This is not my understanding. Most wattmeters that are implemented as you describe require the wattmeter to be the entire target load. Otherwise, there is no error-proof way to know how much power is dissipated in the wattmeter in relation to that which is dissipated in the device.
Typical wattmeter instruments have a similar deficiency as was ascribed as an error by McKubre - I.E. taking the average of the voltage and the average of the current and multiplying them to get the power at a sample time. These wattmeters have a sample period, during which the voltage is averaged and the power is averaged as an input to the power and RMS calculations. Good power meters will have a very short sample period. The PCE 830 used by the Lugano authors had a sample period of probably about 100 microseconds. Only frequencies above about 5 kHz will cause an error in power measurement for such an instrument because higher frequencies than this get averaged out during the sample period causing an error. On the other hand, many high resolution DAQ systems sample the current and voltage with a sample period of 1-3 seconds. So AC signals riding on the input of frequencies greater than about 1 Hz will cause an error in assessing the power. If these AC signals are of small amplitude, the error will be small in amplitude. That's why researchers are admonished to check their waveforms with a digital scope to look for AC signal components impressed on their otherwise presumed DC sources and assess the impact of the AC on the measurement of power. Bob Higgins On Mon, Oct 27, 2014 at 1:37 PM, Jed Rothwell <[email protected]> wrote: > > Well, a watt meter, not a volt meter. Yes, you can, and the instrument > makers do. All high-quality, high-powered wattmeters use calorimetry. That > is to say the heat up a resistor wired in series with the load, measure the > temperature and convert that to power. This method eliminates any > possibility of exotic waveforms or extremely rapid changes in electric > power being missed by the instrument. This method detects every joule of > electricity, no matter what. > > - Jed > >
