Power Measurement Basics

[poynt99]

Poynt99 we are in violent agreement that following a consistent convention is essential.  The convention that I have always seen with respect to power is with reference to the loads.  I have never seen a load referred to as a negative power source.

I think what you are trying to say is that in your work, you are never interested in measuring the input or source power; you are only interested in measuring the power dissipated in a circuit's loads, correct? I have never suggested that loads compute to a negative power, they don't. However, power sources DO, if measured correctly.

To establish what is needed to insure we don't mess up the phasing, take for example a sinusoidal source with an offset voltage capability.  Our load will be a simple 1 ohm resistor, and we will place in series a 1 mOhm current sense.  The current sense is low-side and we assign node 0, and the connection points for all of our instrument references to the low side of the power source.  Channel 1 measures voltage from reference to the top of the circuit where the source connects to the load.  Channel 2 measures voltage from the reference to the junction of the CSR and the load.

Set the source to 2V p-p with zero offset.
Channel 1 shows 2V p-p in phase with the power source.
Channel 2 shows ~2mVp-p in phase with the power source.
Measured power magnitude at each peak is:  2mV * 1000A/V * 2V = 4W, and -2mV * 1000A/V * -2V = 4W.

Offset the source positive by one Volt:
Channel 1 swings from -1V to +3V in phase with the power source.
Channel 2 swings from -1mV to +3mV in phase with the power source.
Measured power magnitude at each peak is:  3mV * 1000A/V * 3V = 9W, and -1mV * 1000A/V * -1V = 1W.

Offset the source negative by one Volt:
Channel 1 swings from -3V to +1V in phase with the power source.
Channel 2 swings from -3mV to +1mV in phase with the power source.
Measured power magnitude at each peak is:  1mV * 1000A/V * 1V = 1W, and -3mV * 1000A/V * -3V = 9W.

Where you have to watch yourself is when you use the trick of moving the reference node so that it is at the junction of the CSR and the load, rather than the CSR and the power source.

I'm not sure I understand your point. Have I shown the reference between the CSR and load?

In your scenario above, you are measuring voltage directly across the voltage source, and yes due to the probe configuration your voltage and current traces will be in-phase.  However, your probes are placed in a series-opposing configuration, which means that one channel signal is inverted wrt the other.

So since you are measuring directly across the voltage source, you are actually measuring the source power, not the load power (even though they are close to the same in magnitude). When you invert one channel in the scope to correct for the series-opposing probe configuration, your power will compute to be negative, as it should be for where you are measuring, i.e. source power.

In order to properly measure the 1 Ohm load power, you must move your voltage probes either directly across the load resistor (differential probe), or from the source to the reference, which is the connection we had previously. Now your probes are either in series-adding (differential probe scenario), or they are in a "fixed reference configuration" whereby the CSR drop must be compensated for accuracy. In either of these last two cases, there is no need to invert one channel because the probes are physically "in phase".

How would you go about measuring the power in each of several loads that were in series? What if you were only interested in the total power being used, how would you measure that?

Keep in mind that in these forums, it is important that folks understand how to measure battery/source power and load power, because often the load they are interested in may not be the only component dissipating power in their circuit. Comparison between "input" and "output" power then becomes very relevant.

And there's this:
http://www.overunity.com/14220/power-measurement-basics/msg383962/#msg383962

[MarkE]

Yes, absolutely.

I hope there will be some time for questions and answers.

For example, a proper Current Probe (Hall effect - transformer type) matched to the oscilloscope clips around a circuit wire at the measurement point and usually needs no ground reference connection to the circuit. The probe body is generally marked with the correct orientation wrt conventional current flow. How does the signal from a probe like this compare in polarity/phase with a reading of voltage drop from an inline CSR at the same location?

Another "poynt" or demonstration that might be nice would be an explanation of the use of differential voltage probes in situations like this one, and also how two passive probes can be used in place of one differential probe to measure signals between arbitrary points in a circuit.

