Audibility of distortion in horns!

[Pano]

I had a pair of Pioneer bent horns that were driven by more of a cone driver than a compression driver. They are fairly common on the used market. Maybe I could dig up measurements, but I doubt I recorded the SPF of the test.

 

[peteleoni]

In many way those old cone tweeters from jbl and Pioneer were nicer than what followed.

 

[BYRTT]

It's hard for me to understand a 20 dB difference in 2nd harmonic at about 1 kHz between the two devices. This is not a small difference and something of note between the two different designs should be apparent. I am not familiar with either design so I have to ask if anyone else can explain it.

More info xrk971's Trynergy build combined SS 10F below.

My thinking is that it comes down to diaphragm or membrane motion. Displacement ties in directly with distortion (more motion, more distortion). How much of the diaphragm motion (in % of xmax) on a compression driver is used to achieve 114dB? xmax (or maybe xmech) on a compression driver, I am told, is circa 1mm before it bumps against the phase plug. On a full range cone driver designed for direct radiator duties, the driver has a lot more motion capability. My simulations show that the horn loading on the cone of the 10F is so efficient, it moves circa 55 micrometers (that is 55/1000 of a mm) to achieve 114dB. That is 2.2% of xmax for the driver - which could be considered the infinitessimal low motion limit and distortions due to non0linearities in suspension, magnetic fields, hysteresis, etc are essentially zero. Another way to look at it is to use a Taylor series approximation of the non-linearities, and in the limit that x tends to zero, the only significant term left is the linear term. 2.2% of xmax, can probably be safely called motion that is tending to zero. This may explain why Art noticed that the distortion was scaling linearly with SPL at these small cone motions.

As a direct radiator, the 10F already is one of the lowest distortion drivers. Now put a horn on it to reduce the motion and it gets better.

Trynergy build thread:
Swapped out the 3FE22 with the 10F/8424 (#784/785/787) - http://www.diyaudio.com/forums/full...ll-range-tractrix-synergy-79.html#post4508525
First measurements 10F/8424 (#798) - http://www.diyaudio.com/forums/full...ll-range-tractrix-synergy-80.html#post4509142
Polar response measurements (#840) - http://www.diyaudio.com/forums/full...ll-range-tractrix-synergy-84.html#post4512637
Distortion Measurements at 114dB (10F high passed via serial 6,6uF capacitor) (#1040) - http://www.diyaudio.com/forums/full...l-range-tractrix-synergy-104.html#post4539890
Distortion talk (#1045/1046) - http://www.diyaudio.com/forums/full...l-range-tractrix-synergy-105.html#post4540781


Driver info SS 10F/8424G00:
SCANSPEAK datasheet - http://www.scan-speak.dk/datasheet/pdf/10f-8424g00.pdf
KLANG+TONG review 2/2010 - http://www.scan-speak.dk/datasheet/reviews/10F-8424G00-KlangTon2010-2.pdf
KLANG+TONG review 2/2012 - http://www.scan-speak.dk/datasheet/reviews/10f8424g00-KlangTon2012-2.pdf
Hobby HiFi review 2/2012 - http://www.aos-lautsprecher.de/pdf/test/2012/10F8424HH.pdf


Zaph|Audio shared a test of 4 ohm version 10F/4424G00 and made below comment for power handling simulation.

I did some power handling sims and they show that this driver does not need to be treated like a dome midrange with minimal excursion. Even at 300Hz LR2 electrical filtering, calculated Xmax is not exceeded even with 256 watts. Power handling will always be a thermal issue with this midrange unless you try to treat it more like a small woofer than the midrange it is.

 

[turk 182]

Brytt
i'm not quite sure what your post is meant to convey or that Dr Geddes is still following the thread,

audibility of distortion in horns (or in general) is still a highly contestable and ambiguous topic that does require more work and understanding.
perhaps if distortion was not such a source of confusion and a devise topic we would be further along with methods of measuring it, qualifying it, and reducing it.

 

[gedlee]

The difference in distortion most likely is related to the compression ratio, the cone driver Trynergy has a very low ratio (around 2/1) while the DH1 (or most any HF compression driver) must have nearly an order of magnitude higher compression ratio.

Art

Hi Art

I was thinking about this last night. Even if air nonlinearity were the answer, which is hard for me to accept, it could only account for the difference in the 2nd order harmonics since air nonlinearity is purely 2nd order. There is still a 20 dB difference in the 3rd harmonic and this cannot be due to the compression ratio.

