Are RMS multimeters a con?

[DAIRVINE]

I have an application for measuring phase angle on an electrical supply by measuring voltage, voltage across a resistor for current and the vector sum of both. Use of trigonometry derives phase angle. But you need to use an averaging meter for good results. When the current and volts are in phase the averaging meter will show this, but the rms meter could give a false phase.

Does anyone ever need RMS AC voltage measurements? OK, for RF power, but multimeters are no good for this. Most equipment now is electronic and if you want power delivered to load you need to multiply current and volts. RMS volts does not cut it except for a constant resistance load, which barely exists anymore, maybe an electric heater element is as close as it gets, but an averaging meter works fine for this. A non sinusoidal power supply into a resistor load? Never needed that in my career.

RMS has complexity and battery life costs and different meters could give different results with high crest factors.


Additonal (corrected):
Instead of RMS why not true power?
Instead of SQRT(average  (V^2)) or SQRT(average (I^2))
Please could meter manufacturers do
Average ((V*I))

[electr_peter]

I have an application for measuring phase angle on an electrical supply by measuring voltage, voltage across a resistor for current and the vector sum of both. Use of trigonometry derives phase angle. But you need to use an averaging meter for good results. When the current and volts are in phase the averaging meter will show this, but the rms meter could give a false phase.

You need simultaneous measurement setup of voltage and current (preferably an oscilloscope) to get a phase angle or power factor. I don't know what you do, but it sounds like TRMS or AVG DMM would make no difference.

Does anyone ever need RMS AC voltage measurements? OK, for RF power, but multimeters are no good for this. Most equipment now is electronic and if you want power delivered to load you need to multiply current and volts. RMS volts does not cut it except for a constant resistance load, which barely exists anymore, maybe an electric heater element is as close as it gets, but an averaging meter works fine for this. A non sinusoidal power supply into a resistor load? Never needed that in my career.

You do understand that for power measurement it is not enough to multiply V and I (whether TRMS or AVG), right? Have you heard about devices called SMPS? They load line with non-sinusoidal load. For simple applications and sine waves TRMS=AVG, so what is the problem?

RMS has complexity and battery life costs and different meters could give different results with high crest factors.

It depends, see specs.

I don't get your rant. TRMS DMMs are here to provide TRMS results for somewhat non-sinusoidal waveforms because simple AVG meters flat out lie for anything other than sine wave. AVG type can be a benefit for some setups but you need bench type >6 digits DMM with special settings configuration to get any benefit of it. Power factor and phase angle has to be measured by checking voltage and current with high sampling rate, TRMS vs AVG has little difference here.

TRMS meters have their place, AVG meters have their place, TRMS is superior in many cases. Choose right meter for the job according to your specs and right measurement techniques.

[Ericho]

A Con..No not at all

It just tells you what the number on the display means, nothing more.

There are plenty of electrician meters for sale that use good old fasion averaging acV

In the field I couldn't care less if the dmm is true rms or not, as long as the number is in the ballpark 

[rsjsouza]

Some articles may help understand the benefits of TRMS.

http://www.newark.com/pdfs/techarticles/fluke/whyTueRMS.pdf

http://www.testequipmentdepot.com/application-notes/pdf/hvac/hvac-applications-where-true-rms-rest-tools-matter_an.pdf

These ones are a bit more technically deep:
http://www.escoglobal.com/resources/pdf/white-papers/True_G2.pdf
http://educypedia.karadimov.info/library/5988-6916EN.pdf

[rs20]

I have an application for measuring phase angle on an electrical supply by measuring voltage, voltage across a resistor for current and the vector sum of both. Use of trigonometry derives phase angle. But you need to use an averaging meter for good results. When the current and volts are in phase the averaging meter will show this, but the rms meter could give a false phase.

You need simultaneous measurement setup of voltage and current (preferably an oscilloscope) to get a phase angle or power factor. I don't know what you do, but it sounds like TRMS or AVG DMM would make no difference.

Eh? The term "phase angle" implies that we're dealing with sinusoidal signals, and with sinusoidal signals, you can indeed (at least theoretically) perform the vector sum of voltage and current and compare it against voltage and current signals alone to derive phase angle. E.g.

If the phase angle between voltage and current is 0 degrees, they combine constructively, so the RMS of (voltage + IR) will be (RMS of voltage) + (RMS of IR).
Phase angle is 90 degrees, RMS of (voltage + IR) will be hypotenuse (  (RMS of voltage), (RMS of IR))
Phase angle is 0 degrees, destructive inteference,  RMS of (voltage + IR) will be |(RMS of voltage) - (RMS of IR) |.

