Measuring speaker's impedance using a DMM? - Is Erin serious, or am I missing something?

[PHD]

Sorry, but I assumed that Erin has some knowledge of basic electricity. Have a look at this:

He's advising people to use a DMM to measure the speaker's impedance? DMMs Ohm measurements are a simple DC measurement, not actual impedance.

Even with an LCR meter, you'll probably get the speaker's impedance wrong because a speaker + passive crossover is a complex circuit which have inductance, capacitance, and stray resistance.

It also has an induced back EMF, which together with the drawn current forms the main portion of the Ohmic (real) part of the impedance. The back EMF depends on the output power of the speaker and of course, the frequency.

When you measure the DC resistance of a commercial loudspeaker, all you get is the static equivalent resistance of the crossover + voice coils of the woofer and midrange drivers (not the tweeter because of the DC blocking capacitor).

I've just measured the resistance of my Forac Aria 906 speakers using a DMM and got a value of 3.6Ohms which isn't even close to the nominal 8Ohms.

I always assumed that to measure nominal impedance you need to drive the speaker near rated power using an amplifier and a tone generator, measure voltage and current per frequency, and then calculate the impedance and stage an average. I'd expect the Z(f) graph to change if you change the power because the higher the motor's excursion, the higher the back EMF would be.

I'd love to learn how exactly impedance graphs are generated.

 

[DonH56]

Sorry, but I assumed that Erin has some knowledge of basic electricity. Have a look at this:

He's advising people to use a DMM to measure the speaker's impedance? DMMs Ohm measurements are a simple DC measurement, not actual impedance.

Even with an LCR meter, you'll probably get the speaker's impedance wrong because a speaker + passive crossover is a complex circuit which have inductance, capacitance, and stray resistance.

It also has an induced back EMF, which together with the drawn current forms the main portion of the Ohmic (real) part of the impedance. The back EMF depends on the output power of the speaker and of course, the frequency.

When you measure the DC resistance of a commercial loudspeaker, all you get is the static equivalent resistance of the crossover + voice coils of the woofer and midrange drivers (not the tweeter because of the DC blocking capacitor).

I've just measured the resistance of my Forac Aria 906 speakers using a DMM and got a value of 3.6Ohms which isn't even close to the nominal 8Ohms.

I always assumed that to measure nominal impedance you need to drive the speaker near rated power using an amplifier and a tone generator, measure voltage and current per frequency, and then calculate the impedance and stage an average. I'd expect the Z(f) graph to change if you change the power because the higher the motor's excursion, the higher the back EMF would be.

I'd love to learn how exactly impedance graphs are generated.

I did not watch the video. You could use an RMS DMM (or pair of them) to measure voltage and current at various frequencies to judge the magnitude of steady-state impedance, I suppose. It has been ages since I measured speaker impedance, and I used a four-port VNA (Kelvin fashion, sometimes with a high-power amplifier, which also enabled a two-port VNA measurement with decent accuracy but complicated the test and post-processing), something beyond the reach of most hobbyists. Then converted from S-parameters to regular (vector) impedance. I think maybe there are inexpensive PC-based versions now? I believe there are speaker measuring setups from places like Parts Express.

I am not sure you need to measure at near rated speaker power? At least not for nominal impedance measurements. I do not recall the standards, was thinking it was usually measured at either 1 V or 1 W? Not my field... IIRC, the curves did not change significantly until power was high enough to be near compression and/or high and long enough that self-heating affected the drivers.

 

[Blumlein 88]

I am very surprised at this from Erin. Very wrong.

If you know a bit it can be done with resistors and a meter or two. I'm not going to explain, but you don't need much to get very accurate results. The thread below has some simple software that makes it much easier. REW is one that is widely available.

Is there easy way to measure speaker impedance

Is there an easy way to, without more sophisticated equipment, measure the speaker impedance at certain frequency? Let's say I have a frequency generator (read 'a laptop'), a DC/AC multimeter and an amplifier... ??

www.audiosciencereview.com

 

[Blumlein 88]

I did not watch the video. You could use an RMS DMM (or pair of them) to measure voltage and current at various frequencies to judge the magnitude of steady-state impedance, I suppose. It has been ages since I measured speaker impedance, and I used a four-port VNA (Kelvin fashion, sometimes with a high-power amplifier, which also enabled a two-port VNA measurement with decent accuracy but complicated the test and post-processing), something beyond the reach of most hobbyists. Then converted from S-parameters to regular (vector) impedance. I think maybe there are inexpensive PC-based versions now? I believe there are speaker measuring setups from places like Parts Express.

I am not sure you need to measure at near rated speaker power? At least not for nominal impedance measurements. I do not recall the standards, was thinking it was usually measured at either 1 V or 1 W? Not my field... IIRC, the curves did not change significantly until power was high enough to be near compression and/or high and long enough that self-heating affected the drivers.