 

[xrk971]

Maybe worth repeating myself:

My simulations show that the horn loading on the cone of the 10F is so efficient, it moves circa 55 micrometers (that is 55/1000 of a mm) to achieve 114dB. That is 2.2% of xmax for the driver

It all comes down to tiny motion = tiny distortion - simple as that.

 

[Pan]

Hi Art

I was thinking about this last night. Even if air nonlinearity were the answer, which is hard for me to accept, it could only account for the difference in the 2nd order harmonics since air nonlinearity is purely 2nd order. There is still a 20 dB difference in the 3rd harmonic and this cannot be due to the compression ratio.

But what happens to the asymmetric nonlinearity when two relatviely similar air springs are positioned one in front of, and one behind the moving element?

The extreme and not very realistic case would be a driver mounted with one closed box in front and one behind the driver. The net nonlinearities will be symmetric = no even order distortion products.

 

[Eldam]

Is this happen when the two closed boxes are rooms with a speaker in the wall between (infinite bafle load) ?

 

[peteleoni]

How do you then explain the complete lake of interest in this subject by people like Tool and Olive?

Complete lack of interest? Maybe it is the least offensive because it is well controlled now.
Klipsch heritage speakers are probably only still selling because those drivers move very little and produce so little distortion. I know I can barely stand to listen to my 801s at anything approaching loud. Not so with good horn loaded even with a bad mid like the khorns. And yes I have great amps. I appreciate your study Dr. Geddes but I am certain the last shoe has not fallen on audibility of non linear distortion.
One clue is how a good horn loaded speaker never sounds as loud as it really is.

 

[turk 182]

ok those last two posts are a little out to lunch.

and i still can't find any information that confirms that air non linearity is purely 2nd order.

correction post 407,408
peteleoni's post appearred as i was posting

 

[turk 182]

wait is that 801's reference mean that Bose 801's are somehow a reference for the audibility of distortion in horns?

 

[xrk971]

I think reference was to B&W 801's?

 

[peteleoni]

I should say so!! No Bose near this studio.
B&W Matrix 801's per Abbey Road and countless other mastering houses and the intrepid Don Keele, till he created his CBT's Bose....huff huff ���� The dawn of Bose created the "Man Cave" and reduced testoterone levels worldwide by 18.5 per cent. ��

 

[badman]

That's a terrific thing to say Pete, re: the dawn of Bose.

I may well repeat it without attribution as my memory is terrible. I trust you'll forgive me 🙂

 

[Pan]

Is this happen when the two closed boxes are rooms with a speaker in the wall between (infinite bafle load) ?

With volumes of that magnitude you have practically no effect of air loading. For the air load to be a factor you need a volume small enough to have low compliance in relation to the drivers suspension. Or low enough compliance to shift the system fundamental resonance up in frequency.. effectively turning mass controlled working range into compliance controlled.

 

[peteleoni]

That's a terrific thing to say Pete, re: the dawn of Bose.

I may well repeat it without attribution as my memory is terrible. I trust you'll forgive me 🙂

Damn truth. Those little boxes staring showing up. Men started wearing bowties and sitting
In their cars wth the stereo cranked smoking cigars and listening to (Rush.... no the one with drummer Niel Pert) From there he was basically not allowed to have male odor of pheremones anywhere but the garage. Yep those nasty little oversold "Waf boxes" have doing yoeman's work for the feminization of the westen world.
Gentlemen take back your Dens give them a maidens cave to listen to Aaerosmith or Yanni or thst infernal suprano sax bitch guy.

 

[turk 182]

still can't find any information that confirms that air non linearity is purely 2nd order.

and what's horn loaded on B&W 801's ?

 

[peteleoni]

On another note what is stopping us from throwing three mid cones around the shallow of some these ready made OS shallow waveguide thingies perhaps making a bespoke unit with all frustrums done and including mounting guide

 

[turk 182]

loss of pattern control at lower frequency due to lack of overall size comes to mind.

what's your beef with Rush?

 

[gedlee]

But what happens to the asymmetric nonlinearity when two relatviely similar air springs are positioned one in front of, and one behind the moving element?

That's true to the extent that the nonlinearities on the front and back are the same. But the air compression on the back is no where near what it is on the front, so in realty this will not occur.

Complete lack of interest? Maybe it is the least offensive because it is well controlled now.

That is precisely the point that I have been making. The nonlinearities are well controlled and as such don't interest those of us who are working at the for front of the problems that remain. Even if NLD has not fallen below audibility it is not a significant factor any longer.