Everything inbetween is given by the cosine rule. So, with three DMMs, you can measure phase difference. Is it accurate or an advisable way to do it? No comment 

The thing I don't get is why using an RMS meter would cause problems in particular. If your signals are truly sinusoidal, the amplitude ratios between the summed and non-summed signals will be the same, whether or not the DMM is RMS. Replace all instances of RMS above with mean-of-absolute-value, and it still works just fine. Any function which obeys k*f(X)=f(k*X) (i.e., double any waveform, and the reported voltage doubles) works fine.

EDIT: Just to be crystal clear about this, if you have one signal that the DMM reports as 12V, and another that the DMM reports as 1V, and they're identical modulo scaling ("in phase" in the sinusoidal case), the summed signal will report as 13V, both for an averaging meter and an RMS meter (even though different DMMs might have a different opinion on what a given signal is, they will all [at least theoretically] follow the summing rule I've just stated).

[DAIRVINE]

"You need simultaneous measurement setup of voltage and current (preferably an oscilloscope) to get a phase angle or power factor. I don't know what you do, but it sounds like TRMS or AVG DMM would make no difference."

Yes I do get that. I am looking at vectors here and using trigonometry. What I am doing is measuring voltage and current spearately and then both together and deriving power factor from trigonometry. For sine waves you can use AVG or RMS, for SMPS with peaky current I think AVG gives a more correct phase angle answer than RMS. So I questioned where RMS is an advantage and don't really see any advantage for RMS.

"simple AVG meters flat out lie for anything other than sine wave."

In the context of power transfer to a resistive load this is correct if you want the equivalent DC voltage. I do understand this, just questioning whether it is still relevant. For anything complex I would use a scope.

For DC voltages I am quite happy with average, even for 6.5 digit meters. For measuring battery consumption I would use average so I can work out the battery life. For DC I prefer average even when the load is peaky.

Where is RMS important?
An article linked by rsjsouza said:
"The purpose of RMS value is to simplify power calculations by providing equivalent DC voltage that would develop the same average power in resistive load."
When do we ever see resistive loads? RMS is pointless if the load is not resistive.


  "if you have one signal that the DMM reports as 12V, and another that the DMM reports as 1V, and they're in phase, the summed signal will report as 13V, both for an averaging meter and an RMS meter"
True if both are sine waves. It think it does not work for RMS is one wave is, for example, smps.

[rs20]

"if you have one signal that the DMM reports as 12V, and another that the DMM reports as 1V, and they're in phase, the summed signal will report as 13V, both for an averaging meter and an RMS meter"
True if both are sine waves. It think it does not work for RMS is one wave is, for example, smps.

What does phase angle even mean in the context of non-sinusoidal signals? Is that a well-defined concept? I'm curious to see a definition of phase angle and an associated proof that an averaging meter gives a closer value to this definition than the RMS meter.

[electr_peter]

"You need simultaneous measurement setup of voltage and current (preferably an oscilloscope) to get a phase angle or power factor. I don't know what you do, but it sounds like TRMS or AVG DMM would make no difference."

Yes I do get that. I am looking at vectors here and using trigonometry. What I am doing is measuring voltage and current spearately and then both together and deriving power factor from trigonometry. For sine waves you can use AVG or RMS, for SMPS with peaky current I think AVG gives a more correct phase angle answer than RMS. So I questioned where RMS is an advantage and don't really see any advantage for RMS.

I get that method, but I doubt accuracy of such technique for unknown loads.

As for TRMS, it does not give you any advantage over the AVG when both methods are appropriate (TRMS=AVG in that case), but TRMS is TRMS for non-sine signals (AVG does not provide correct results for non-sine wave). Also, remember that AVG meters are designed to provide TRMS reading for a sine waves. TRMS type result is useful for AC analysis and measurements.

However, for any given signal type or waveform (except very high peaks) TRMS = AVG * constant, so any trigonometric operations are not affected. So I still do not see the problem in your case. If V is sine wave and I is not sine wave, V/I/phase angle method is simply wrong.

"simple AVG meters flat out lie for anything other than sine wave."
...
For DC voltages I am quite happy with average, even for 6.5 digit meters. For measuring battery consumption I would use average so I can work out the battery life. For DC I prefer average even when the load is peaky.

