phones in their charging bases pull 1
>watt each, a digital clock that draws 1 watt....  and so-on-and-so-forth....
>But all of those devices do not total up to anywhere near 100 watts - much less

Typical.  I bet your parasitic load is significantly less than 100
watts.  I think that it would still be interested if you rigged up a
Jesus cord to power the various branches through the KAW and see what
the parasitic loads really are.


From: John De Armond
Newsgroups: rec.outdoors.rv-travel
Subject: Possible inexpensive E-meter alternative
Date: Wed, 18 Apr 2007 00:14:59 -0400
Message-ID: <>

I mentioned almost a year ago buying a DC Watts Up

A DC volts, amps and watt meter.  Steve Wolf asked if it would be
suitable for use as an RV battery monitor.  Unfortunately not, as it
registers current in only one direction.

Tonight a friend sent me a link to this instrument:

About the same price ($60) but capable of bi-directional (charge and
discharge metering) measurements.  Add the $40 USB cable or go with
the Pro Basic package

And you have data logging.

I think I may have to buy one of these :-)  Meanwhile, if the
instrument performs as advertised then it should be a viable
alternative to the E-meter/Link-1000 at a fraction of the cost.


Subject: Re: Kil-a-watt usage / refrigerators
From: John De Armond
Date: Tue, 17 Jul 2007 16:36:54 -0400
Message-ID: <>

On Mon, 16 Jul 2007 21:51:44 -0400, "Mark Fields" <> wrote:

>Right now I am making an EXCEL spreadsheet to capture to populate with the
>date te Kil-a-watt was started, stopped, the readings, model no or
>appliance, and comments.  I've also got a column for the date the appliance
>was purchased.  Over time I home to keep reducing my electric usage.  In the
>past 2 months I have replaced 20 incandescents with CF bulbs.

Save yourself some effort and feel free to use my spreadsheet.

The spreadsheet is set up to be used with revenue meters as shown on that page but it
works just fine for KAWs.  Just enter the start value as "0" and the ending value as
whatever the KAW reads.  I keep track of the start and end time/date because I don't
entirely trust the KAW's internal clock.  I don't know whether the clock can be
affected by surge/sags and other glitches so I simply track the calendar time.

The major flaw in the KAW IS that it can be reset by even the briefest power
interruption.  I've modified one to add a backup battery to let the thing ride
through the interruption.  I'll try to get a page up on my website about this mod in
a bit.


From: John De Armond
Subject: Re: Kil-a-watt usage / refrigerators
Date: Thu, 19 Jul 2007 21:59:36 -0400
Message-ID: <>

On Thu, 19 Jul 2007 21:17:50 -0400, "Mark Fields" <> wrote:

>I like the web page and the measuring device.  I may try to build one.

If you can, locate a 100 amp/CT metering base.  It's round and just a little larger
in diameter than the meter itself. Much lighter and easier to handle than the large
base I show in the photos.  I got a bunch of those bases free which is why I used

One of my client utilities has a major retrofit underway to convert over completely
to self-reporting meters.  About 40,000 meters are involved.  I haven't found out
what they're going to do with the old meters yet but if we're lucky, they'll show up
on the surplus market.  Maybe I can get a few to pass on to others.

>The battery backup on the KAW is a great idea.  Your measuring device being
>mechanical has a built in memory so there is not any concern like this.

Yup.  I have the backup system worked out (two rechargeable Li button cells, a diode
and a NC pushbutton but I want to test it a bit before I publish.  In particular, I
want to make sure that fractional second dropouts (called "chugs") don't affect the
memory.  It shouldn't but one never knows.

There's a big unmarked surface mount chip in the KAW.  Given the cheap shunt, 5%
resistors and given how accurate every KAW is that I've tested, they almost have to
be calibrated on the production line.  That means loading in calibration factors.  I
bet there's an undocumented serial port on that chip.  Maybe I can find it.  If so
then there's a possibility that the KAW can be turned into the front end of a data

In this late model KAW that I'm working on there is also space on the board for a
power relay and driving circuits.  It's missing and the "contacts" are simply
jumpered on this one.  I wonder what that's for.  perhaps to turn the power off after
a designated amount has been used?  Anyone heard of any variations of the KAW with
this kind of functionality?

>My plan is to measure items for about a week and keep them in the
>spreadsheet I already had set up before seeing your post.

Those sound like handy sheets.  You ought to post them somewhere.  If you don't have
a website, send 'em to me and I'll host them for you.


From: John De Armond
Subject: Re: ping John, was: Kil-a-watt usage / refrigerators
Date: Fri, 20 Jul 2007 00:07:01 -0400
Message-ID: <>

On Fri, 20 Jul 2007 02:32:37 +0000 (UTC), danny burstein <> wrote:

>Have you tried using a KAW on the 240 circuit?

Not yet.  See my other post on this subject.

>I've attempted it here in NYC (yes, I was
>willing to risk it dieing) and I got some
>pretty wierd readings, including _120_ Hz.
>My electrical supply is based on 3-phase coming
>into this apartment complex, thus giving
>me a (nominal) 208V hot<->hot. (This 3-phase
>arrangement, with first apartment getting
>legs A and B (plus neutral and safety ground),
>second apartment getting B and C, third getting
>A and C, rinsel, lather, repeat... is pretty
>common in NYC. It's pretty rare in the rest
>of the country).

You mentioned that before and it STILL seems strange to me.  I *HATE* 208, especially
for heating appliances that are designed for 240.

One nice thing, though.  If you needed 3 phase power and had a friendly neighbor and
a little hole through the wall for a wire... :-)

>I periodically check friends' wiring looking
>for a true 240V, but haven't found one.
>Anyway... I suspect that the 120 hz is courtesy
>of the wierd wave form that two legs of
>a 3 phase circtui will give you. I haven't
>seen an actual oscilliscope wave form, but I
>suspect it would be two peaks (and a skip)
>in each cycle.
>Note that an EE friend of mine strongly disagrees,
>so I could be wrong...
>Oh, the KAW survived that 208V test fine.

The waveform is a simple sine wave.  That's easy enough to see if you draw it out and
manually sum enough instantaneous points on the waves to produce the difference of
the two.  It plots out to a nice sinewave.

I suspect that the double reading is resulting from the internal A/D converter
saturating for a period during each half cycle.  I can see how that might confuse any
of several math functions that they might employ to calculate frequency.

That doesn't bode well for me theory that a regular KAW can be used on 240.  Sure
wish I had an European one to dissect and compare to a US one.


Index Home About Blog