Two-way speaker system, WAY horn-loaded

tubino

A guy with too many tubes
I'm on the verge of putting together a horn-loaded two-way system. This is (almost) all new to me, but I thought it might be worthwhile to share and document. I'm not starting entirely from scratch, though. I have a pair of cabinets that stand 62" tall x 38" wide x 25" deep. They appear to be built almost exactly to these plans:

212422188-L.jpg


So far I have:
  • yanked out the old odd mismatch of drivers
  • mounted the cabs on casters to make it easy to move them around till I'm done dinking
  • cut new baffle boards to use standard 15" drivers
  • put in wiring to bi-amp or use an external crossover for dual concentrics
  • tried several drivers, including Tannoy dual concentrics
  • been wowed by the results! :banana:
  • vowed to see if I can get better still by learning how to maximize the cabinet's loading for particular drivers
  • determined that I should try a dedicated 15" and high-quality horn-loaded compression driver on top to see if I can outdo the Tannoys
  • acquired or am acquiring the parts to accomplish that last item (details to follow)
  • built first- and second-order basic crossovers
  • started getting a handle on BASIC formulas for determining throat dimensions
 
Parts I will have to try

To date I have tried Tannoy Gold 15's, Tannoy K3826, Tannoy 3833, and some odd 15" woofers. All the Tannoys did very well. High Qts woofers from consoles do not do well. A random EV 15" did pretty well. A JBL E-130 was a disappointment, BUT that might have a lot to do with the cabinet's throat and compression chamber dimensions! (more on that later)

What I have to try (or will get very soon):

On the bottom:
  • JBL 2205J (Fs=30, Qts=.2, should be great)
  • Altec 515E (similar specs, should be great)

On top:
  • Edgarhorn 500hz salad bowls, with...
  • JBL 2441 (on the way)
  • Great Plains Audio 399 (to be borrowed)

I'll try a crossover point of ~800, at first anyway, with passive components.
 
Doing the calculations

The BLH cabinets are supposed to load down to 30hz, and they really might do it, when dialed in! They are enormous, and I get incredible low clean clear bass to 110db+ with MINIMAL cone excursion on drivers that are loading properly.

But I was pointed in the direction of UltraFi Times issue #8, in which John L. Hasquin II walks through design of a back-loaded horn for the Fostex FE168E Sigma. Larry Moore suggested I use the formulas to determine if the throat and chamber dimensions are right, with a strong hint that they quite likely too large.

Larry was right. Hasquin boils down the formula, and gives it like this:

The old formula to calculate throat area (St) is (0.8*Fs*Qes*Vas) = St. Note that St is in square inches and Vas is in Cubic feet. This normally will get you close. For the FE 168E Sigma, this works out to 9.2 square inches.

So I did the math for some of the drivers I have or will use, and came up with some numbers I'll put in the next post. But what I'm starting with, in the existing cabinets, is close to this:

251595621-L.jpg


In other words, I think the existing throat constriction is about 3" w x 16" h on each side = 48 x 2 = 96. I need to measure to confirm this, but it's in the ballpark.
 
Looks interesting. So the existing cab allows for one 15" driver right? I would think that the Altec 515 would be ideal seeing as how Altec used it in some of their horn loaded cabs and it's Q. How the other parameters work into that formula I have no idea.

I have always wondered how you would go about designing a horn loaded cab around a particular driver or vica versa.

RC
 
Specs of good candidate woofers

Looks interesting. So the existing cab allows for one 15" driver right? I would think that the Altec 515 would be ideal seeing as how Altec used it in some of their horn loaded cabs and it's Q. How the other parameters work into that formula I have no idea.

I have always wondered how you would go about designing a horn loaded cab around a particular driver or vica versa.

Yes, the cab was designed for one 15". (Steve Deckert's version, the Decware Imperial, is designed to handle as much as 2 x 18", with larger compression chamber ... maybe throat too, not sure.) I got the cabs with Stephens Tru-Sonics, which are cool, but a mismatch. As you indicate, low Qts is wanted here, and that they weren't. Here are the numbers for good candidates, with the throat size (St) in the last column, using the formula given above.

---------------Fs Qes Vas Calc St, sq in
Tannoy
Gold 15---------26 0.22 12.5 57
---------------------------------------------------------
Tannoy
DU386/K3828---22 0.2 17.2 61
---------------------------------------------------------
Altec 511E-----26 0.24 20.4 102
---------------------------------------------------------
JBL 2205J------30 0.22 10.5 55


So if I haven't screwed the pooch too badly, the range of recommended throat sizes goes from ~55 to 102 for my drivers -- IF the drivers operate to original specs! If Fs has gone up, up goes the St calculation. I will use Woofer Tester 3 to check that out.