You should differentiate between measuring technique for a single measurement and measurements over time period. High end >6 digits DMM may use AVG(multiple samples for a single measurement) or TRMS circuitry to get TRMS measurement for AC waveform, and then integrate/add/avg/TRMS resulting series of measurements. Initial measurement type is still TRMS type usually (for a sample time scale).

Where is RMS important?
An article linked by rsjsouza said:
"The purpose of RMS value is to simplify power calculations by providing equivalent DC voltage that would develop the same average power in resistive load."
When do we ever see resistive loads? RMS is pointless if the load is not resistive.

"Pointless" for you, maybe not for others. AVG is considered misleading with some signal types. AVG DMM may miss peaks due to limited BW anyway, so what is the advantage then?

Phase angle for SMPS is not important. Power factor is important which consists of phase angle + sin wave distortion which you cannot determine from trigonometry - you need to get real power, voltage and current (many methods, most reliable with oscilloscope).

[DAIRVINE]

rs20:
My interest is power factor correction. So magnetizing current and other currents are superimposed. So really I am only interested in the phase angle of the fundamental frequency. So, add in power factor capacitors and look for zero phase.

I REMOVED an example which was faulty.




Even if RMS and AVG gave the same result, what is the advantage of RMS?

[dom0]

Have you heard about devices called SMPS? They load line with non-sinusoidal load. For simple applications and sine waves TRMS=AVG, so what is the problem?

See attached pic. This is the mains current drawn by a radio set. And it doesn't even have a SMPS, it's just a standard 50 Hz transformer and some DC/DCs for logic supply behind that.



As one can easily see the current waveform isn't even close to what you'd expect after reading a text book 'bout transformer-linear-supplies. It just goes like "wardagblablahurrdurrwardasgabl"¹.

¹ no, that isn't a German word

[electr_peter]

10 peak with 1% duty cycle:

• AC AVG = ~0.22
• AC TRMS = ~0.995
• AC+DC TRMS = 1

20 peak with 1% duty cycle is exactly double the value for all methods.

When you have to pick one number to describe waveform you are measuring, TRMS is much better than AVG. AVG may be very confusing, TRMS at least indicates something (actual power of signal). TRMS is used in context of AC mostly because

Code: [Select]

power = voltage(TRMS) * current(TRMS) * fnc(power factor or phase angle)Equation

Code: [Select]

power = voltage(AVG) * current(AVG) * fnc(power factor or phase angle) is not valid

[TimFox]

Not your application, but in my former day job I needed to measure noise voltage accurately (roughly audio bandwidth) and a true rms meter was essential.  I relied on a Marconi 2610, with its variable averaging time (after the "square", before the "root") to obtain stable results.

[macboy]

rs20:
My interest is power factor correction. So magnetizing current and other currents are superimposed. So really I am only interested in the phase angle of the fundamental frequency. So, add in power factor capacitors and look for zero phase.

A simple example where RMS does not add up simply is a peak of 10, 1% of the time, for example like a SMPS current. rms=SQRT(10^2/100)=1. AVG gives 10/100= 0.1
Add two together and get a peak of 20, 1% of the time.
rms=SQRT(20^2/100)=4. AVG gives 20/100=0.2
RMS gives 1+1=4.
AVG gives 0.1+0.1=0.2
Because AVG adds up better and places less emphasis on peaks it is closer to the sine wave result.




Even if RMS and AVG gave the same result, what is the advantage of RMS?

Have you ever looked at the waveform of coming out of an electrical socket?
If the supply is heavily loaded with devices using rectifer-capacitor supplies (which is most electronics except the rare one with PFC), then the peaks are heavily loaded and flattened.

An average responding meter (which has a scaling factor to give an "RMS" result) will give a value which is higher than the true RMS value. A peak repsonding meter (again with scaling for RMS result) will give a value lower than the true RMS value. The True RMS responding meter will give a reading which can be used to calculate the actual power dissipation in a resistive load. The others give a number which is useful for .... very little if anything.

[rsjsouza]

Where is RMS important?
An article linked by rsjsouza said:
"The purpose of RMS value is to simplify power calculations by providing equivalent DC voltage that would develop the same average power in resistive load."
When do we ever see resistive loads? RMS is pointless if the load is not resistive.

The way I learned is: since the amount of current/voltage provided to a resistive load is purely real (no imaginary component), measuring RMS values of non-resistive loads would "automagically" represent the real component of the power triangle in Watts. In Portuguese we call RMS something similar to "effective value" - the amount of energy/power that yields work.

