That's really interesting, dude!

VERY GOOD INFO

Nice description of RMS.

One of the best multimeter reviews, can u compare with FLUKE?

Nice explantion of RMS. On the symetric waveforms (symetricly around zero) you are correct. But on the asymetric waveform (only 0 to plus or 0 to minus) you are making a mistake. Simply expland by a square wave , 50% duty cycle, 1 volt peak. Power generated in a 1 ohm resistor is 1 watt. But that is only during 50% of the time. Over a full period on average only a 0.5 watt is generated in the resistor. The RMS needeed to produce this 0.5 watt over a full period is the square root of 0.5 = 0.707 Volt (0.707 Volt in 1 ohm gives 0.707 Amp is equal to 0.5 watt.) Not half the peak value. The general formula for RMS on an asymetric square wave is: U= Ui times square root of (ti devided by T) U= Urms Ui= peak value ti= time of the impuls T= time of a full period (in fact ti/T is the duty cycle) Why does the multimeter displays the symetric RMS value while measuring an asymetric waveform ? Simply because on AC somewhere in the meter a capacitor is put the signal path. You don't measure DC while the meter is in the AC position. So DC components are filtered out and the asymetric signal is automaticly turned into a symetric signal with half the peak value. (Peak to peak stays the same) In the video the DC ofset you put to the square wave is filtered out by the meter and altough this should be calculated into the true RMS value only the ac component is measured. And the ac component, because of the behaviour of the meter, is also measured wrong for the asymetric waveform. Conclusion: RMS on a multimeter is only valid for symetric waveforms with no DC in it. Actualy asymetric waves are not real AC but more a kind af amplitude modulated DC. Sorry for my lame english.

I would expect the AC measurement to be AC coupled and ignore any DC bias. The duty cycle affecting the readings I would not expect though and i find it rather strange especially since the meter does give correct values for the most part. It'd be interesting to see this kind of test with a higher dollar fruke meter or such in the lineup.

@randomtronic

5 жыл бұрын

Yep. Would be interesting to see some more expensive gear in same test. But my budget is blown for some time now :) Other commenter suggested another type of true RMS: AC + DC , (ut71a from unity) which I think would show expected values. But that renders all "regular" true RMS meters not so true...

@netman69

5 жыл бұрын

I guess there is no way around meticulously getting to know the meters behaviour before we can trust it :D. Personally I find it more useful to have the AC coupled RMS value, as it allows in a pinch to know what the RMS value of noise on a power rail is for example, whereas AC+DC does not really seem like something I would use (maybe it's very useful in some situations, but I have no idea). I may try to repeat what you did with a few meters if i get around to cobbling together or buying a signal generator.

24:50 That's because those DMMs are AC coupled, if you want to take into account that DC offset you need a AC + DC TRMS like the Uni-T UT71A

@randomtronic

5 жыл бұрын

Thanks, so I need one more multimeter after all :)

@modorangeorge4991

8 ай бұрын

@@randomtronic Or use 2 measurements with UT61E, one for AC, one for DC and do the math yourself. You can test with your scope AC coupled and then DC coupled and get the same results.

I wonder if this RMS behavior with square waves is so the reading on distorted AC Mains makes sense to say Electricians. eg SMPS without PFC, where the current draw is uneven and thusly the voltage waveform is not a perfect sinewave. Could you try feeding a distorted AC waveform ? would be interesting to see how clipping on a valve amplifier compares to a solid state amplifier, perhaps a viewer could do that ?

A square wave with a positive DC offset such that no part of the wave goes below zero is no longer an AC waveform, it's a changing DC waveform. Perhaps if you put the meters in DC mode, they will integrate the waveform and give you a more accurate value, even with a changing duty cycle. You may even find that higher frequencies work *better* in this case. Just a theory; most of my gear is in storage, so I can't test it myself...

@randomtronic

5 жыл бұрын

Thanks. DC readings will be way off in this case. The inaccurate readings with other than 50% duty are quite common with many "true RMS" meters. After reading all comments so far, I am inclined to believe, that only true RMS AC+DC meter would show correct measurement in this scenario.

I googled the question your title has posed us, when I saw other videos like yours on this quirk of a product description. I wonder how much appetite there is for producing a true true rms multimeter? I guess like you it doesn't exist - yet scopes seem better at it.

I came here after I got surprisingly poor results, as you did, from a square wave with anything other than a 50% duty cycle and no offset with my Fluke "true RMS" 117 meter. I similarly got preposterously bad results from any offset at all, even with 50% duty. For example, a 0-5V 60 Hz square wave with 50% duty should be 5V/(sqrt(2)) = 3.54 RMS volts. Interestingly my Fluke is also getting the totally incorrect value of 2.5 V, like your two RMS meters did for the test you show starting at 20:46. I'll guess this probably says something about how all these devices badly estimate results. My scope measures RMS fairly well, even with variable duty cycle, presumably because it actually does the integration electronically. Even a reputable Fluke meter is disappointing junk for that kind of analysis.

