Working of True RMS analog voltmeter?

[I_am_learning]

How does a True RMS analog voltmeter work (if it exists)?
(I am not interested about digital Sampling methods)
I have heard about 'absurd' methods like heating up a coil through the sample voltage and using a thermocouple driven simple DC voltmeter to give the reading proportional to the temperature, which in turn is proportional to the RMS current.

Any other method?

 

[I_am_learning]

I have an idea in mind (which I din't got to hear elsewhere in my little search), which I am not sure about working.
1. Take a simple DC Voltmeter
2. Remove the permanent magnet.
3. Wind a field coil of high resistance in place of the permanent magnet to produce the required magnetic filed.
4. Shunt join the two coils.
5. Since the deflection is proportional to the Force which is proportional to => Magnetic Field Strength * Current;
But, Here, both the Magnetic field and the Current is proportional to Voltage, so Here
Deflection <is proportional to>=> Deflecting Force <is proportional to>=> V²
6. Now make a SQRT scale, and calibrate to get the true RMS.

 

[waht]

The RMS is defined as:

$$v_{rms} = \sqrt{\frac{1}{T}\int{v^2(t)dt}}$$


To take Vrms the analog way, you need to make a circuit that

1. Squares input voltage
2. Takes average
3. Takes square root

2 is easy. You could use an op-amp integrator to take an average.

But 1, 3 are kind of difficult, but still possible with very neat op-amp and diode tricks. With op-amps and diodes you can make a circuit that takes an exponential or a log of input. And with those, use a well know algebraic rule to make a square or a square root.

 

[I_am_learning]

Thanks for the suggestion.
In my suggestion in #2, I Did
1. Square input voltage
I suggested making arrangement such that the Deflecting Force is proportional to the square of the applied Voltage
2. Take average
Average Force is automatically obtained due to the averaging effect of the pointer mass, Just like in regular DC voltmeter. Thus deflecting force, on average is proportional to the average of the Voltage Squared. So, is the deflection.
3. Take square root.
This I suppose can be done by recalibrating the scale. Say, on top of the original linear scale write sqrt(2) in place of 2; 2 in place of 4; 3 in place of 9 and so on.

Won't it work? What do you think?

 

[Studiot]

I have heard about 'absurd' methods like heating up a coil through the sample voltage

In days of yore this was the only method sufficiently reliable and accurate for a laboratory RMS instrument.

Calorimetric methods are still used for higher frequencies as equipment to perform waht's three steps to heaven just does not exist or is probibitively expensive.

Other RMS methods include twin coil instruments, but these were developed for powerline frequencies and are restriced to that application.

You also asked for a true RMS meter.

Calorimetric methods are the only ones that will work on all waveforms.

In particular recalibring a scale will only work for a particular waveform.

 

[I_am_learning]

In particular recalibring a scale will only work for a particular waveform.

Why do you say that? Perhaps you din't carefully read my previous post.I strongly suggest doing that, and suggest making comments based on that.
I amNOT suggesting to recalibrate the scale so as to convert average reading to RMS, like Multiplying by 1.11 for the case of Sine waves. I am not suggesting this. I know it doesn't work.That would be just a linear scaling.

I am suggesting to recalibrate by replacing every numbers in the scale with its square root. It won't be leaner scale, but I think that is what which should make it work. (Remember, due to two coils, the deflection is already proportional to the average of the square of the Voltage)

 

[Studiot]

Square root of what voltage?

Do you understand what a moving coil meter (without additional circuitry) displays in response to an alternating input?

Nothing!

The way is open for you to implement and prove your design.

I would be interested to see its response to square, triangular or pulsed at 1% duty cycle waveforms.

 

[I_am_learning]

Square root of what voltage?
Do you understand what a moving coil meter (without additional circuitry) displays in response to an alternating input?
Nothing!

Ans: Square root of average of the square of the voltage.
Yes, I do. Nothing = 0 would be right answer if the Magnetic Field was produced by a permanent magnet. But I suggested using field winding, in which case, during the negative half of the alternating current, both the field and current are reversed, so, the deflection will be in the same direction.