We then used a wattmeter that measured what we used to call "true power" in VA and did some simple math to get the apparent or reactive power.

If I haven't missed any detail (it's been 15+ years ago), the RMS measurement is still very relevant. 

[rs20]

I have to endorse the OP's question, and point out that it hasn't been answered yet. Linked documents at the start say that averaging meters give an answer that is "wrong by up to 40%", but only under the presupposition that RMS current is the "correct" one. And the only justification given for RMS presupposes a load that is purely resistive/capacitive/inductive; and that is not what an SMPS (or even a classic rectifier-after-transformer approach) does, as illustrated in this thread.

How can TRMS ever be useful, when the most common reason for a non-sinusoidal signal is a non-resistive load, and TRMS is founded upon the assumption of a resistive load? The only use I can contrive is I^2 R losses in cables. And measuring square-wave output inverters with a view to powering resistive toasters or incandescent light bulbs. Which seems very niche these days, even a compact fluoro bulb blasts your assumptions to oblivion, and renders TRMS and averaging equally  useless.

[AG6QR]

"if you have one signal that the DMM reports as 12V, and another that the DMM reports as 1V, and they're in phase, the summed signal will report as 13V, both for an averaging meter and an RMS meter"
True if both are sine waves. It think it does not work for RMS is one wave is, for example, smps.

What does phase angle even mean in the context of non-sinusoidal signals? Is that a well-defined concept? I'm curious to see a definition of phase angle and an associated proof that an averaging meter gives a closer value to this definition than the RMS meter.

You can certainly argue that phase angle is an ill-defined concept for non-sinusoidal signals.  For many purposes, I'd agree wholeheartedly.

But I've seen phase angle defined as the arccosine of the power factor.  In other words, the arccosine of the true power divided by apparent power.  True power is defined as the integral of instantaneous power over a full AC cycle.  Apparent power is defined as Vrms * Arms.

This definition of phase angle matches the traditional definition when the waveforms are all sinusoids, but this definition is independent of the waveforms.  Voltage, Current, and Power may all be non-sinusoidal.  Notice that this definition doesn't distinguish between leading and lagging phase angles, and you can construct non-sinusoidal waveforms for which the concept of "leading" versus "lagging" doesn't make any sense.

To measure this definition of phase angle, you've got to have an RMS measurement of voltage and current, and a true measurement of power.  The true measurement of power can come from measuring heat dissipated in the load, or can be approximated by measuring instantaneous voltage and current at many points along the power cycle, an integrating them using a numerical method.  The RMS voltage and current measurements come from a true RMS meter, thus answering OP's question about why an RMS meter is needed.

[rs20]

What does phase angle even mean in the context of non-sinusoidal signals? Is that a well-defined concept? I'm curious to see a definition of phase angle and an associated proof that an averaging meter gives a closer value to this definition than the RMS meter.

You can certainly argue that phase angle is an ill-defined concept for non-sinusoidal signals.  For many purposes, I'd agree wholeheartedly.

But I've seen phase angle defined as the arccosine of the power factor.  In other words, the arccosine of the true power divided by apparent power.  True power is defined as the integral of instantaneous power over a full AC cycle.  Apparent power is defined as Vrms * Arms.

The OP responded to say that he was defining phase angle as the phase angle between the fundamentals; which may be nonstandard* but it seems very relevant and well-suited to his (nevertheless questionable) use case of finding the optimum amount of capacitance to add to achieve a maximum power factor. Furthermore, even more interestingly, I've given it some thought and an averaging meter is indeed much more likely to ignore harmonics (which is what the OP wants) than a TRMS meter.

* Unless it turns out to be identical to your standard definition through mathematical magic, which I suspect may perhaps be true since everyone's trying to maximize power factor and doesn't care about harmonics** at the end of the day here.
** And we don't care about harmonics if either V (as is in this case) or I is harmonic-free.

OP: You realise that your method can't distinguish between leading and lagging phase?