It looks as though my estimate of 96 sq in for the throat is pretty close to what's wanted, or at least being 1.5 to 2 times the lowest of these low Fs/low Q drivers, is consistent with the drivers recommended by Jensen in the plans (namely Fs of 45, with up to 55 giving "reasonably satisfactory results").

So it may be that I make some gains by reducing chamber and throat size a bit, for the lowest Fs/Qe/Vas drivers. Empirical tests will show!

What I DO know from what I've done so far is that there is absolutely nothing subtle about the difference between the horn loading vs not. At one point I had a Tannoy on one side and a borrowed professionally-redone JBL (D-130, not 2205J) on the other. I was using mono amps, and had the Tannoy side playing while I popped in the JBL. I flipped on the JBL amp, and in 2 seconds I noted 2 things: the JBL went up higher in frequency as a direct radiator (I had no coil or crossover on it) than I expected, and the horn wasn't loading with it.

Now, this JBL had been redone, and might have been the only driver I was using that was close to original specs. I'll measure free air Fs etc. and see if maybe it's the only one so far that really wanted a smaller throat. It would really be something if I tried it again with throat and chamber adjustments, and heard it really using that horn.
 

Attachments

  • throatcalc.txt
    348 bytes · Views: 143
I wonder if a 15trxb would work, especially for mono if i cant find another driver.
 
I wonder if a 15trxb would work, especially for mono if i cant find another driver.
Got specs? This one works for many:
http://www.thielesmall.com/database.asp
EV DL15: Qe = 0.3, Fs=42, Vas = 195. (101db/1w/1m!)

Those are the kinds of numbers Jensen had in mind, probably would work, but personally I would feel better with lower Fs and Q. Easy for me to try it, if I had the driver though.
 
I wonder what happens with a 12trxb..

12trxb

Re = 6.7
Sd = 535 cm2
Bl = 8.76
Vas = 167 liters
Cms = 411 uM/N
Fs = 45 hertz
Qms = 5.9
Qes = 0.75
Qts = 0.665
Pe = 65 watts
Xmax = 3.3mm
1w SPL = 96.22 db*
 
I wonder what happens with a 12trxb..

Fs = 45 hertz
Qms = 5.9
Qes = 0.75
Qts = 0.665

I think that's very similar to one 12" EV co-ax I tried... and I believe I know the answer: BOOMY BASS. It's not a good match, with that high Q.

There's a reason why I'm picking drivers that have been used in big horns before: they work! When Tannoy pulls out the stops with those drivers, they put them in big folded BLH.

Today I started up the learning curve with the program HornResp. When I feel reasonably secure that I know how to model this cabinet in there, it will be very easy to simulate response with different drivers. Here's a first crude attempt.
 

Attachments

  • HornResp1.jpg
    HornResp1.jpg
    26.9 KB · Views: 82
Last edited:

Nice work there, Paul! I have a Tannoy GRF-R that is really quite similar in size to yours, and though designed for a 15", it works quite well with an Altec 414. Look, I even put the same Altec 902-8 on top (thanks for the loan, Ron!)
211520295-S.gif
227582972-S.jpg


I thought about putting that EBP figure out there... it all comes out eventually.

I probably should point out that I'm getting really phenomenal sound, esp. with the Tannoy Golds. I could just stop there and be really, really, happy, and just move on to reinforcing the cabs and prettying them up. But before I call it quits, I want to see if I can go up one more notch with dedicated drivers, with a suitable HF horn. I've heard a friend's setup with JBL 2226 and GPA 399 + short horn, and there's something about that big compression driver...

But I think by learning to tweak these cabs a bit, I will be able to get extended flat response. Do you know what kind of frequency response you get with the combination of direct and horn output with the scoops?
 
Last edited:
man. i have a carpenter right in my subdivision...! ohh the temptation for HPD's!
 
Nice work there, Paul! I have a Tannoy GRF-R that is really quite similar in size to yours, and though designed for a 15", it works quite well with an Altec 414. Look, I even put the same Altec 902-8 on top (thanks for the loan, Ron!)
211520295-S.gif
227582972-S.jpg


I thought about putting that EBP figure out there... it all comes out eventually.

(snip)


Those are actually Jaymanaa's speakers. I just did the article.

The 15" scoops typically go to about 40-45 hz, but very solid down to there. If you stack them in groups of four (outdoor/PA use) they go down to about 35 hz.

