How can I test if a DMM is reading in true RMS ?

[opablo]

I'm about to buy a ridiculously cheap DMM with true RMS (Uni-T UT61D)
It is so cheap that I'm skeptical and want to test if it is really measuring in true rms.

How can a create the most simple test without a function generator ?
Maybe creating an AC square wave with a 9+9 transformer and checking if the rms meter in ac mode matches the peak voltage and verifying that my other meter has a lower value ?

Please don't tell me that it's stupid to buy a measuring device that you don't trust... I know that.
I'm on a tight budget.

[opablo]

Is there a more simple test like measuring the mains having a diode in series with one of the probes to break the sine wave ? Would that have different reading between an rms vs. non-rms meter ? what should that reading be ?

[IanB]

Is there a more simple test like measuring the mains having a diode in series with one of the probes to break the sine wave ? Would that have different reading between an rms vs. non-rms meter ? what should that reading be ?

I think that should work. Wire up a circuit with a bulb to provide a load and measure the voltage across the bulb terminals. Let's say that is 120 V. Now insert a diode in series with the bulb and measure across the bulb terminals again. A true RMS meter should read 84 V more less (which is 120 / sqrt(2) ).

To eliminate the effect of the diode drop you could put two diodes back to back in the first circuit and just remove one of the diodes for the second.

Obviously this is safer with a low voltage AC supply. Use a transformer and a small bulb for the load. Make sure the diode reverse voltage is rated for at least twice the AC voltage.

[T4P]

Signal gen in square wave, that should defeat most non-trueRMS meters since they are all sine-calibrated but hey the UT61D uses a AD737J, no worries about not reading TrueRMS because that's a step higher than commonly found industrial standard AD637J TrueRMS converters

If you are thinking about the UT61D just for the trueRMS, okay. But if you are using it as a all-purpose


The continuity is slow, autoranging is slow but if you need the backlight very badly by all means
compared to the UT61E which pretty much has the same TrueRMS circuits.

[opablo]

Thanks T4P... That information about it's chipset gave me a lot more confidence about it.
My other meter is a Mastech MS8229. Not fast refresh rate, fast but not latched continuity test, no analog bar, no rms.

The cheap rms meter that I -would- love to buy is the BK 2709B
But I live in Argentina and the only way for me to buy a DMM that matches my budget is through sites like dealextreme, dinodirect, tinydeals, aliexpress, etc. And you have to pick out of whatever happens to be available there.

I'm thinking in trying to mod/hack my MS8229 to add latching to it's high responsive continuity tester.

[iloveelectronics]

Thanks T4P... That information about it's chipset gave me a lot more confidence about it.
My other meter is a Mastech MS8229. Not fast refresh rate, fast but not latched continuity test, no analog bar, no rms.

The cheap rms meter that I -would- love to buy is the BK 2709B
But I live in Argentina and the only way for me to buy a DMM that matches my budget is through sites like dealextreme, dinodirect, tinydeals, aliexpress, etc. And you have to pick out of whatever happens to be available there.

I'm thinking in trying to mod/hack my MS8229 to add latching to it's high responsive continuity tester.

Try the Mastech MS8250C, it has an eqivalent chipset as the BK 2709B, and is available at DealExtreme at almost half the price of the BK. It even has a dual display. Main downside is that the Mastech defaults to AC measurements. But thanks to the dual display you would quickly recognize that you are in AC modes as you would see Hz in the secondary display.

[Lightages]

I can assure you that the UT61D/E will do True RMS properly. They do not like DC offsets of more than around 7.5V however so be aware of that. If you insist on buying a UT61E please read everything available here on the forums about this meter. Yes it is good value for the money but it has its problems. I have one and I still like it but it has many little problems.

If you can at all find a way to buy one and you feel it worth the extra money to you then have a look at the Amprobe AM-270. At $90 US it is one of the best deals available in a multimeter. You can get one from ebay from testequipmentdepot and yes they ship worldwide.

[Wytnucls]

They do not like DC offset at all!
I've just run a test with oscilloscope and function generator on 61E and 71D on AC+DC.
With a 100Hz 5Vpp sine wave or square wave at 50%, the 61E only measures TRMS accurately if there is no DC offset.
This is quite normal, since the meter doesn't have the AC+DC feature.
If you have tested the 61E yourself, I'd like to hear what your settings were.

