Horn/Driver Comparisons

[whgeiger]

WW,

Comments follow your text.

"When using a horn as a test jig, this is not happening; e.g., horn artifacts, at the bottom end, are manifesting themselves as distortion products due to the driver being operated below the horn’s effective pass band, which begins at least one to two octaves above horn [fc]."

I thought we discussed this already. The point is, those artifacts will affect all of these drivers in the same manner.

Sometimes the only way to get a point across is through repetition where the issue is recast in different words. Will try one more time:

Since our purpose wasn't to determine the absolute value for each driver but rather the relative performance between them

1) If the ‘what’ to be discovered and measured is hidden or suppressed under horn artifacts, then the ‘what’ remains un-observed (masked) or at least misrepresented.
2) Using either test jig (horn or PWT), the measurements are relative because
a) the unit used is dB (which is a ratio), and
b) it is cast with respect to the parameters frequency and driver.
So, in either case, the data set for each driver contains only “relative” measurements not “absolute” ones.

I think you are going over board.

That is your opinion, so be it.

Of course if you would like to make us a PWT, I'll be happy to use it in future tests. It would be nice to have one for 1", 1.5", and 2" drivers.

Cost more to ship a completed unit than to simply perform a local build. As the procedure for building them is sufficiently ‘Gerberized’, the DIY route would be the most cost effective approach to acquisition. For most, if not all, JBL compression drivers, a PWT power response sweep is published. Some are also provided in the referenced articles [1] & [2] and JBL authored technical notes.

BTW: your link didn't work... I did try to check it out.

A complementary copy of the AES standard [1] may be obtained at the following (AES revised) URL: http://www.aes.org/publications/stan...rtesy.cfm?ID=1
If you want a copy of the companion article [2], will be more than happy to e-mail it to you.

Regards,

WHG

[Ian Mackenzie]

"When using a horn as a test jig, this is not happening; e.g., horn artifacts, at the bottom end, are manifesting themselves as distortion products due to the driver being operated below the horn’s effective pass band, which begins at least one to two octaves above horn [fc]."

I thought we discussed this already. The point is, those artifacts will affect all of these drivers in the same manner. Since our purpose wasn't to determine the absolute value for each driver but rather the relative performance between them, I think you are going over board. Of course if you would like to make us a PWT, I'll be happy to use it in future tests. It would be nice to have one for 1", 1.5", and 2" drivers.

Widget

BTW: your link didn't work... I did try to check it out.

Excellent work Zilch and The Widget,

I agree and I think it makes more productive discussion to look at this from a more practical perspective. Why would you attempt to use the horn all the way down that far anyway unless the woofer didn't have the legs? Besides the artifacts if there are any would be normally be attenuated by the high pass crossover filter..that's common sense. In anycase we can't do anything about the garbage down there so its a dead issue imho.

Of more interest as illustrated by these nice curves is the behavioural relationships b/n the various horns and drivers and their interface with real world acoustics. I would like to see some subjective appraisals at some stage, altough given the reasonable cost of these horns no doubt we can find this out for ourselves!

Ian

[Steve Schell]

Hi Ian,

You asked "Have you ever considered a conical/ differaction hyprid horn?"

Not really. My understanding of diffraction horns and how they operate is far from complete, but I've never cared for their sound. I once bought a pair of the big JBL Biradials, and couldn't leave them in the system for more than a day or two. I have always regarded as suspect horns that have large and/or sudden changes in cross sectional shape or flare rate; they strike me as sounding confused and harsh somehow.

You may be familiar with the work of Dr. Earl Geddes. His theory is that Higher Order Modes (HOM) are very problematic. He regards HOMs as any wavefront that bounces off the horn walls and creates a delayed wave launch, much like edge diffraction on a flat baffle. He feels that these closely timed reflections confuse the sound and disrupt imaging; this all makes sense to me. In his horns, or "waveguides" as he calls them, he introduces a bit of curvature in the initial throat section to establish a spherical wavefront. The waveguide is basically close to a conical horn, with a large radius applied to the mouth to control reflections.

You can read an explanation of his theories here:

http://www.gedlee.com/downloads/Cum%20laude.pdf

J.M. LeCleach has also attacked this problem and the flare profile he has developed seeks to create a pure spherical wave launch. I don't understand the math behind his flare, but the mouth shape indicates he is attempting to control mouth reflections. His horn can be seen here, near the bottom of the page:

http://www.arduman.com/aa/Sayfalar/l...h/lecleach.htm

Discouraging HOMs does sound to me like a great way to make a horn. The relative lack of HOMs in even rudimentary conical horns likely contributes to their clarity of sound.

While I was out surfing for links, I also ran across this golden post from Tom Danley. Whenever he is talking I try to be listening...

http://www.audioasylum.com/forums/hu...ges/85136.html

[Zilch]

2435HPL's are back from JBL - "Cleaned gaps and added ferro-fluid." Passed final test. We'll see if they pass the Zilchster test, now.

[speakerdave]

I and my 2435's are waiting patiently for the result. Maybe it would be worthwhile for me to send mine to summer camp at Northridge.

David

[Mr. Widget]

WW,

Comments follow your tex....

I guess I missed these comments.

Bill,

I must say that I agree with every one of your points. That said, I am reluctant to build a PWT since any tube I made would always yield results that were at least partly clouded by the question of whether or not I got the foam or fiberglass damping load properly tapered and therefore the results would still be only relative and not truly absolute. I agree that it would be closer to absolute than the use of my 500Hz horn... but the horn is here and ready to go... the PWT isn't. I may venture down that path at some point and make one for kicks. Unfortunately there are only so many audio hours in a day.