I would be interested to see its response to square, triangular or pulsed at 1% duty cycle waveforms.

This type of reply was what I was seeking.
I want to know the scenarios where the design succeeds and where it fails. Could you elaborate your this point, please?(I know you guys are pretty intelligent. I simply had to use rather harsh language so as to provoke this reply. Otherwise you guys are too kind to point out my mistakes! :) cheers. )

 

[Studiot]

One of the desirable characteristics of a voltmeter is that it has very high input impedance; ie it does not draw appreciable current from the measureand.

Have you considered how much current you will need to replace a permanent magnet and howmany turns that will need? How about the effect on input impedance?

 

[sophiecentaur]

Do you know, from the tone of responses of some 'questioners' to some of the answers / responses, you'd really think they were paying a highly paid consultant instead of visiting the department of free opinions.
There is no answer to the OP. If there were, you'd be able to buy one of these fabulous , ideal, meters in every electrical shop in the world. You clearly can't achieve what's been demanded with simple passive circuitry so why not grasp the nettle and go for the DSP approach which will give you as good an answer as you are prepared to pay for?
It's like asking for a design of a rocket which uses gunpowder, to launch a satellite.

 

[Studiot]

I realize that you want to do the signal processing by analog means.

However electromechanical meters are several time more expensive (these days) than digital panel meter. This is even unmodified ones.

Possible analog approaches to solving waht's equation are:

To take the log of the voltage, halve it and take the antilog.
This gets you the square root.
Log / antilog amplifiers are easy to implement with op amps. All you need is an element with an exponential response curve (eg a pn diode or transistor junction) in the feedback loop.

Use a multiplier.
A full four quadrant multiplier will take care of any waveform. A two quadrant one will work if the waveform positive and negative excursions are symmetrical.
Again such multipliers are readily available.

The above circuitry will take you to frequencies well above those reachable by magnetic coils.

 

[The Electrician]

I have an idea in mind (which I din't got to hear elsewhere in my little search), which I am not sure about working.
1. Take a simple DC Voltmeter
2. Remove the permanent magnet.
3. Wind a field coil of high resistance in place of the permanent magnet to produce the required magnetic filed.
4. Shunt join the two coils.
5. Since the deflection is proportional to the Force which is proportional to => Magnetic Field Strength * Current;
But, Here, both the Magnetic field and the Current is proportional to Voltage, so Here
Deflection <is proportional to>=> Deflecting Force <is proportional to>=> V²
6. Now make a SQRT scale, and calibrate to get the true RMS.

Meters such as this, with two coils, have been used since the earliest days of electricity. They are called "electrodynamometer movements", and they do respond to the RMS value of the applied current. They do draw substantial current in operation; much more than a modern DVM. See:

http://www.engineersedge.com/instru...s_measurement/electrodynamometer_movement.htm

Thermocouple meters are not at all absurd. They have the advantage that they can be used at frequencies ranging up into the RF range (typically up to 65 MHz). They respond to the RMS value.

So called moving iron meters respond to the RMS value.
See:

http://avstop.com/AC/apgeneral/acmeasuring.html

Finally, electrostatic meters respond to the RMS value.

 

[I_am_learning]

Meters such as this, with two coils, have been used since the earliest days of electricity. They are called "electrodynamometer movements", and they do respond to the RMS value of the applied current. They do draw substantial current in operation; much more than a modern DVM.

So, the answer is that the design works.
However its drawbacks are (As I can see)
1. Relatively low input impedance.
2. Works best for only low frequencies and relatively high duty cycle.
(I would like further explanation on this.)

sophiecentaur:
I am not here to actually buy or use out something. So, no point in suggesting me to use this or that DVM. Also, in addition to telling me that "the Design Doesn't work because if it would work, we could see a lot of them in the market", you could explain why it doesn't work. Don't it work better than the normal Analog Voltmeters (that multiply average reading by 1.11)?

 

[The Electrician]

So, the answer is that the design works.
However its drawbacks are (As I can see)
1. Relatively low input impedance.
2. Works best for only low frequencies and relatively high duty cycle.
(I would like further explanation on this.)

