Temperature Control In Speaker Measurements

[Holton]

I recall Harbeth's Alan Shaw cautioning John Atkinson in the pages of Stereophile about the pitfalls and limitations of an individual attempting to measure speakers. Shaw argues that there is a reason real speaker companies invest so heavily in measuring facilities, or pay for the use thereof.

In that context, burried at the bottom of Amir's review of the Outlaw 2200M was a seemingly innocuous comment that, if true, begs for clarification and discussion. Here it is:
"Freezing in the garage measuring speakers. Need to get a nice supply of hot chocolate to warm me up".

Are we to understand that ASR's speaker measuremnts to date were conducted in an unheated garage during winter?

Here is a link to very introductory level comments by Shaw about the effect of temperature on speaker performance in the listening environment,
which has obvious implications for the testing environment. Screenshot attached.
https://www.harbeth.co.uk/usergroup...ing-room-and-why-it-should-be-controlled.124/

Does anyone here believe that temperature is not a controlled variable in the manufacturer's design/testing process?

Since the comment about testing in a cold garage was just posted, at least some of the speakers must still be on hand.
I would suggest that ASR retest whater speakers possible at Shaw's recommended 18-20 degree celcius temperature (which he claims is common design assumption). Publish the results, with an analysis of whatever differences exist (or lack therof).

Let's see if Shaw is mistaken, or if ASR should control this variable moving forward.

 

[scott wurcer]

I agree at an extreme this is very important, that's why scientific measurements specify a standard temperature and pressure (RH matters too to some degree).

EDIT - BTW that article is wrong 20C to 30C is not a 50% rise. Temperature rise in this application only makes sense in degrees Kelvin which is the basis for most physical processes.

 

[briskly]

Viscoelastic materials, such as in the flexible surround or extensional/CL damping, are rather temperature sensitive. These materials provide the most effective damping in a narrow temperature band, where the material stiffness is the most sensitive to temperature changes.
A change of about 20C can knock the damping layer clear out of its sweet spot. IME, the most obvious change is in the suspension stiffness rather than "break-up" resonances. This usually means a loss of bass/increased bass with lower/higher temps, as noted.

 

[headshake]

Washington state does not get so cold (unless you are high up, which he is not).

 

[restorer-john]

I think the impedance sweep on a woofer straight from the fridge vs off the bench would be quite different. Tweeters not so much.

I'll have to make space alongside the meat and vegetables for the woofer first.

 

[Mawclaw]

Temperature definitely plays a roll but I personally doubt that it has a measurable effect within a reasonable range, 50F to 90F for example. Also if it was a big issue I am pretty sure every professional loudspeaker company would tell you to buy an AC.

 

[Mawclaw]

Also I go into a lot of chemical plants/refineries for work and a lot of products that we think are held to exact properties are often just tested to a relatively loose range. I would propose that if you could hear vast difference between 40F then different batches/vulcanization times would also be audible.

Just my two cents. Not a chemE.

 

[A800]

Makes sense.
But what about humidity and atmospheric pressure?

 

[amirm]

As with "break-in," the mechanics of drivers can change but once you put them in the box, those effects become negligible. Alan can say all he wants. As long as he is not posting measurements, his comments are without value.

People listen to speakers at all kinds of temps. In hotel rooms at shows temps vary hugely as does humidity level. Never heard of one exhibitor complaining about the temperature impacting speaker performance.

Regardless, it is not much of an issue. The temp is around 55 to 57 degrees F or about 13 degrees C. Yes that is "cold" to us that live on the west coast. The garage is insulated and rarely goes up above 65 degrees F during summer. If a speaker operates worse in such moderate temps, then its measurements better damn it this way....

 

[amirm]

Also, the voice coil temp rises rapidly in use. That is bound to heat up the inside of the cabinet.

 

[Wombat]

Don't expect 'lab' results from an uncontrolled testing environment.

Near enough is probably good enough for general domestic use of loudspeakers.

 

[amirm]

Don't expect 'lab' results from an uncontrolled testing environment.

Testing is a heck of a lot more controlled that just about all other tests you see. This is precision instrument and far better than anyone crawling around measuring speakers indoor and out. Even posted anechoic chambers lack specifications much less temperature used during testing.

The measurement accuracy is 1 degree which if you were to repeat using anechoic chamber at 2 to 3 meters, would require thousands of measurement points.

 

[Wombat]

I was employed for ten years by this organisation: https://www.nata.com.au/about-nata/our-role

I do know more than a little re testing protocols.

 

[restorer-john]

Happy to throw one of the Andrew Jones Pioneer SP-BS21LRs I retrieved from Dad's place in the refrigerator and the other one in the oven (warm, say 60 degrees Celsius) and compare if it helps?

Not much use for them otherwise.

 

[amirm]

I was employed for ten years by this organisation: https://www.nata.com.au/about-nata/our-role

I do know more than a little re testing protocols.

I agree with little. Now study the Klippel system and get back to us...

 

[restorer-john]

Also, the voice coil temp rises rapidly in use. That is bound to heat up the inside of the cabinet.

With respect, that takes a considerable period of time and most of the heat is dissipated rapidly into the pole piece, the basket, the entire magnet structure and is often pumped out the vented pole piece or the reflex port.

Sealed, small, infinite baffle speakers heat up rapidly, I agree, but medium sized, bass reflex speakers with decent magnet structures take quite a while.

 

[Wombat]

I agree with little. Now study the Klippel system and get back to us...

Please set us straight with your inside knowledge. Klippel are very tight-lipped with detail on the WWW.

 

[Wombat]

I agree with little. Now study the Klippel system and get back to us...

Snitchy, condescending and unprofessional 'come-back'. You can do better than that on a 'science' based forum(your premise) re your testing.

You are claiming the accuracy of the Klippel rig in an uncontrolled test environment is unfazed. Please enlighten us all with Klippel support information.

 

[thewas]

Some of my work colleagues research on rubber behaviour, most such polymeres show both a temperature depending as well as a short and long term change of parameters. When such polymere is deformed also part of its mechanical energy is transformed into heat, heating it up. Also some parameters change after some hysteresis curves (what is known as "break-in") but also some return partially back after resting some time. Some holds also for other deforming parts like the loudspeaker spider.
All those reasons show that depending also on the long term output/music signal you will never be in the exact same hysteresis behaviour region, but a good engineer should chose the material and design the parts such that this change will relatively small compared to other sources of "distortion".
Thus I personally don't think that measuring a well-engineered loudspeaker with the Klippel NFS system in 13° instead of 22°C should make a big difference in the acoustic measurements, especially since it has a longer and continuous excitation. Of course extremes like temperatures over 60° or subzero should be avoided for such measurements. Would be interesting maybe also to measure the temperature difference of the rubber loudspeaker surround in the cold garage before and after the long Klippel testing with a infrared thermometer, especially for a small woofer which does large excursion I guess it would be measurable (plus also the heat from in the inside housing of the voice coil and crossover or amp heating it also up).

 

[Jon AA]

Please enlighten us all

Normally, the burden for "enlightenment" rests with the one making the accusations. You have not produced a single measurement of a single speaker showing the output variation vs. temperature when measured at the distance the Klippel measures them. Show us one of those so we can see the magnitude of the variations at the temperatures in question (is it 10 db or 0.1 db max deviation?) and a constructive conversation might begin.