Test frequency response/noise immunity of multimeters

[NiHaoMike]

3 years ago, a friend of mine wrote an article about how cheap multimeters might not read properly if there are harmonics or noise in the input. I actually bought a Mastech when I entered college because it was on special at a local electronics store at the time and my friend really likes it.
http://www.refrigeration-engineer.com/forums/showthread.php?t=8200

Here's a reason why you should not buy a cheap DMM. The input filter in cheap meters easily gets confused and attenuates the signal, and worse than that, the amount of attenuation is unpredictable so you have no way to know the real measurement!
To demonstrate, I'll use a cheap meter (about $35) and an expensive meter (about $80) to measure the normal AC line.

So far so good, right? That's because they're measuring an almost pure sine wave.
But measure the UPS output and look what happens!

That's because they're measuring a square wave, which confuses the input filter in the cheap meter. The expensive meter has better input filters that are not confused by high frequencies in the signal being measured.
This inaccuracy is confusing at best and dangerous at worst. A cheap meter can easily cost you a lot in the long run!

(Note that what she refers to as "accuracy" is better described as "measurement confidence".)
However, only two meters were included and the actual shape/harmonic content of the waveform is unknown. Perhaps you can do a more controlled test using a function generator and many multimeters, along with an oscilloscope to display the actual waveform? Maybe even set the function generator to sine (no harmonics) and crank up the frequency until the measurement is out of tolerance.

[squeezee]

The difference is between Average Responding and True RMS meters.

A short explaination probally isn't a bad idea though, perhaps in the $100 multimeter shootout or as a short supplimentary blog, as all of those are True RMS.

[alm]

But frequency response is a separate issue, true-RMS won't do you any good if a signal is outside the DMM's bandwidth.

[Kiriakos-GR]

But frequency response is a separate issue, true-RMS won't do you any good if a signal is outside the DMM's bandwidth.

That's true , but all the inverters with the output of 220 V or 110V 50-60Hz,  why they need anything more than 1KHz as max bandwidth?
With the center of the frequency at 50-60Hz  up to 1KHz you have capture even the smallest harmonic.
 

[Time]

Saying a filter is "confused" is poor use of anthropomorphism.

[vk6zgo]

Another interesting point is that DMMs sometimes give incorrect readings on the DC voltage ranges if there is an unexpected large AC component present at the point being probed,

Quite some years back,I was confronted with the problem of reduced vertical scans on a TV set.(My Mother in law's TV ).
The DC supply to the vertical circuits came from an overwind on the Horizontal output transformer,plus a 1/2 wave rectifier/filter cap setup.
The voltage was supposed to be around +130 volts.
The DMM showed about +108 volts-----low,but not that low,so I went looking elsewhere.

Next weekend,I borrowed an Oscilloscope from work & started looking around.
When I looked at the "+130 volt" test point,the fault was immediately obvious---virtually unfiltered 1/2 wave rectified 15kHz!
The filter cap was dead,but the DMM (Fluke 77) read it as DC!

VK6ZGO

[Kiriakos-GR]

The filter cap was dead,but the DMM (Fluke 77) read it as DC!

Now days we have the Agilent U1272A  AC+DC ..  nothing can fool it,  plus 100KHz bandwidth. 
Last night I was playing with AC+DC measurements on my sound generator ( sinusoidal - square - saw) at the 27KHz.
I got the output that it was specified in the specifications paper.