Constrained throat riddle

[hetmann]

Does anyone have the Bruce Edgar articles about his studies of the Klipschorn throat? There were two published in Speaker Builder magazine. The first was in the 4/90 issue and the second in the 6/90 issue.

------------------

Klipschorns and Moondogs

[Klewless]

hetmann,

I have the two articles you mentioned.

------------------

John P

St Paul, MN

[hetmann]

This forum is great! I've already received the second of the two articles by email. Thank you William. For you builders out there, it appears as though you can use either a 3" or 6" throat with no detectable difference in response.

------------------

Klipschorns and Moondogs

[WMcD]

Here is a paragraph from the first article (4/90) and corrected drawing (5/90). The second article (6/90) retracts much of what was concluded in the first but does not comment on the data supplied by Hunter.

Maybe Mr. Hunter can comment.

Note: I originally pasted the URL into the text here but the image is too wide. I'll just leave it as an attachment rather than post a smaller version.

Regards,

Gil

This message has been edited by William F. Gil McDermott on 04-19-2002 at 05:20 PM

[Klipschguy]

Funny, looking at the graphs, it looks like the earlier K-horn has a better response.

Andy

[johnnyp]

Nice post, Gil.

I am surprised Dr. Edgar/Hunter didn't perform (or include) the frequency comparison from 200hz down to 35 hz. The huge trough at 250hz of the 3" x 13" restrictor slot is disconcerting. If this is due to frequency cancellation, what is the 1/4 length measurement of 250 hz?

------------------

John Packard

This message has been edited by johnnyp on 04-19-2002 at 08:29 PM

[WMcD]

Here is a more compact form of just the graph so we can look at it and discuss things.

Wavelenght = velocity / frequency.

Velocity is 13500 inches per second.

So, 1/4 wavelenght of 250 Hz is 13.5 inches. You can check my mental math.

This message has been edited by William F. Gil McDermott on 04-19-2002 at 09:21 PM

[WMcD]

I note the dip at 250 Hz is there to some extent with or without the restriction.

Why would the data presented not go down to the lower limits of the K-Horn? The measurement of the K-Horn bass unit in its full bandwidth and natural setting the corner of a room is a challenge.

I.e. it is simple to put a measurement microphone in one position in one room and run a curve with the most accurate equipment. But that doesn't tell the story for all rooms, or all positions. The lower end is very hinky.

What to tell people in all honesty? Run curves in 5 rooms in 5 positions, average everything? You can take only the peaks, or smooth it. Then it doesn't tell you what is going on in your room.

I suspect that Mr. Hunter ran his tests in conditions where the reponse above 200 Hz was easy to obtain, because these were the frequencies of interest. Maybe in the anechoic chamber at Hope.

The results are the only ones I know of published about the effect of the restriction of the throat. Again, I'd point out that the size of the duct beyond the restriction is effictively 6 x 13.

Unless Mr. Hunter or Dr. Edgar can give us some further information, I'm a bit suspicious about the corrected curves because they are contrary to comments by Dr. Edgar in his article.

Gil

[johnnyp]

Your math is good, Gil.

Then, 1/4 of 250 hz = 13.5" and 1/2 of 250 Hz = 27". Hmmm, it is not clear to me what would be causing the cancellation.

If the graph is valid, what do you think causes the poor response at 250 hz?

I have wondered how they determined the frequency response of the K-horn. If one stands on axis with respect to the squawker and tweeter, say 1 meter out and 1 meter vertical, the bass output is more subdued than when one approaches either wall. At first I thought this was because the output of the squawker/tweeter was "hotter" on axis, but then Idisconnected them and tried the experiment again and the phenomenon persisted.

------------------

John Packard

This message has been edited by johnnyp on 04-20-2002 at 12:03 AM

[WMcD]

Well, I dunno. Scratch head. How come I'm suppose to know everything? I should keep my mouth shut and keyboard quiet.

I'm not questioning the overall curves, just which is which.

We may be hunting for witches for the source of the dip.

One thought is that the we can contemplate the path of the sound along folded geometry of the horn. It goes up and down two ducts behind the front face, making a turn, then to the back to the tail board, then up to the front.

The overall depth of the K-Horn is 24 inches. So two of these paths, the last two, are close to 27 inches. The turns might create some reflections and gang up on each other, impedance mismatch along the transmission line.

Gil

[WMcD]

Hmm, looking at things again, the dip(s) is a bit above 250 Hz and so might correspond to the 24 inch baseline of wavelength computations, rather than the 27 inches. So it could make sense.

Gil

[djk]

I would consider the LMS outdoor response, 1/8th space, microphone centered on axis, at a distance of 2 meters, to be the most accurate data (JAES vol. 48, no. 10, p. 926, fig. 9). This data is un-smoothed and looks like the SB curve identified as 6 X 13 '50s K-horn in the "corrected" text. This leads me to believe that the original text was correct. The peak-to-trough on the current Klipschorn from 100hz to is about 9dB, the revised Klipschorn (dual 12") is about 13dB through the same region (fig. 8, ibid).

[John Warren]

the 3x13" baffle board slot is a consequence of the k33e woofer. decreasing the slot area increases the shunt capacitance at the throat making the slot/throat combination more effect as a low pass filter. if a "regular" woofer is used the 6x13" slot is needed. the slot attenuates the upper end output to make the

low cost (small motor) K-33E woofer sound like its putting out bass.

looking between 200 and 400Hz is looking at the the tree tops and missing the contour of the mountain range. Look at the output above 200Hz and compare it to the output at 60Hz and below. 13+db DOWN.