[dacman]

1 VRMS Sine Wave
Peak to Peak: 2.8284
Peak: 1.4142
RMS: 1.0000
Average (Absolute) Value: 0.9003
Average Responding; RMS Sine Wave Calibrated; Meter Indication: 1.0000

1 VRMS Square Wave
Peak to Peak: 2
Peak: 1
RMS: 1.0000
Average (Absolute) Value: 1.0000
Average Responding; RMS Sine Wave Calibrated; Meter Indication: 1.1107

1 VRMS Triangle Wave
Peak to Peak: 3.4641
Peak: 1.7321
RMS: 1.0000
Average (Absolute) Value: 0.8660
Average Responding; RMS Sine Wave Calibrated; Meter Indication: 0.9619

For a waveform, if each point along the waveform is squared, then these squares are averaged, then the square root is taken of this average, the result is the RMS value, for any waveform, including DC.  The RMS value is also the heating value.  This is important for many reasons, such as the ampere rating for the wiring in a house would usually be limited by heat caused by amperage.  A power transformer would be limited by electrical heating.  Wiring in an airplane needs to be sufficient to not overheat.  A power resistor (or any resistor) is rated so that it is not overheated.  A fuse mainly works by heat.  Breakers work by heat (usually).  The list goes on and on.

[rs20]

For a waveform, if the square root is taken of the absolute value of each point along that waveform, then these square roots are averaged, then this average is squared, the result is the RMS value, for any waveform, including DC.  The RMS value is also the heating value.  This is important for many reasons, such as the ampere rating for the wiring in a house would usually be limited by heat caused by amperage.  A power transformer would be limited by electrical heating.  Wiring in an airplane needs to be sufficient to not overheat.  A power resistor (or any resistor) is rated so that it is not overheated.  A fuse mainly works by heat.  Breakers work by heat (usually).  The list goes on and on.

Yep, that's a good point. Even if loads are typically much more interesting than they used to be, and therefore RMS is no use in describing how much power draw/heating they experience, the wires, breakers and fuses leading to that load do behave according to the RMS current running through them. I think that fully justifies TRMS being standard for current measurement.

Just for the sake of argument, what's the least contrived reason someone can come up with for a true RMS voltage measurement? I submit an square wave inverter powering a toaster.

[ BTW, pedantically, you got the square root and square the wrong way around there. ]

[mjkuwp]

This is a discussion that comes up a lot.  True RMS is not a con but maybe some day it will be a dated term.  It just means that the reading will be correct.  Older meters used tricks to accommodate lack of affordable technology at the time and would not give a true reading on all types of waveforms - as others have pointed out.

Power factor as it is taught in school assumes some idealized and simplified situations that don't often  happen in the real world anymore (pure sine waves shifted merely by a phase angle).  I think this academic description of power factor causes confusion.  With switching power supplies the current drawn by a load is quite a complex waveform.  Incandescent light bulbs draw nearly a pure sine wave and 1.0 power factor but these light bulbs are going away, albeit slowly.

Power factor is the ratio of Watts / V*A

Power factor can be negative and this means power is sent back instead of being drawn. (older terminology assigned a sign to power factor to indicate leading or lagging current)

Power measuring ics sample current and voltage at the same time - at several khz sample rate and multiply the values together to obtain the active power.

[dacman]

Yep, that's a good point. Even if loads are typically much more interesting than they used to be, and therefore RMS is no use in describing how much power draw/heating they experience, the wires, breakers and fuses leading to that load do behave according to the RMS current running through them. I think that fully justifies TRMS being standard for current measurement.

Just for the sake of argument, what's the least contrived reason someone can come up with for a true RMS voltage measurement? I submit an square wave inverter powering a toaster.

[ BTW, pedantically, you got the square root and square the wrong way around there. ]

You’re right, I got it backwards.  I went back and corrected it.

One point I was trying to make is that average responding meters indicate about 11% higher than the average so that they will indicate the RMS value of a sine wave.

[G0HZU]

Just for the sake of argument, what's the least contrived reason someone can come up with for a true RMS voltage measurement? I submit an square wave inverter powering a toaster.

I have an old analogue meter here which is designed to 'just' measure Vrms. That's all it does. It works up to a 20MHz bandwidth and can be useful for lots of measurements. eg complex signals, noise, signal to noise ratio and it can measure down to about 30uV with this 20MHz bandwidth.

Meters like this have sold in their thousands for many decades. You can do similar measurements with a decent Trms DMM but usually they only work for bandwidths up to a few hundred kHz.

[dom0]

Yep, even a 3400A from the 60s does a good job at measuring wideband (~25 MHz bw on mine) noise to reasonably low levels.

[Guni]

Can you explain difference between RMS DMM and TRMS DMM ?

[Guni]

Answer to my own question:

TRMS for correct measurement of mixed-mode (AC+DC) voltages and currents,
RMS for correct measurement of possibly non-sinusoid AC (without DC offset)

http://www.digitek.com.hk/en/newcon.php?id=12

It seems that some TrueRMS DMM on the market are RMS only.