I don't think you can use a DDS (direct digital synthesis) signal generator for these tests. I'm not sure but I don't think the voltages are correct regarding real power etc. Maybe use stepped down mains or a normal oscillator. I might be wrong, but my meters do not read the voltage from a DDS properly, and remember they are floating and have no earth ground reference

@randomtronic

5 жыл бұрын

Thanks for the comment. I think this is not the issue though. After all, floating or not floating, voltage is voltage, right? And DDS is not a problem either, why would it matter how the voltage/signal is created?

I wonder if it best to think of these meters measuring the RMS value of the wave after it has gone through a bandwidth filter. When you change the duty cycle of a square wave then there will be much more energy in the harmonics (especially even harmonics). So I wonder if there is enough in these harmonics that they are getting filtered out at the about 1khz cutout (That is how the ANENG is behaving at any rate). So how much energy is above the 1khz cutoff for a 10% duty cycle square wave. Not sure right of the top of my head.

@randomtronic

5 жыл бұрын

Interesting theory! But that would mean, that aneng and uni-t would have to respond differently (uni-t has clearly higher bandwidth) but they respond the same.

@argcargv

5 жыл бұрын

But the uni-t is weird. Why would it read a higher value at higher frequencies if there was a bandwidth filter? I suspect the uni-t has no bandwidth filter so is susceptible to aliasing errors. However, as you say the two behave the same way which I can't explain.

@randomtronic

5 жыл бұрын

I suspect once the signal went past the bandwidth, the readings become undefined. Can be high or low, doesn't matter which way it goes. The true RMS converter inside just was not able to cope with it.

@argcargv

5 жыл бұрын

Ok, so I can explain why both meters read the same on the 90% duty cycle 10v p-p square wave. It is because it is AC RMS. The dc bias for this signal is 0.9*5+0.1*(-5) = 4, now the AC RMS will be sqrt((5-4)^2*0.9+(-5-4)^2*0.1) = 3. This is pretty close to what both meters are reading so the meters are reading true AC RMS. However this is different than true RMS. I tested my ANENG and it was able to both give an accurate measure of the dc bias point in DC mode and the AC RMS value in AC mode. Note, to get the true RMS value we we would take sqrt(4^2 + 3^2) = 5.

This is clearly a case of not knowing how meters works. True RMS meters will only measure the AC part of a signal, if you had put the scope on AC the readings would have matched. To get the full RMS value you have to add AC-RMS and DC value with sqrt(sqr+sqr). A DC+AC rms meter will do that for you. When you change the duty cycle the DC value will change, that is the reason scope and meters do not match. You cheap meter is probably using a single diode for rectification and the diode drop will affect the result significantly at low voltages. Many True RMS meters do also have problems with low values, precision is often rated from 5% or 10% of range and up.

@randomtronic

5 жыл бұрын

That's not the point I wanted to make. When I started to mess with the signal and adding DC offset, it was curiosity to see what will happen (I thought I had the scope on AC, I have to double check the video). However, a symmetrical square wave, RMS value should be equal to amplitude, but RMS meter shows correct reading only at 50% duty cycle. My question is: what it the point? If one can't measure signal other than square 50% or a sine, then it's not a true RMS, right? I know some more expensive meters have the ability to measure RMS AC + DC (haven't got one though), but why meters like those in video even exist? In my opinion, a true True RMS meter should not have AC/DC switch at all. It should show RMS (heating value) of what you stick the probes on. I think the True RMS label is used in a bit misleading way. Thanks for your comment! 👍

@henrikjensen3278

5 жыл бұрын

You do not get the point: When duty cycle is different from 50% there is a DC offset implied, you can measure that with the meter on DC. The AC do not include any DC. If you like that or not do not really matter, it is how multimeters works.

@randomtronic

5 жыл бұрын

@@henrikjensen3278 ok, I get what you are saying, non 50% square has a non-0 average value, therefore it has the DC offset, fine. However still I do feel that labeling a meter like this as RMS is used in a misleading way, but that's just my personal opinion.

@henrikjensen3278

5 жыл бұрын

If you ever get around to play with PWM you will find that they are closely related. I will suggest you connect two meters to your generator, one in DC and one in AC and then play around a bit, maybe you could learn something.

@randomtronic

5 жыл бұрын

@@henrikjensen3278 thanks, I somehow didn't think of duty and offset in those terms, but it does make sense. Thank you for taking time to get your point across. I think I understand something a bit better now. Still think cheap RMS meters suck though 😁

The average meter can't sample the output of DDS signal, mine don't, not even the Fluke.

@randomtronic

5 жыл бұрын

Which suggests some flukes have the same problem. Output of DDS is just as good as any other gen for purposes of this test.