Power - W, VA and power factor measurement.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I bought a power meter and it arrived two days ago.
Cut a hole in a plastic box to mount all the connections safely inside with a pair of cables for Plug top (power input) and socket outlet (to load).

AC 110V 220V Digital LED 100A watt power meter volt amp Ammeter Voltmeter W/V/A | eBay

The readouts are
xxx.x Vac, resolution to 0.1Vac
xx.xxx Aac, resolution to 1mAac
a choice of Watts, VA, or cos(phi) (power factor)

I plugged in my various bits of audio gear.
Quad FM tuner - - - - - - - - -0.02A, 6.5W, 7.2VA, 0.90cos
DCB1 vol pot+Buffer - - - - - 0.04A, 7.2W, 10.2VA, 0.72cos
STX 100W+100W 8+8ohms - 0.60A, 102W, 149VA, 0.69cos
NS 95W+95W 8+8ohms - - - 0.48A, 78.8W, 120VA, 0.66cos

Turning up the volume from silent to pretty loud made no difference to the power amplifier W/VA readout

I can see that the cosphi power factor for all my home built capacitor input filters have a range of 0.66 to 0.72. Pretty low
The Quad tuner is much closer to 1 @ 0.90 and yet it is also a transformer feeding a capacitor input filter.
I wonder what they have done differently?

Using 16p/kWhr and 18hours a day, that electricity bill comes to ~£200/annum.

Going back to the power factor of capacitor input filters.
The Quad uses quite small value capacitance feeding a 3pin regulator and the load is clearly very low.
My wondering may be explained by the difference in post rectifier smoothing capacitance.
Anyone tried to measure, or explain this?
 
Last edited:
A small reservoir cap means a long charging pulse which means better power factor. It also means high ripple, but that can be dealt with. It also means smaller charging pulses, which can ease grounding and magnetic induction problems.

It is reasonable to assume that Quad carefully designed their PSU using engineering principles, rather than audiophile recipes.
 
A small reservoir cap means a long charging pulse which means better power factor. It also means high ripple, but that can be dealt with. It also means smaller charging pulses, which can ease grounding and magnetic induction problems.

It is reasonable to assume that Quad carefully designed their PSU using engineering principles, rather than audiophile recipes.
I agree.

Time to experiment with the DCB1 smoothing capacitance and see whether power factor varies inversely with capacitance.

There is sufficient voltage drop across the DCB1 regulators to cope with much increased ripple voltage when small caps are substituted.
 
Andrew,
There are various definitions of power factor when taking into account nonlinear loads. Based on the measurements you provided, I would say this particular meter measures power factor NOT by the cosine of the angle between voltage and current - even though they use the cos(phi) notation. It appears the meter measures voltage mag, current mag, and watt mag. From there, all other measurements are calculations. VA is simply V*I. Power factor is watts/VA (which is a better measurement than cos(phi)).

If you looked at the phase angle between voltage and current on a fundamental basis, it would be nearly 0 degrees, resulting in a very high power factor. This is because the current is nearly zero most of the time, but then exhibits a high crest factor near the peak of the voltage wave. cos(phi) becomes 1.0. But this is with the traditional measurement of power factor based on angle only, which is meaningless with non-linear loads. Those meters that measure phase angle with comparators can be even more meaningless, as the current signal ends up being a pulse train rather than a square wave.

It would follow, however, that I wonder if the meter is measuring true-rms on current. Not too concerned about voltage, as the difference between AC and true-rms is minimal. But it is significant with current, given the harmonic footprint. So it is possible your measurement for current is incorrect on magnitude, and therefore VA would be wrong, and therefore power factor. Who knows how they are measuring watts, it isn't all that simple to do accurately.

So I would say your power factors either could be right or wrong, depending on measurement technique. It would be suggested that if you can measure current with either a shunt or clamp, input into your scope, you could determine if the meter is measuring AC or true-rms. That would be my first step. Makes it really easy if your scope has measurement functions, but not necessary. As long as you have a really peaky current wave you can get the answer.
 
My wondering may be explained by the difference in post rectifier smoothing capacitance.
The capacitance has a fairly small effect on the PF (but yes, less capacitance does mean better PF). The bigger player is the ratio of the source resistance of the transformer (and any other resistance that comes before the cap) to the load resistance. The lower traces on this graph would represent PF if you inverted them (Idc/Irms). You can see that when the source resistance of the transformer is much larger than the load resistance, the PF will approach unity.