Erin is showing measuring the DC resistance with a meter and then calculating power using voltage and that resistance. He should have done much better than that. At a bare minimum, too bare in my opinion, there is a rule of thumb it might average 1.5x the DC resistance. Even that is going to be far from accurate though closer than what Erin is showing in his video.

 

[DonH56]

Erin is showing measuring the DC resistance with a meter and then calculating power using voltage and that resistance. He should have done much better than that. At a bare minimum, too bare in my opinion, there is a rule of thumb it might average 1.5x the DC resistance. Even that is going to be far from accurate though closer than what Erin is showing in his video.

Thanks and Ach, no!!! Yeah, that won't be close except in a few lucky instances, and only for the woofer at that... And as I was recently chastised about, just measuring rms voltage and current gets you average power but more accurately VA since phase is not considered.

 

[Cbdb2]

The DC resistance is usually the resistance of the woofer voice coil and the lowest impedance of the speaker. Which can tell you how hard it is to drive.
I didn't watch the video but yea impedance measurement needs voltage, current and phase.

 

[NTK]

Please be aware of the context of this video. What Erin was showing was how to very roughly "calibrate" the meters on the Fosi LC30 to show the approximate amplifier power outputs. That's it. It is not intended to show how to measure the actual impedance of the speaker. Nor it is to enable the Fosi LC30 to accurately measure the amplifier output power, which I don't believe the Fosi can do since I doubt it measures both the voltages and currents of the amplifier outputs, and uses both to drive the meters.

He just needed a very approximate speaker impedance value to estimate the amplifier power output, and used it to "calibrate" the readings of the meters in the LC30.

 

[Scholl]

A viewer mentioned this thread in the video comments and this is Erin's answer :

I see someone else replied with the same thing so I'll copy/paste my reply here:It's been a while so help me out here... was it me saying to take the impedance of the speaker and calculate power based on that? If so, my rationale is that there's no way to calibrate this unless it's for a static load. So, what you're doing here is calibrating it for the Re of the speaker at Fs. That's the point where power is consumed the most. I don't know of an easier way for the modern consumer to "calibrate" this thing unless they used a static load which most aren't goin to run out and buy. I agree it's not ideal but I don't think it's so far from reasonable that it should be called "wildly inaccurate".

You're not actuallly measuring the entire impedance; you'd need something like the Dayton DATS or a painstaking process of playing different tones and measuring the resistance over a lot of frequencies (I used to do this before the advent of the Woofer Tester / DATS). If this is one of those "gotcha" moments then trust me, it's not that. There's just no way for someone to measure the entire bandwidth of a speaker and calibrate this for that. You have to assume some static resistance and you can get that at Re/Fs using a DMM which most have around. Alternatively, they can sweep the impedance with a DATS and then find the frequency for which they want to calibrate. I actually have a video on the Dayton DATS. Hope that helps.

 

[Cbdb2]

A viewer mentioned this thread in the video comments and this is Erin's answer :

"So, what you're doing here is calibrating it for the Re of the speaker at Fs. That's the point where power is consumed the most."
Well hes right and wrong. The DC resistance, Re a pure resistance, what hes measuring, is the lowest impedance of the woofer (unless the xover is doing something unorthodox) which dosnt change with freq. and that is the largest load. The speaker impedance is highest at Fs, the resonant freq, so its the point of least power consumed. This seems to be a common mistake. Its always easiest to move something at its resonant freq. Try to get a clock pendulum moving at anything other than 1hz its resonant freq.

 

[J.M. Noble]

You're better off using the manufacturer's nominal impedance for a voltage-based power estimator (I hesitate to say "meter") than the driver's DCR.

 

[Blumlein 88]

A viewer mentioned this thread in the video comments and this is Erin's answer :

I disagree even with his explanation. I agree there is no simple answer. That is what he should have said. An old rule of thumb would be 1.5 times DC resistance. Don't like that either. Both can be described only as wildly inaccurate.

Better would have been the nominal rating of the speakers. That is supposed to be 1.25 times a speakers minimal impedance. Impedance not DC resistance.

I see JM Noble just posted the same thing.

 

[PHD]

"So, what you're doing here is calibrating it for the Re of the speaker at Fs. That's the point where power is consumed the most."
Well hes right and wrong. The DC resistance, Re a pure resistance, what hes measuring, is the lowest impedance of the woofer (unless the xover is doing something unorthodox) which dosnt change with freq. and that is the largest load. The speaker impedance is highest at Fs, the resonant freq, so its the point of least power consumed. This seems to be a common mistake. Its always easiest to move something at its resonant freq. Try to get a clock pendulum moving at anything other than 1hz its resonant freq.

No, he's only wrong! Can you tell the nominal power of an electric AC motor just by measuring the DC resistance of its stator coil?

Of course not because an ideal AC motor would have 0 DC resistance and still be capable of generating MW of real power.