The small slot throws away efficiency.

This message has been edited by John Warren on 04-20-2002 at 06:47 PM

[hetmann]

Sometimes ignorance is bliss. Sometimes its just annoying...

Ok, here it is. I have K-33-E woofs and I am building khorns from scratch. I have two sets of plans, 1946 PWK show the 6" slot and recent vintage speakerlab plans show a 3" slot.

What size slot should I cut?

------------------

Klipschorns and Moondogs

[Misery Guest]

K-33 E NEEDS THE 3" WIDE SLOT!!!!!!!

This message has been edited by Misery Guest on 04-20-2002 at 08:16 PM

[johnnyp]

What an ugly graph. Reminds me of the Sawtooth Mtns in central Idaho. How do the good folks at K&A get away with calling it a +/- 5dB speaker?

------------------

John Packard

[Al Klappenberger]

Everybody!!!

That curve posted by John is of a test of my prototype crossover network in a Khorn that was kicking around the shop at the Klipsch lab. It was posted for a short time on my web site. I WAS ASKED TO REMOVE IT BY KLIPSCH BECAUSE IT WAS NOT CORRECT! It is NOT a curve of a good Khorn! Don't not take it as the last word for ANYHTING! It is only one of a group made to avaluate the various tap setting on the transformer. It shows what happens when the squawker is OUT OF PHASE! AGAIN: It is NOT a good curve!!!

Al K.

This message has been edited by Al Klappenberger on 04-21-2002 at 08:22 PM

[WMcD]

Thanks for the data djk.

I've often commented on some of the implacations of the AES article. It tells the story of the creation of the present Jubilee. There are graphs for the current K-Horn, presumably with the 3 x 13 restriction, a first prototype for the Jubilee, and then what is probably the current Jubilee.

From the article, and as djk points out, the curves are made with the bass units in a corner outdoors. I can only visualize a building with such a corner. Two exterior walls coming together to form a room type corner; the corner is open to the outdoors.

It seems to me that this gives a uniform testing environment. For good or for bad, it elimiates the room gain in the bass. So the K-Horn response at the bottom end is poor. I.e better response as we see on spec sheets from Hope and the Richard Hyester review from Audio magazine which rely on room gain.

I am guarded in making any definitive statements in this area. Please excuse that.

I think it is not a matter that the K-33 is a low cost driver. Certainly PWK is more than capable of putting any driver in there.

Dr. Edgar was a big fan of very low Q drivers with a Qts less than 0.2. and high Fs. That is important for the high end output, in theory. You have to study an AES Preprint by Don Keele to get the story because it also involves the Fs of the driver. The high end output 3 dB down point with an optimal throat size, not restricted:

High (3dB down)= 2 Fs/Qts.

Note, that is the output of the driver itself; not after it comes out the horn, into the room.

The K-33 has a higher Qts and lower Fs than what Dr. Edgar would like. That gives, in theory, poor high end response from the driver. However, there is not anything I can find specifically on the issue of Qts and low frequency performance in a bass horn.

Again, a guess. I think the higher Q of the K-33 of about 0.35 is there to give an improved response where the load from the bass horn is falling at the bottom end. Specifically, a high Q driver like the K-33 is more "floppy" and under damped. Box enclosures have poor low frequency response when the driver is over damped electrially with a low Qts. Same when it is looking into a classic, pure exponential horn, like the K-Horn.

So you can see the initial issue. A high Q driver like the K-33 could be necessary to give good low end response in the horn. Remember, as the impedance created by the horn falls at cut off, it is looking like there is no horn loading. So we need a driver which would be floppy enought to give a good box response.

But, same issue, that higher Q is not good for the high end output of the driver.

So, is there a way to make the bass horn have better "gain" at the high end. Looking at the progression of designs for the Jubilee, it seems that front facing output ducts are the answer. Here the ducts don't fire along the side walls like the K-Horn; rather forward, like the LaScala and Belle. Hence you get some "beaming". This means, some gain in the higher frequencies.

But lets go back to the curve about the restriction, and the LMS curve both for the classic K-Horn. Let's assume the graph I posted is correct, as corrected. Let's also assume the LMS curve does use a throat restriction.

Taking them together, it looks as if the restricion allows a relative improvement of some merit (some merit!) from 275 to 400 Hz. You have to imagine what the LMS response would look like without the restriction. Arguably, the restriction gives some smoothness and extended response in a given plus or minus window.

In painting this mental picture, we have to appreciate that the restriction knocks down high end response in the band pass in the Edgar curve. But that means it is closer in level to the low end. So it allows a better plus or minus 3dB.

What to do when building. I'd say keep the 3 x 13 slot.

I built my horns with a 6 x 13 in both the motor board and the piece it mates with. Then I went back and replaced the motor board with one with a 3 x 13 slot. So you can make two motor boards and compare and contrast. I'll never have the time.

Food for thought.

Gil

[John Warren]

Al-

The woofer output is only of interest here, the curve above is a resonable representation of that. But, to address this, I have the anechoic response of the woofer wired directly to the amp without the inductor.

I'll post it here and delete the one above.

This message has been edited by John Warren on 04-22-2002 at 06:35 PM

[Stig]

For those wanting another opinion go here....

WWW.artandaudio.com/constrained.html

don't shoot me I'm only the piano player,

Stig