Multimeter specific measure problem - zero instead of mV value

[mikolajj]

Hello,

I want to put together a simple solution to determine the impedance of a loudspeaker line from 2-4 loudspeakers connected in parallel either in series or in parallel + series to fit within the 4-8 ohm range.
I found on the internet a very cool way to measure impedance of speaker. Simple and very cheap ;-) especially as it doesn't require super equipment (just to know what you are doing) and has sufficient accuracy.
What I found was to calibrate the miltmeter to the right scale and then take measuremant - this is what I wanted to adopt
 


I have two multimeters Extech EX230 and Voltcraft VC880 - nothing fancy.

The Extech shows the result without any problem.
Whereas the Voltcraft calibrates normally, it shows 0.0000V (zero) during measurement and a frequency of 0 (it shows frquency during mesurement).

The question is whether I have a faulty multimeter - the mV measures normally (the calibration works after all - its result is 4% different from the Extech - so looks like mV range its working).

I assume it is some parameter specific to this multimeter. The internal resistance, the frequency of measurement, burden voltage problem - but its not a current measurement, so not this - makes something zeros the result in Voltcraft multimeter.
Obviously for work it is enough to give such a proven multimeter that shows the value but I would like to know why this might be happening and would like to understand it.

It would be great if someone with experience (which I lack at the moment) could explain to me how this is possible and maybe point out what can be done to use this multimeter that does not show a result after all?

Thank you in advance for your help and explanations.

Best regards
MJ

[donlisms]

I would be curious whether you could calibrate the Voltcraft down to 4mV (by reducing the input significantly), just to see if the meter can see the voltage without the other stuff in the path.  It looks like its lowest AC range is 4v, where the Extech appears to go much lower than that.

[mikolajj]

According to multimeter spec

Alternating voltage

| Range	| Accuracy	| Resolution	| Frequenz range
|   4 V	|   0,0001 V	|   ±(0,5 % + 40) |   ±(1,2 % + 40) |   ±(3 % + 40) |   ±(4 % + 40)	|   45 Hz - 1kHz |   1 kHz - 10 kHz |   10 kHz - 20 kHz |   20 kHz - 100 kHz

For 4Ω - 4mV (0,004V) we are balancing on edge of capabilities of Voltcraft if I understand accuracy 0,0001 0.5% (1kHz) and number of counts 40 - means for 0.0040V could be between 0.0082V (0.004 + 0,5% + 40 counts) to -0.0002V (0.004V - 0.5% - 40 counts) and I get 0.0000 almost instantly when I touch wire.

Extech has resolution of 0.1mV in 400mV AC range with 1% + 6 digits (no frequency) so 4mV could be 0.004V + 1% + 6 digits = 0.005V to 0.003V - hmmm so better then Voltcraft.


I will check if I can mesure 1,2,3,4 mV at all with voltcraft.

[wasedadoc]

According to multimeter spec

Alternating voltage

| Range	| Accuracy	| Resolution	| Frequenz range
|   4 V	|   0,0001 V	|   ±(0,5 % + 40) |   ±(1,2 % + 40) |   ±(3 % + 40) |   ±(4 % + 40)	|   45 Hz - 1kHz |   1 kHz - 10 kHz |   10 kHz - 20 kHz |   20 kHz - 100 kHz

.

The accuracy and resolution columns are transposed.

[Kleinstein]

Some metes have a problem with reading small AC voltages / current below some 1% or 0.1% of there range.  This is a relatively common problem of the rectifier or RMS solutions. For the low signals the internal offsets and tolerances make the result somewhat random - some units will show zero early and other may hardly get to zero at all or have a very distorted curve to force it to zero in some way.

Having only a 4 V range is not really helping the Voltcraft meter. the 4 mV are only 0.1% and even good bench DMMs may work poorly at such a small fraction of the range (e.g. no specs for <1% for FS for the HP34401). The Voltcraft meter also has such a limtation: the specs are only for 10% -100% of FS. In this respect this meter is a really bad choice for small AC voltage.

One may work around by using the AC current range. At least the meters has AC current ranges - some of the very cheap ones may skip those as they can measure 100 mV AC voltage either.

[DavidAlfa]

A speaker nominal impedance is given in DC, so why this?
Just measure it with the DMM in resistance mode.

The dynamic impedance varies a lot over the audible frequency, just an example:

[CaptDon]

The guy makes the statement "This 4 ohms is the resistive component", no it isn't. If you are measuring impedance it is the algebraic addition of the resistive component and the inductive component. Square root of (Resistance Squared Plus Inductive Reactance Squared). Quick way to know the rated impedance of a speaker, it is the measured D.C. Resistance times a value typically between 1.2 and 1.4. Sadly his lash up didn't show the high shift in impedance at the self resonant frequency. Another poster has shown a typical 'real' impedance sweep and it is very plain to see an 8 ohm speaker can easily show an impedance of over 50 ohms.

[Kleinstein]

+1 usually the speaker DC resistance is close to the nominal impedance.  This is specially the case with smaller speakers. So the factor us closer to 1 for small ones like 5 W and below.
The amplifiers are also no that picky about the impedance. They still work with less and more, just clipping a little earlier and thus getting slightly less maximum power out. If in doubt high impedance speaker produce less heat at the amplifier and this way better than too low impedance.

[mikolajj]

The accuracy and resolution columns are transposed.

It is like so in official Voltcfart documentation ;-)

A speaker nominal impedance is given in DC, so why this?
Just measure it with the DMM in resistance mode.