Power in electromagnetic-based electromechanical transducers like loudspeakers and AC motors, is generated by the product of the drawn current and the induced back electromotive force (EMF as in voltage).

The only way to assess the equivalent resistance of the active power drawn is to apply voltage and, measure the current while power is generated.

 

[NTK]

I'd love to learn how exactly impedance graphs are generated.

This.

 

[Cbdb2]

No, he's only wrong! Can you tell the nominal power of an electric AC motor just by measuring the DC resistance of its stator coil?

Of course not because an ideal AC motor would have 0 DC resistance and still be capable of generating MW of real power.

Power in electromagnetic-based electromechanical transducers like loudspeakers and AC motors, is generated by the product of the drawn current and the induced back electromotive force (EMF as in voltage).

The only way to assess the equivalent resistance of the active power drawn is to apply voltage and, measure the current while power is generated.

We are talking about conversion efficiency of less than 5%, your ideal motor is %100. Most of the power going into a speaker is wasted as heat in the Re. The largest speaker load is where the impedance is lowest. The lowest impedance a speaker can have is Re plus the couple percent of acoustic power which can be ignored. In other words the Re of a driver is usually very close to the lowest impedance of the driver. Measure some. The DC resistance of a multidriver speaker is usually the Re of the woofer where most of the power from music goes so the DC resitance is an approximation of the entire load. Not ideal but if all you have is a DMM it will give you an idea. And look at the meter, you'll barely notice a 3db difference.

 

[PHD]

We are talking about conversion efficiency of less than 5%, your ideal motor is %100. Most of the power going into a speaker is wasted as heat in the Re. The largest speaker load is where the impedance is lowest. The lowest impedance a speaker can have is Re plus the couple percent of acoustic power which can be ignored. In other words the Re of a driver is usually very close to the lowest impedance of the driver. Measure some. The DC resistance of a multidriver speaker is usually the Re of the woofer where most of the power from music goes so the DC resitance is an approximation of the entire load. Not ideal but if all you have is a DMM it will give you an idea. And look at the meter, you'll barely notice a 3db difference.

My Focal Aria 906 are rated as 8 Ohms and measure using Fluke DMM at 3.6 Ohms. This seems way off. However, Amir's measurements show a minimum impedance of 4.5 Ohms for these speakers.

 

[Cbdb2]

My Focal Aria 906 are rated as 8 Ohms and measure using Fluke DMM at 3.6 Ohms. This seems way off. However, Amir's measurements show a minimum impedance of 4.5 Ohms for these speakers.

Which makes how many dB of power difference ?

 

[J.M. Noble]

Which makes how many dB of power difference ?

Assuming a 20Vrms amplifier output that doesn't sag at these resistances and purely resistive loads (i.e. the DC resistance = effective impedance):

3.6 Ω = 111 Watts = 20.45dBW
4.5 Ω = 89 Watts = 19.49dBW

Less than 1dB difference.

 

[Cbdb2]

Assuming a 20Vrms amplifier output that doesn't sag at these resistances and purely resistive loads (i.e. the DC resistance = effective impedance):

3.6 Ω = 111 Watts = 20.45dBW
4.5 Ω = 89 Watts = 19.49dBW

Less than 1dB difference.

Now look at the meter he was calibrating, it would barely show up.

 

[solderdude]

VU meters measure a voltage and do so with a certain delay so peaks will not register which is what one would like to see.

DC resistance of a speaker says little to nothing about the nominal impedance which is mostly measured at 1kHz.
So.. to 'calibrate' such meters just apply a known voltage (measure with a MM that has autoranging or at least 2V AC range at say 100Hz (not all Multimeters work correctly at/or above 1kHz but most do, they all will work fine at 100Hz).
Then calculate the power that belongs to that voltage using the manufacturers specs for 1kHz and 'calibrate' to that calculated power level.

Will the VU meter (with Watt scale) actually measure the drawn power from the amp ? No it won't. These meters are just 'fun' to watch and will give an indication of the outputted voltage (minus short peaks).

 

[EJ3]

VU meters measure a voltage and do so with a certain delay so peaks will not register which is what one would like to see.

DC resistance of a speaker says little to nothing about the nominal impedance which is mostly measured at 1kHz.
So.. to 'calibrate' such meters just apply a known voltage (measure with a MM that has autoranging or at least 2V AC range at say 100Hz (not all Multimeters work correctly at/or above 1kHz but most do, they all will work fine at 100Hz).
Then calculate the power that belongs to that voltage using the manufacturers specs for 1kHz and 'calibrate' to that calculated power level.

Will the VU meter (with Watt scale) actually measure the drawn power from the amp ? No it won't. These meters are just 'fun' to watch and will give an indication of the outputted voltage (minus short peaks).

I thought that the object was to have some COOL lighting with dancing needles as ENTERTAINMENT! They'll let you know when the BASS notes hit and that is about as accurate as this meter can be.