A low duty cycle waveform contains high frequency harmonics, so that doesn't work for the same reasons that high frequencies generally don't work well with this type of movement. Things like skin effect, distributed capacitances in the coils, etc.

Most of these meters:
http://shop.ebay.com/i.html?_nkw=pa...igital)&_osacat=0&_trksid=p3286.c0.m270.l1313

are moving iron movements. These meters are used because they are rugged and low cost (but not very accurate), and in situations where their low impedance doesn't matter, such as monitoring the line voltage. You'll notice that the scale is typically compressed at the low end.

This:
http://cgi.ebay.com/GOOD-USED-YOKOGAWA-LABORATORY-STANDARD-VOLTMETER-2013-/230551180638?pt=LH_DefaultDomain_0&hash=item35adebd55e
is an example of a laboratory grade electrodynamometer voltmeter, mirrored scale and all. They are typically 1/2% accuracy.

Thermocouple meters are still being manufactured and used for measuring radio frequency currents. Currently available on eBay are these (some of them are not thermocouple type):

http://shop.ebay.com/?_from=R40&_trksid=m570&_nkw=rf+ammeter&_sacat=See-All-Categories

 

[I_am_learning]

So, I Leave concluding that:
Eelectro dynamometer are little superior than normal power line AC voltmeter in measuring RMS value. But their extra price isn't worth paying for the little improvement they bring; especially when high accuracy Digital VOMs / precise thermocouple meters are available.

Thank you.

 

[Carl Pugh]

To The Electrician, you are totally correct.

The only exception might be that thermcouples can be used to measure microwave power. MANY years ago Litton Industries measured microwave power with thermcouples. They put DC into a waveguide load and calibrated the thermcouple, then used the calibration to determine the microwave power.

 

[sophiecentaur]

sophiecentaur:
I am not here to actually buy or use out something. So, no point in suggesting me to use this or that DVM. Also, in addition to telling me that "the Design Doesn't work because if it would work, we could see a lot of them in the market", you could explain why it doesn't work. Don't it work better than the normal Analog Voltmeters (that multiply average reading by 1.11)?

I just meant that, if it were a practical solution, (i.e. 'works', in the sense that it could be made and sold) meters like that would actually appear in shops (or have appeared in the past).
The theory of multiplying two currents together and producing a force is faultless but there's more to it than theory.

As an aside, I could refer you to the Ratiometer design which was available for use (and I used a number in Analogue days). This told you the ratio between two DC currents (in the mA region). It consisted of a conventional meter, with a permanent magnet for its field but with two windings and no return spring. The windings were at right angles to each other (i think) and 'worked against each other'. The deflection of one winding was balanced by the other winding and, over a fair range of currents but, as they were at right angles, one winding increased in torque as the other decreased so a balance was achieved at a certain angle of deflection. I guess the magnetic circuit didn't use a radial field, as most moving coil meters did - to make the torque per amp dependent on angle rather than independent - which gives conventional meters a linear scale.
The position on the scale gave the ratio of the currents in the two windings. Needless to say, the scale was very non-linear but readable.
So there you had a whacky solution (steam age) which was actually made to 'work' and which was implemented and sold.

 

[The Electrician]

I just meant that, if it were a practical solution, (i.e. 'works', in the sense that it could be made and sold) meters like that would actually appear in shops (or have appeared in the past).

Did you not follow the link(s) I provided? It is a practical solution, and such meters are made and sold, and actually appear(ed) in shops.

 

[sophiecentaur]

Mea culpa!
But I just realized a rotating disc Joulemeter effectively does the same thing with two coils.

 

[kevs926]

erm guys, isn't this just using ac bridge and getting the voltage across rectified waveform

if the frequency is high enough the meter will read the average voltage, rectifying means negating all the negative components of the voltage waveform as we all know.

getting the voltage just put a D'Arsonval galvanometer series with a resistor of known value and just make the scale

 

[sophiecentaur]

It will read the mean of the modulus of the current or voltage. That is not RMS and can only be corrected for if you know that the waveform is a sinusoid.