I've attached an oscilloscope capture using a 100 Ohm resistor driven by a function generator and a Tektronix P6021 transformer current probe.  I have diagrammed the set-up including with the marking orientation of the P6021.  The P6021 output is in phase with the function generator voltage.  This is consistent with current flow in the diagrams that both Poynt99 and I have posted.  The issue is what convention to follow for the voltage.  My experience is that the applied voltage across a load is always used.  IE it is the voltage that with an in-phase current, results in power dissipation.

[MarkE]

In your first diagram, how did you come to the conclusion that the battery power is -10W and the resistor is +10W?

The battery disssipates -10W.  That is the same as saying that it supplies +10W.

[poynt99]

The battery disssipates -10W.  That is the same as saying that it supplies +10W.

You've re-stated basically what the diagram is depicting, but how did you come to the conclusion that the battery dissipates -10W, or the resistor +10W?

Why did you not conclude that the battery was dissipating +10W and the resisitor -10W?

[MarkE]

I think what you are trying to say is that in your work, you are never interested in measuring the input or source power; you are only interested in measuring the power dissipated in a circuit's loads, correct? I have never suggested that loads compute to a negative power, they don't. However, power sources DO, if measured correctly.
I'm not sure I understand your point. Have I shown the reference between the CSR and load?

In your scenario above, you are measuring voltage directly across the voltage source, and yes due to the probe configuration your voltage and current traces will be in-phase.  However, your probes are placed in a series-opposing configuration, which means that one channel signal is inverted wrt the other.

They do not oppose.  Both probes are in phase.  In order to oppose one must be CCW and the other CW from the reference node.

So since you are measuring directly across the voltage source, you are actually measuring the source power, not the load power (even though they are close to the same in magnitude). When you invert one channel in the scope to correct for the series-opposing probe configuration, your power will compute to be negative, as it should be for where you are measuring, i.e. source power.

You have declared that I am measuring across the source and then introduced loss that is not shown in the diagram.  As the diagram is shown there is no distinction between the voltage across the source and the load.  If we introduce a few milliOhms of wiring resistance, then my probes would be on the load side of that wiring resistance.  The magnitudes would then change.  The signs would remain the same.

The issue entirely revolves around whether we declare positive power that which the source supplies, in which case dissipated power is negative, or do we declare positive power that which the loads dissipate, in which case sources "dissipate" negative power.  Industry convention is the latter.  Power meters indicate positive power as the power into the load.  No one refers to the power that a light bulb or any other kind of load dissipates as negative.  If we were to split the 1 Ohm load into two 0.5 Ohm resistors and probe the junction, we would note basically half the source as the positive voltage drop across the bottom 0.5 Ohm resistor.  And we would still report the same current.

In order to properly measure the 1 Ohm load power, you must move your voltage probes either directly across the load resistor (differential probe), or from the source to the reference, which is the connection we had previously.

You are mixing up multiple issues.  The first is what is the convention for positive versus negative power.  The second issue is one of instrumentation options of which there are many for anyone with enough budget, and fewer for those who don't.  It is critical that the convention is settled before worrying about small instrumentation errors.  The convention issue does not change whether we instrument with some piece of junk +/-20% accurate instrument or something good to six digits that is fully isolated and has an infinite CMRR.

Now your probes are either in series-adding (differential probe scenario), or they are in a "fixed reference configuration" whereby the CSR drop must be compensated for accuracy.

I described where the probes are.  They both use the same node 0 reference.  Yes, that introduces a miniscule voltage magnitude error if left uncorrected, but no it has nothing, absolutely nothing to do with the selected power convention.

In either of these last two cases, there is no need to invert one channel because the probes are physically "in phase".

How would you go about measuring the power in each of several loads that were in series? What if you were only interested in the total power being used, how would you measure that?

Dividing a branch into series pieces does not change the convention or methods.

Keep in mind that in these forums, it is important that folks understand how to measure battery/source power and load power, because often the load they are interested in may not be the only component dissipating power in their circuit. Comparison between "input" and "output" power then becomes very relevant.

Which is the reason that I object to your choice of positive power as that supplied by a source when the common convention for positive power is the quantity dissipated by loads.  If one is intent on educating folks, which is a good thing, teaching them to go against accepted conventions is a recipe for confusion and dissent.

And there's this:
http://www.overunity.com/14220/power-measurement-basics/msg383962/#msg383962