I understand this. For our field service technicians I used this simplified method. I have made 3 or 4 ways how to connect speakers, I have measured resistance with DMM and prepare them table with ranges they have to check depending on schema they chose (2, 3, 4 speakers). Resistance is not very far away from nominal impedance. Speaker with impedance of 8Ω has DC resistance 7.3Ω - for speakers we use.
So I can measure 2 speakers in series 2s, 2s2p or 2p and for a technicians its enough. They just have to check it to be sure where is no short or impedance is no to low by mistake.

If in doubt high impedance speaker produce less heat at the amplifier and this way better than too low impedance.

Do you mean, that with higher impedance there is less current flowing through and that's why amplifier doesn't heat up that much?
Do I understand it right, that it's always safe to connect 32Ω and higher impedance speaker but its unsafe to connect any lower impedance than minimal - lets say 4Ω, because this will make to big current flow through amplifier and burn it?


Do you have any experience how can I generate 1mV AC signal with some regular price hardware (not 300k EUR Rohde & Schwarz generator) to check this DMM capabilities - I know it's an academic problem, but I would like to dig it to the bone ;-)

I was also wondering If it's possible to calibrate DMM with 100Ω resistor and then measure voltage with operational amplifier (let say 100x gain) to get a different scale? What should I look out for?
(again I understand it can be a waste of time for geeks like you ;-) but I'm curious)

BTW I have found very interesting discussion about measure problems here
 

[Kleinstein]

With a higher impedance load the current will be lower and thus less power drawn from the supply. The amplifier will reach it's voltage limits and this way clip at high amplitude.
With a lower than optimal impedance the amplifier will have to provide a high current and may reach current limiting type clipping, before taking full use of the voltage. So more current and in addition relatively littte voltage drop at the load and more at the transistors in the amplifier.

At least transistor amplifiers are OK with a higher load. Some tube based amplifiers may be different and not like too high a load however.


For generating a low level AC signal one could start with a more normal voltage range (e.g. 1 V) and use a voltage divider, like 1 K and 1 ohm or 10 K and 10 ohm to get a 1:999 divider. If needed one could measure the resistors to maybe get a little better than the resistor specs. One could also check the divider with DC.

To generate an approximate sine one could start with a square wave (e.g. from a NE555 or 74HC14) and than add plenty of low pass filtering (e.g. 3rd oder passive or active with 1 OP-amp). For crude test also the output side from a transformer can be sufficient - though mains can change and may need checks before and after.
Another option is the sound card output from a PC.

[mikolajj]

Thank you for all answers.

Another option is the sound card output from a PC.

I was trying to do so with Function Generator for Android, but two phones I was trying, ware not able to make a wave as "quiet" as 1 mV. Second problem is, that I used my RIGOL 1054 (with all options on) but still it is not accurate enough - as I can see it on the oscilloscope screen. 1mV/div is almost impossible to measure - or I can't do it ;-)
So I have no rock-solid reference point.

I like the idea with voltage divider. I will try it first - so thank you for the hint.

[CaptDon]

Trying to measure 1 millivolt on many DMM's is useless if the most sensitive scale is .001 volts. Your readout can only indicate .000 (infinity to 1 error) .001 which is questionable to +/- .0005 volts (50% error) or .002 which is 100% error. If your meter can read down to .0001 volts then you could get a reasonable measurement of .001 volts with good accuracy. The original YouBoob speaker test was total crap to start with and there is so much bullshit half baked half correct info on youtube videos to start with!!! He should have set up his test to produce more like 40mv at 4 ohms or 80mv at 8 ohms. If you really want a meaningful impedance measurement then it should be taken 'in-situ' with the speaker mounted in the cabinet it will be used in. Then you do a sweep across the frequency range the speaker will actually be asked to reproduce. No sense testing the impedance of a 15" woofer at 15,000Hz!! The 'RATED' impedance of a speaker is seldom more than about 1.2 X D.C.R. and never less than the D.C.R. B.T.W., testing a speaker's impedance does not provide the total picture of the speaker's condition. It may have voice coil rub or the voice coil may be stuck due to overheating but the D.C.R. could still be close to the correct value. The impedance will probably look very wrong however when the speaker is driven at 1v.r.m.s. because a stuck voice coil won't show the impedance peaks at the several 'normal' mechanical resonances. Remember that the term impedance indicates that current and voltage are out of phase so you must measure both and calculate!!!

[CaptDon]

B.T.W., you mentioned this all pertains to your 'field service' team making the measurements? Far better to have your technicians also inject 1v.r.m.s. into the speaker line and actually listen to the speakers. Even do a sweep to listen for buzz and rattle sounds indicating damaged voice coils , torn cones, damaged surrounds or loose hardware. What are you maintaining?? Drive-in movies? Church Pa's? Night Club house sound systems?
An impedance measurement will barely scratch the surface of what is laying at the other end of that wire!!

[mikolajj]

He should have set up his test to produce more like 40mv at 4 ohms or 80mv at 8 ohms.

I think it is possible, but not with phone. I can use a waveform generator and regular amplifier - I think I can also make test like this.
BTW how does this PEAK measure speaker impedance? Isn't it the same way what I'm trying to do manually?

I know this is far from perfect measurement. Drive-in movies and PA's usually need 100V system and this is for small places 4-8 speakers.
Oh and we do a sweep and we listen to this peeeeee sound.

It all started to check if it's possible to measure impedance with DMM.
I know that 4-8mV is hard to be measured but I can see a difference when I'm trying to use 4Ω, 8Ω, 16Ω speakers so it kinda work.

I started this tread more to understand why one DMM showed (some) value and the other did not.

On the other hand I got a lot of ideas and comments I can work with to understand it better - thank you.