This is from an old JBL 4530 scoop ad:

attachment.php


That would be a fairly typical response. Notice the up and down response, peaks and dips, where the woofer is in phase and out of phase with the sound radiation from the mouth of the horn. That is typical response of a "transmission line" speaker cabinet. That is the reason for the damping material in transmission lines, and the damping gets rid of those peaks and valleys in the upper end, while acting as a low pass filter, and still allowing reinforcement of the bass under 100 hz.

The Eminence/Fane/JBL type plans that Jay used are a little simpler to build in the upper end, not having that labyrinth just behind the woofer.

With 811b horns I suggest crossing over at 1000 hz - 1200 hz, depending on the horn driver.

The requirements for woofers are spelled out in that Dancetech article. There are plenty of woofers that will work.
 

Attachments

  • Scoop Response Curve.jpg
    Scoop Response Curve.jpg
    56.8 KB · Views: 2,490
That would be a fairly typical response. Notice the up and down response, peaks and dips, where the woofer is in phase and out of phase with the sound radiation from the mouth of the horn. That is typical response of a "transmission line" speaker cabinet. That is the reason for the damping material in transmission lines, and the damping gets rid of those peaks and valleys in the upper end, while acting as a low pass filter, and still allowing reinforcement of the bass under 100 hz.

The Eminence/Fane/JBL type plans that Jay used are a little simpler to build in the upper end, not having that labyrinth just behind the woofer.

Did you know that there is a company selling those style of scoops on ebay?

Maybe this is a minor terminology nit-pick, but why call it a transmission line rather than simply a backloaded horn? I thought that the expanding flare would make that a BLH, where the TL is intentionally a narrower-bandwidth device, more like a port than a flared mouth...

I understand about the peaks and dips (notice I posted a similar shot above), but how do those vary with distance from the speaker?

One interesting distinction between the scoops and the giant BLH I'm monkeying with is that the BIG cancellation dip is moved down significantly in the longer horn (= longer wavelength that arrives out of phase), with a lot of output (over 100db) at 40 hz. The longer horn also seems to greatly reduce the valleys past the first big one, even without absorbent material, just through length and folds...

When you mention NOT having the labyrinth behind the woofer, I would really question how that's loading the woofer. You're not trying to get output below the Fs of the driver, is that correct? In the case of the Imperial design, output below Fs is entirely possible.

I would add one more observation. The Tannoy GRF-R cabinets, I think you'll agree, are very much the same design as your "scoops", including the labyrinth (aka chamber and throat) behind the driver, and I own a pair. I also have the significantly larger Jensen, with 3 times the horn length and horn mouth size, so I have a basis of comparison between the two.

The big Jensen design, using the same drivers (Tannoy Golds) is in a completely different league. Granted, it comes at a tremendous price of size and weight, but I repeat, it is in a completely different league of presenting live-like dynamics and authority in the room.

The driver demands are similar, yes, (low EDIT: OOPS I mean HIGH EBP) but the loading and cutoff frequency etc. are different enough that I'm not yet convinced that all drivers that work in one design will work in the other.
 
Last edited:
Did you know that there is a company selling those style of scoops on ebay?.

A number of companies have made variations of the "scoop" or "rear loaded horn" cabinets over the years.

Maybe this is a minor terminology nit-pick, but why call it a transmission line rather than simply a backloaded horn? I thought that the expanding flare would make that a BLH, where the TL is intentionally a narrower-bandwidth device, more like a port than a flared mouth....

A transmission line has a more or less constant cross section. A rear loaded horn expands as it gets farther from the woofer. The "scoop" is both a transmission line AND a horn. It exhibites the same up and down response as a transmission line.

I understand about the peaks and dips (notice I posted a similar shot above), but how do those vary with distance from the speaker?.

They don't vary with distance, they vary with frequency. At some frequencies the wave at the mouth is in phase with the woofer, causing a reinforcement, or peak in response. At other frequencies, the wave at the mouth is 180 degrees out of phase with the woofer, causing cancellation, or a dip in response. Damping the line kills those in phase/out of phase peaks and dips at higher frequencies, but the longer waves pass on through, reinforcing that bottom peak. Which is what we want to happen.

One interesting distinction between the scoops and the giant BLH I'm monkeying with is that the BIG cancellation dip is moved down significantly in the longer horn (= longer wavelength that arrives out of phase), with a lot of output (over 100db) at 40 hz. The longer horn also seems to greatly reduce the valleys past the first big one, even without absorbent material, just through length and folds....

When you mention NOT having the labyrinth behind the woofer, I would really question how that's loading the woofer. You're not trying to get output below the Fs of the driver, is that correct? In the case of the Imperial design, output below Fs is entirely possible..