Ideally we will be able to get Zilch's drivers back over here to do a before and after measurement so that we can see exactly what may have been "improved" by the cleaning and reapplying of the ferro fluid. Zilch's subjective results will also be interesting.

Referring to questions about the validity of measurements made on the Ring Radiator Thread. I agree with those that said or implied that our onboard measuring devices (ears) are the ultimate judge of quality. That said I found the measured response that we posted are for the most part a validation of JBL's claims and not a contradiction. I guess there is interpretation even in objective measurements. As for which devices are superior as was pointed out by some, it primarily depends on your intended application and also to a some degree personal preferences.

As for whether or not one device measures better than another is obviously not entirely objective either. The curves that looked best to me were not the curves that looked best to Zilch. We approach the design problem differently and have different goals and therefore come to different conclusions while looking at the same objective data. There rarely is one right answer, though it is easy to forget that.

My understanding of diffraction horns and how they operate is far from complete, but I've never cared for their sound. I once bought a pair of the big JBL Biradials, and couldn't leave them in the system for more than a day or two...

I must say that based on my own experiences (both objectively and subjectively) I agree with the notion that there is a problem with sudden changes in cross sectional flare rate. In my limited experience with horns that feature an abrupt change in flare rate like the Biradial 2344 and H9800 I agree that there seem to be qualities in their reproduction of music that I subjectively find problematic. On the other hand the 2397 diffraction horn I find reasonably pleasing and also a fairly good imaging horn in a room with adequate acoustical controls. (The very wide dispersion can be problematic in live rooms.) Objectively this diffraction horn has a series of notch cancellations most likely caused by the parallel surfaces. Fortunately it is much harder to notice a notch than a peak so subjectively the horn sounds reasonably smooth. Other horns that may measure "smoother" may have slight peaks that can be objectionable. This again is where the "ultimate" test must be done by ear and is obviously also a matter of personal preference.

Widget

[Zilch]

They sound fine. They have the flatness Mr. Widget likes, but they don't have the HF extension of 2431's. The two units are different at the high end. I don't pretend to hear any difference between Al and Be.

I conclude JBL doesn't spec 'em up there. We'll run again on CLIO to see if they are really any different from before they went in for service. A 2430H is on the way from tomp787 to complete the 243x picture.

#5 (top) and #6, both on PT-F waveguides, biased "AM" crossovers:

[Mr. Widget]

Here are plots of Zilch's 2435s after being refurbished at Northridge. The red and blue plots are of the two drivers before they were sent back to JBL. The green plot is of one of them after the servicing. When we saw the two drivers performing so differently we suspected that they were out of spec. They were also suspect since they had been an eBay purchase.

[Mr. Widget]

Here we see both of the refurbished 2435s are now nicely matched. We can also compare them to a pair of nicely matched 2431s on the same 100° by 100° PT-F wave guides. The 2435s appear to be slightly smoother in response and also about 3dB more sensitive.

[Mr. Widget]

Forum member tomp787 was kind enough to lend us one of his 2430 drivers. Here we see it compared to a 2431. The 2430 has a different diaphragm giving it a lower FS and a bit less UHF extension. We can see the effects on the HF extension, but on this horn the lower frequency response is completely dictated by the horn cut off.

[Mr. Widget]

Here is the 2430 on the PT-F and PT-H 100° by 100° horns and also on the PT-F 60° by 40° horn.

[edgewound]

Here we see both of the refurbished 2435s are now nicely matched. We can also compare them to a pair of nicely matched 2431s on the same 100° by 100° PT-F wave guides. The 2435s appear to be slightly smoother in response and also about 3dB more sensitive.

Wow....3dB is huge!!! check out that knee at about 18khz? to maybe give the 2431 an impression of better UHF response...which one sounds better to you? have you done critical listening? female vocals? violin? sax? reed breath sounds?

Edgewound

[Mr. Widget]

Here we compared the 2435 with the 435Be, the 2430 and 2431 drivers all on the H9800 horn. The response of the 2435 and the 435Be were essentially identical. The effects of the larger back cap and lack of ferro fluid of the 435Be was not noticeable in the frequency response measurement. It is quite possible that a subjective listening test would show a difference, but from strictly a frequency response measurement, they do appear very similar.

This series of measurements has shown us that there are potential hazards in buying used and untested merchandise, luckily JBL was able to come to the rescue. It also shows us a nice comparison of almost the full line of JBL's current crop of 1.5" exit 3" diaphragm drivers.

Widget

[Mr. Widget]

check out that knee at about 18khz? to maybe give the 2431 an impression of better UHF response...

That spike at 18KHz is the result of the mass break point. All compression drivers exhibit that type of curve to some extent. Some listeners complain about the distortion or hard edge that is heard when a driver is run out to the resonant peak caused by it's mass break point others do not seem to be affected by it. If it is high enough in frequency it is typically less of a problem.

...which one sounds better to you? have you done critical listening? female vocals? violin? sax? reed breath sounds?

With the exception of the 435Be I have not listened to any of these drivers critically.

Widget

[B&KMan]

Hi ,

I read with a very great interest the series of tests which you expose
here... It is an enormous work with the implementation of much of resource.



after having read and observed the different plot ones I believe
that it would be very advised for more objectively seizing the
difference which you expose the response in distortions of different
the driver. I believe that the level of distortion of the harmonics, 1,3,5, with step eloquence would expose new light of the differences of each driver...

WHAT think about it???

Jean.