A PF of 0.66 is exactly the sort of figure we expect for an off-the-shelf transformer operating at its full design rating, since it will usually have an internal resistance that is about 0.15 times the load resistance, giving a transformer regulation figure close to 15% which is typical. Looking at the graph, Rs/Rl = 0.15 shows the RMS current in the transformer secondry will be about 1.5 times greater than the DC load current, which is a PF of 1/1.5 = 0.66. (Primary magnetising current also worsens PF, but this should be a fairly small contributor when under load)
psu3.jpg
 
Last edited:
zigzagflux said:
There are various definitions of power factor when taking into account nonlinear loads.
I thought there was only one definition of power factor but (like RMS voltage, for example) there are various ways of estimating it from measurements of something which is not power factor. These estimates may be good or bad, depending on the waveforms (just as RMS voltage 'measured' via peak voltage is only good for a sine wave). So the question is: what exactly does the meter measure?
 
Hello,

I bought exactly same power meter some months ago to install it in an isolated variac to replace the galvanometers.
I made on it some modification for power it with 5V supply allowing to measure voltage down to zero
(otherwise it start to ~80V because it is self powered by line voltage you measure).
I also replace the voltage measurement circuitery by a differential amplifier to improve accuracy.

Then, i compared it with a Zimmer LMG95 precision power-meter and i can confirm that it probably
really compute the PF with sampled signal, the error is very low for all various load i had tested.

Frex
 
I thought there was only one definition of power factor but (like RMS voltage, for example) there are various ways of estimating it from measurements of something which is not power factor. These estimates may be good or bad, depending on the waveforms (just as RMS voltage 'measured' via peak voltage is only good for a sine wave).

This might not answer the question, but here goes from IEEE references.
The conventional definition of power factor was the cosine of the angle between voltage and current. At the time of its definition, it was understood these quantities were fundamental (50Hz) only. With this, it was correct and true also that watts/VA would produce the exact same power factor quantity, so choose your method of measurement. Conceptually, the watt/VA method carries a degree of comprehension in how 'effective' or 'efficient' the particular circuit was. This is important.

With the increasing harmonic distortion environment, this conceptualization had to change to consider the signal's rms values, which takes into account the fundamental and its harmonics. So power factor was redefined into Displacement Power Factor as being the cosine of the angle of the fundamental quantities only.

True Power Factor was introduced (sometimes called Distortion PF, but doesn't help in having the D prefix), which takes into account all harmonics, being defined as Prms/VArms. Note how there is now a separation between the concept of watt/VA (which really carries the most meaning) and the angle measurement. Angle remains as a fundamental-only quantity, and True Power Factor all quantities.

If dealing with a harmonic-free environment, DPF=TPF. If harmonics are present, they are necessarily different.

Not really used much, PFtotal=DPF*TPF. I think it was intended to indicate that the only way to have PFtotal=1 was to have a harmonic-free resistive circuit. As soon as you introduce harmonics, it will always be less than one, even if you still have a resistive circuit (i.e. distorted voltage serving a resistor). Don't see a ton of value in using it, but it's defined.

Note that one can 'improve power factor' by adding capacitors to a system, but this will only change the Displacement Power Factor. If you want to improve TPF (also called Distortion PF above), you have to use harmonic filters. Very significant in my line of work, hence why I bother to bring it up.

Further support in IEEE Std 141-8.2

So the question is: what exactly does the meter measure?
I think that question is most important. If we don't understand how it measures, how can we assign any value to its meaning? If the measurement is based on the fundamental angle between voltage and current, then power factor in this application is meaningless. Why bother concerning yourself about the differences in various pieces of equipment when you can't determine what the measurement means. There is no question the current has significant harmonics when feeding a cap input filter. One might as well ignore the power factor entirely if the meter provides DPF.
 

PRR

Member
Joined 2003
Paid Member
I bought a power meter.......

Being in 120 land, I bought two similar units, and wired them into my cellar fusebox.

(Lurkers: do not try this at home! Do it at someone else's house!)

No "power factor" (the traditional meaning of large shifty motors has been clouded by distortion; and I hardly care).

Knowing my voltage and load real-time has been moderately useful.

For your purposes, a "Kill A Watt" monitor is a pre-made and useful tool. I have used that to estimate the duty-cycle of a refrigerator. I have also cut a cord so I could use my Fluke to get the starting-current of the reefer.

FWIW: the two I bought, one had a flaky back-light. I bullied it into working, but 2 years later it is out again. I have got 2 more of the "same"(?) unit, but not yet replaced the dim one.
 

Attachments

  • Monitor-assem.jpg
    Monitor-assem.jpg
    148.7 KB · Views: 123
  • Monitor-live.jpg
    Monitor-live.jpg
    78.7 KB · Views: 121
Member
Joined 2004
Paid Member
There are a few audiophile products that claim to improve power factor (and sound) by adding big caps across the line. I measured one that had a reactive current of around 3A. In effect reducing the max available current and heating up all the connections in the loop. It should have a small impact on the Watt-hours since its reactive but also has little contribution to the power quality or the sound.

Here is a real passive power factor corrector for rectified loads: Patent US5251120 - Harmonic noise isolation and power factor correction network - Google Patents Not trivial to implement but it does work as advertised.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.