[T4P]

The manual didn't explicitly say "NO DC" but they did say "Input Amplitude: (DC electric level is zero)" under voltage measurement  which doesn't really make sense under the Hz% area for voltage

[Lightages]

I did a rough test with my function generator and guestimated where the reading went way off the rail. Yes, any offset starts inducing errors but I was allowing a bit of error before it went way off.

[tom66]

Why don't they use an ac-coupling cap inside the meter?

[Lightages]

Actually, I redid my test using a oscilloscope and I made an error. It was my function generator clipping with the levels I was using. Within the limits of my measurements, my UT61E has no problems with DC offsets.

Anyway back to how to test True RMS on a meter. Feed it any logic level square wave of a low frequency, say 100Hz, and the meter should read 1/2 the logic level. It is that easy.

[Lightages]

Why don't they use an ac-coupling cap inside the meter?

To get a 6Hz cutoff high pass you need a capacitor of around 2.7 nanofarad and it needs to be very high working voltage, around 3kv? Easily done but it would need to be switched in and out of the circuit. If they had a dedicated position on the selection switch for AC then it would be easy. But because they have a soft switch to AC, it would need a relay.

[IanB]

Anyway back to how to test True RMS on a meter. Feed it any logic level square wave of a low frequency, say 100Hz, and the meter should read 1/2 the logic level. It is that easy.

Is that right? If you have a positive going square wave with a 50% duty cycle the voltage will be on for half the time, which means it will deliver half the power of a continuous voltage. That means the true RMS voltage should equal the logic level divided by the square root of two.

[Lightages]

You are correct if the meter is a True RMS +DC reading meter. If the meter is an AC only reading meter then it effectively shifts the waveform to a bipolar waveform, rectifies it, then measures the resulting "dc" level. I know this is a simplified explanation but it demonstrates the idea.

Think of it this way, the meter could be considered as AC coupled.

[lewis]

Anyway back to how to test True RMS on a meter. Feed it any logic level square wave of a low frequency, say 100Hz, and the meter should read 1/2 the logic level. It is that easy.

Is that right? If you have a positive going square wave with a 50% duty cycle the voltage will be on for half the time, which means it will deliver half the power of a continuous voltage. That means the true RMS voltage should equal the logic level divided by the square root of two.

You're right, RMS voltage of a logic level square wave is Vpk x sqrt(d)  where Vpk = logic voltage, d = duty cycle. This holds as long as the square wave does not go negative (ie only between logic high and ground).

If the square wave has no DC offset (ie goes above ground then below by the same amount) the RMS value is equal to peak voltage regardless of duty cycle.

(Someone can probably tell me what the RMS of a square wave is with an arbitrary DC offset, I don't know off the top of me noggin.)

[alm]

V_AC+DC = sqrt(V_AC² + V_DC²) (all RMS voltages, though DC RMS voltage is obviously just the DC voltage). In the case of a 0 to 5 V square wave, this is a square wave with 2.5 V amplitude (2.5 V RMS) plus a 2.5 V DC offset, or sqrt(2.5² + 2.5²) V = sqrt(2) * 2.5 V.

[Wytnucls]

I'm about to buy a ridiculously cheap DMM with true RMS (Uni-T UT61D)
It is so cheap that I'm skeptical and want to test if it is really measuring in true rms.

How can a create the most simple test without a function generator ?
Maybe creating an AC square wave with a 9+9 transformer and checking if the rms meter in ac mode matches the peak voltage and verifying that my other meter has a lower value ?

Please don't tell me that it's stupid to buy a measuring device that you don't trust... I know that.
I'm on a tight budget.

