why so much different impedance curves of same model?

[skyfly]

Spendor A7

The frequency of the lowest peak in impedance magnitude curve is below 25 Hz in Stereophile, but above 25 Hz in another review in youtube. Usual product by product difference, or usual measurement by measurement difference? Or difference caused by "burn-in"?

 

[kschmit2]

both curves are actually pretty much identical.
Top graph has y-axis of 0-50 Ohm, bottom graph has 0-20 Ohm

Top graph x-axis is 0 to about 25 kHz, bottom graph is 10 tp 50 kHz

 

[skyfly]

both curves are actually pretty much identical.
Top graph has y-axis of 0-50 Ohm, bottom graph has 0-20 Ohm

Top graph x-axis is 0 to about 25 kHz, bottom graph is 10 tp 50 kHz

Lower 20 and upper 20 are not identical.

Both graphs start at 10 Hz.

 

[tmtomh]

Lower 20 and upper 20 are not identical.

Both graphs start at 10 Hz.

@kschmit2 is correct - the graphs are essentially identical, with the two impedance peaks at the same locations in each graph. With such different vertical scales to the graphs, you can't just look at the impedance at a single frequency, especially one on a major impedance slope like 20Hz, and say "they don't match up." There is no doubt some minor smoothing, interpolation, or some such in the creation of such graphs, and so to go from less than 10 ohms to presumably 30 ohms in a range of about 15Hz, as in the top graph, the upward curve is going to look more extreme and that's going to impact the reading at 20Hz. But if you look at the locations of the peaks - and the values of the peaks like at just over 2kHz where the upper graph's peak does not go off the chart - they are the same.

The difference you're seeing at 20Hz is about how exactly one draws the arc of the slope from 10Hz to that peak at about 26 or 27Hz.

 

[dfuller]

There's a little bit of wiggle room at the raw driver level for impedance traces, FWIW. Voice coils aren't wound exactly identically every time. The port's peak may be influenced by air temperature and humidity, measurement technique, and level.

 

[MZKM]

Near identical.
I agree with @tmtomh ,the resolution and program used (plus some unit-to-unit variance) will cause the difference around 20Hz

 

[tmtomh]

View attachment 148067View attachment 148068


Near identical.
I agree with @tmtomh ,the resolution and program used (plus some unit-to-unit variance) will cause the difference around 20Hz

Thanks! And I did neglect to add the important additional factor of unit-to-unit variation that you and @dfuller have noted.

 

[skyfly]

Both graphs cut: 20 - 30 Hz, 0 - 20 Ohms.
The second one was horizontally scaled by 400% to make the width of the displayed pictures approximately identical.

The difference of lower 20 and upper 20 is very easily recognizable.

 

[Rick Sykora]

While environment and component variation can contribute to impedance measurement differences, often the differences are due to different drive levels used and/or the value of reference resistor used in the measurement apparatus.

 

[MZKM]

Both graphs cut: 20 - 30 Hz, 0 - 20 Ohms.
The second one was horizontally scaled by 400% to make the width of the displayed pictures approximately identical.

The difference of lower 20 and upper 20 is very easily recognizable.

View attachment 148069

View attachment 148075

Please see my previous response with images.

As stated, these aren’t different measurements of the same unit, that alone could cause this very minor difference. @Rick Sykora also correctly pointed out that even the voltage they tested at matters.

Load both my images in full screen and switch between them.

 

[skyfly]

MZKM's two pictures put together. The difference of lower 20 and upper 20 is not small.

 

[skyfly]

Does anyone have any evidence that the difference of lower 20 and upper 20 cannot be due to, for example, a design change without model name change by Spendor?

 

[MZKM]

Does anyone have any evidence that the difference of lower 20 and upper 20 cannot be due to, for example, a design change without model name change by Spendor?

Heck no. Again, just the fact that they are different models would be enough to explain the difference, let alone testing parameters and software.

 

[Rick Sykora]

…for example, go look at any of Erin’s reviews where he measures at both .1V and 2.83V.

For example, here is Elac speaker ones…

and these diffs are in the same environment on the same test equipment.