True power measurement

1. Digitize the signal. Do the math in a cpu.
2. Analog equivalent of the above.
3. Measure a property that is proportional to power, such as heat generation in a resistor.

Thanks. I was hoping someone would suggest #1. The real problem I am trying to solve is a 60 Hz. sinusoidal voltage and a highly non-sinusoidal current waveform. Several people have said "multiply the true RMS values...wrong. I concluded that the only solution is to do a full digital integration and I think you just suggested the same solution. I wonder if anyone has test instruments available to do this or do I have to build my own?

Shel

in article %1S6b.543$ snipped-for-privacy@newssvr23.news.prodigy.com, Shel Sherman at snipped-for-privacy@sbcglobal.net wrote on 9/7/03 7:44 PM:

Good old fashioned watt meters do exactly that. In the electrodynamic movement type, one coil has a current sample moving through it while the other coil has a voltage sample going through it. A torque is produced that turns one of the coils against a spring. Inertia of the movement smoothes out the positive and negative torque pulses produced. If you look at the scale, it is linear with power.

Harmonics, to first order, are no problem. Each frequency produces its own torque. At high enough frequencies, the impedances of the meter coils can affect the current samples through the coils.

Bill

YOur third suggestion is pretty close to an invention of mine.

We had to measure RF power and the load impedance was always 50 ohms.

What I did was use a very low powered flash light, bulb across a 52 ohm resistor.

Then the bulb was placed in one end of the tube. A photo sensitive resistor was placed in the other end so that it was about 1 inch from the bulb.

Next the scale was calibrated in watts with a dc voltage.

The bulb current was a very small portion of the total current and the bulb never really emmitted any light. The filament just sort of glowed.

Once the scale was calibrated now rf power could be applied to the watt meter and an accurate measurement of the RF wattage could be had. I use such a meter for years to check the output of CB radios and when compared to a standard,. it was quite accurate. . . I DO NOT FOLLOW MANY OF THESE NEWS GROUPS To answere me address mail to snipped-for-privacy@aol.com

If you don't mind something that only works at line voltage, get a Kill-A-Watt. It's about $50 or so online. It's got a custom chip that will do the calculation.

Bill: When you refer to an electrodynamic movement are you talking about the current rotating disk watthour meter or another type? I have been trying to gather data on the rotating disk type under nonlinear load conditions and would really appreciate any info or data sources you have.

dave y. schrieb:

Hello,

of course the voltage and the current must be digitized. The sample rate must be high enough for the non-sinusodial current properties.

Bye

That principal (keeping the resistor at a constant temperature by reducing the direct current to compensate for the RF has been around a LONG time.

I used one in 1962. I think it was called a bolometer or some such. It had a tendency to drift (it had a DC - or maybe a low frequency _loop that was closed by vacumn tube based amplifiers. Where measuring power in the mW range you had do zero, do a quick measurement, and re-check the zero.

"Way back" my brother picked up a WWII surplus "Antenna Current Meter" which used a small resistor in series with the antenna feed. A thermopile sensed the heat and its current drove a small meter.

instantaneous

Oops!! I'm sure you meant 'instantaneous voltage and instantaneous current' ;-)

daestrom

Yes indeed. Thanks.

--------- A standard wattmeter will do the job. You imply no appreciable harmonics in the voltage waveform so higher harmonic powers should be negligible so only the fundamental power is present (typically a wattmeter will not respond accurately above about 130Hz). However, these days digital wattmeters are cheaper.

-- Don Kelly snipped-for-privacy@peeshaw.ca remove the urine to answer >

in article 2G27b.11594$ snipped-for-privacy@newssvr27.news.prodigy.com, s falke at snipped-for-privacy@pacbell.net wrote on 9/8/03 10:06 AM:

I would have worded it somewhat differently: Instantaneous power (watts) is the product of instantaneous voltage and instantaneous voltage. That is averaged to get average power. I have trouble with the term *RMD power* if that indeed is what was meant.

Average power can also be calculated by summing up the products of rms voltage and rms current time cos(phase) for each harmonic present. Each harmonic contributes power as if the other harmonics are not there. Dynamometers do that calculation.

Bill

---------- An electrodynamic wattmeter is essentially a two coil instrument -a fixed current coil and a movable voltage coil constrained by a spring as in a conventional voltmeter. No iron, no disc. Weston, GE, Westinghouse and others made these. A good Weston was typically as 1/4 of 1 % of full scale from DC to 133Hz and even 40 years ago was about $1000. As for the harmonic sensitivity of disc type KWh meters, try and find some papers by Jerry Heydt who did work on this at Purdue at some time in the

-- Don Kelly snipped-for-privacy@peeshaw.ca remove the urine to answer

True

but my device was used to measure the output of cb radios which was in the 5 watt range and was nothing like a bolometer.

The only purpose of the dc was to calibrate the scale. The device was quite accurate down to about a half watt. . . I DO NOT FOLLOW MANY OF THESE NEWS GROUPS To answere me address mail to snipped-for-privacy@aol.com

Actually for ac it is not. It is the Root mean sqaure of each individual frequencies voltage times it current.

If the voltage is not a pure sinasoid then you can not just multiply the rms voltage times the rms current.

In my lab I have a Weston meter (I believe it is Weston, but it couldd be General) that has 4 input leads. One measures the voltage and another and is on a moveable coil. Another measures the current and is on another stationary coil. This gives the true RMS value of the wattage taking into account any distortions in the voltage or current.

The Meter movement does the RMS'ing of the current and voltage. If you feed in a voltage of one frequency and a current of another frequency the meter does not move. . . I DO NOT FOLLOW MANY OF THESE NEWS GROUPS To answere me address mail to snipped-for-privacy@aol.com

Almost forgot. It even takes phase angle into account. . . I DO NOT FOLLOW MANY OF THESE NEWS GROUPS To answere me address mail to snipped-for-privacy@aol.com

in article v587b.918868$ snipped-for-privacy@news2.calgary.shaw.ca, Don Kelly at snipped-for-privacy@peeshaw.ca wrote on 9/8/03 4:16 PM:

I got a good old fashioned wattmeter and a current transformer just for show. They look so elegant.

Bill

RMS voltage * RMS current, no. Instantaneous voltage * instantaneous current (integrating over the cycle), yes.

Integrating the product of the instantaneous voltage/current has become dirt-simple since the advent of the microcontroller. ;-)

Well, we were *not* talking RMS, we were talking 'instantaneous'. And if you sample it often enough, instantaneous voltage times the instantaneous current *is* the power. Even if it is not sinusoidal. If the current is not the same frequency as the voltage, the technique still is valid.

daestrom