Be aware that although the 61E measures true RMS, it is defeated if any DC offset is introduced, like most TRMS meters are.
If you want to measure true RMS for a waveform with a DC offset, then you need a meter with an AC+DC option, like most bench meters have.
Crest factor is also to be considered, with most middle of the road meters giving accurate readings up to a factor of 3.
Accuracy gets worse as the signal frequency increases.
http://masteringelectronicsdesign.com/how-to-derive-the-rms-value-of-pulse-and-square-waveforms/
http://masteringelectronicsdesign.com/how-to-derive-the-rms-value-of-a-sine-wave-with-a-dc-offset/
http://en.wikipedia.org/wiki/Crest_factor

[Lightages]

As I have pointed out, my UT61E has no problem with a DC offset up to the limits of my signal generator. When feeding mine a 1V RMS sine wave it stays within 2% up to the 10V offset limit of my function generator, and within 3 or 4% correct with square and triangle waves. I was actually surprised that mine has no real problems after all the reports.

[Wytnucls]

You're not measuring it right.

[opablo]

Wytnucls: I had to do some reading and some googling to fully understand your explanation but now I have a much better understanding of all this.

Thanks for taking the time to reply.

This means that it's not a good idea to use a 0->5v square  logical signal to test it.
And it's also not a good idea to use half of the waves of the main using a diode to break the sine wave.

I have to really have a non-sine AC signal with a balanced positive and negative stage.....

.... so... what could be the easiest way to create it -without a func gen- ?

...what about having the 5v AC sine wave of a transformer passing though 4 diodes connected in parallel with another 4 in the opposite way like this:

Code: [Select]

---+---|>|----|>|----|>|----|>|----+----
    \--|<|----|<|----|<|----|<|---/


thus erasing big/long parts of the wave in which the voltage is low maintaining the balance of the parts untouched.

is that a good idea to create a clean non-sine AC wave with no DC offset ?

[Lightages]

You're not measuring it right.

Please elaborate on how it should be measured.

[Wytnucls]

I'm glad you asked:
Fire up your oscilloscope and make sure the channel trace is DC coupled. Install a BNC T-junction on your function generator and connect both your oscilloscope and the meter under test. The meter should be set on the ACV range.
Feed a 5Vpp sine wave at 100Hz to both of them, with 0 DC offset.
Note that TRMS is showing correctly on both oscilloscope and meter, namely 1.77VDC.
Now, increase the DC offset to 1V. The 61E will show something like 1.7016V, while the oscilloscope will display TRMS at about 2VDC.
With 2V DC offset, the 61E shows 1.706V  instead of TRMS 2.63V. With 3V offset, it displays 1.705V instead of 3.46V TRMS, and so on.
If your oscilloscope doesn't show TRMS, work it out with the formula for a sine wave with a DC offset.
If you can, get your 71E out of the cobwebs and it will show the correct TRMS with any offset in the AC+DC mode.
This holds true, for any function, as long as the crest factor is kept below 3.

[Wytnucls]

Wytnucls: I had to do some reading and some googling to fully understand your explanation but now I have a much better understanding of all this.

Thanks for taking the time to reply.

This means that it's not a good idea to use a 0->5v square  logical signal to test it.
And it's also not a good idea to use half of the waves of the main using a diode to break the sine wave.

I have to really have a non-sine AC signal with a balanced positive and negative stage.....

.... so... what could be the easiest way to create it -without a func gen- ?

...what about having the 5v AC sine wave of a transformer passing though 4 diodes connected in parallel with another 4 in the opposite way like this:

Code: [Select]

---+---|>|----|>|----|>|----|>|----+----
    \--|<|----|<|----|<|----|<|---/


thus erasing big/long parts of the wave in which the voltage is low maintaining the balance of the parts untouched.

is that a good idea to create a clean non-sine AC wave with no DC offset ?

Well, as TP4 was saying earlier, the easiest way would be to feed the meter any waveform that is not a sine wave, since the non-TRMS meter is calibrated to give you a voltage based on the sine wave TRMS formula.
However, you need to be able to work out the real TRMS of that waveform for comparison, so a square wave or triangular wave would be the easiest.
http://masteringelectronicsdesign.com/how-to-derive-the-rms-value-of-a-triangle-waveform/
I'm not sure how you would do that with diodes only, but it may be possible.

[Lightages]

My apologies for confusing people, I was wrong. This is not the first time I have confused myself with this issue.

Of course it is not a True RMS reading unless it does account for the DC offset, which the UT61X series does not.