No, I was just pointing out that there are some simpler scoop designs, easier to build, that are quite similar to what you are building.

I would add one more observation. The Tannoy GRF-R cabinets, I think you'll agree, are very much the same design as your "scoops", including the labyrinth (aka chamber and throat) behind the driver, and I own a pair. I also have the significantly larger Jensen, with 3 times the horn length and horn mouth size, so I have a basis of comparison between the two..

Yes, quite similar, which is why I was making the comparison. And Jay mated that scoop bass cabinet with an Altec horn and driver. A neat setup.

The driver demands are similar, yes, (low EBP) but the loading and cutoff frequency etc. are different enough that I'm not yet convinced that all drivers that work in one design will work in the other.

Yes, but you want, for horns, woofer EBP of about 100 or higher, and Fs below the horn cut off.
 
They don't vary with distance, they vary with frequency.

Sure, I understand that. That’s why the big first out-of-phase valley is lower with the longer horn, as shown in the HornResp screen shot I posted above. But is that cancellation completely independent of distance? I thought so at first, but now I’m not so sure. The wavefront coming from that big mouth is not going to be identical with the one coming from a 15” driver, and the cancellation will vary with the incidence of overlap, which has got be different at 12’ than 6’…

As some have been quick to point out to me, if I want to use these without any cancellation, the solution is simple: bolt the woofer firing INTO the compression chamber & throat, enclose it in a sealed or vented back chamber, and let it be a great woofer, period. (Lots on this on the Decware site regarding the Decware Imperial.)

No, I was just pointing out that there are some simpler scoop designs, easier to build, that are quite similar to what you are building. .

Those designs are similar in scale to the Tannoy GRF-Rs, but not to the 30-cu-ft Jensen Imperials – yet there is definitely a family resemblance! Note that I'm not building cabinets -- I have them already.

Yes, quite similar, which is why I was making the comparison. And Jay mated that scoop bass cabinet with an Altec horn and driver. A neat setup.
Yes, with the horn on top that looks much like my current Tannoy GRF-R setup. Funny that we both went with Altec. Very nice, I like it a lot, very enjoyable to listen to, and yes the bass is there to ~40... but it's really not much compared to the Jensen Imperials! I like it enough to consider a better more appropriate horn for it.

Yes, but you want, for horns, woofer EBP of about 100 or higher, and Fs below the horn cut off.

Yes, well, that may be the scoop rule. For the big BLH, conventional wisdom in some quarters is for a still larger EBP. (I made a typo above and put LOW instead of high, trying to get it done without being late somewhere...) If you look at the specs I posted about the drivers I have tried out, they are in the neighborhood of Qe = .22, Fs = 30, so I’m edging closer to 150.

As you say, it isn’t a big deal to find a woofer that works. Just looking for very low Qts and low Fs, I found SEVERAL that work well without much looking! (I'm not building cabs, I'm just trying to figure out how best to use them.) But I decided to get serious about this, and really try some alternatives. The on-the-cheap route, using some 80s Tannoy dual concentrics that I got for under $300/pair years ago, is really quite astounding in quality, with so much of that point-source coherence and imaging combined with a presence and dynamics that is hard to convey. But just how good can the sound be with these crazy-big cabinets?
 
Here's what they looked like when I got them:

28221343_85Ask-M.jpg


A Stephens Tru-Sonic co-axial, a Calrad midrange mounted on the horn flare mouth (???), and a University tweeter in between. These giants lived in my garage for a good while till I engineered a way to get them down my stairs into the basement.

I still haven't painted them, but by next summer I'll be ready to do it.

The removeable baffle board made things easy. The thing is built out of 3/4" plywood, but I made the new baffle boards out of 3/4" MDF. Samples of drivers mounted on some of the several pairs of boards I cut:

215866056_9dvqU-M.jpg
215866111_2WnU3-M-1.jpg
230829030_mmdj9-M.jpg
 
Yes, but you want, for horns, woofer EBP of about 100 or higher, and Fs below the horn cut off.

That second part probably explains something. A friend loaned me a beautiful pair of reconed and maybe re-gaussed JBL E-130's to try (photo above) Specs are:

Qe = 0.21
Fs = 40
Sensitivity = 105

I *wanted* these to work. With that sensitivity, that much less padding of a horn tweeter, and they look GREAT. 40/.21 = EBP 190 (!)

But they just didn't load the horn, at all, and the bass just wasn't there. Maybe the Fs at 40 was too high... but I think the horn DID work with at least one other driver (EV) that had an Fs at least that high... so it might not be a single-parameter answer.
 
Back
Top Bottom