Front loaded and back loaded horn

The reasoning is that adding a front horn to the driver will increase the mids and highs, but the lows will be disproportionate to the intensity of the whole. Re the response plot. By adding a rear horn you can turn out of phase sound into reinforcing in phase sound by passing it through a quarter wave length passage where the sound appears in phase. Horns can use a fut off frequency pass filter In their design to eliminate other frequencies that will travel the same back loaded tunnel to emerge as out of phase distortions to the front wave. When all sound is added together from the front and rear horns you should get a new frequency plot that has more pressure for the lower frequencies which were lower when just the plot of the front horn were used, yet the rest of the frequencies hace the same pressure as before.
 
Hornresp, TL.app and Akabak can all model a front and rear loaded horn load. In Hornresp it's called a compound horn.

Usually the front horn is much smaller so the rear horn is responsible for the lower frequencies and the front horn handles the higher frequencies. If you make the front and rear horn identical and the horn mouths are close together they will just cancel each other out to a large degree (theoretically perfect cancellation but not perfect in real life). So they have to be different to avoid cancelling each other.

There are many examples of compound horns, both diy and commercial. It was a fairly common thing a few decades ago.
 
3fa2cb5b2bdb4debe0926b15743d84f6.jpg


It's not clear from this image whether the rear enclosure is sealed, ported or horn loaded but it doesn't matter, this is an example of a compound design and there's lots of examples of this kind of thing. As long as the front and rear are tuned differently they won't cancel each other (much).
 
Hi Jag,

Usually for answers found, I have even more need.

In a Dinsdale article on horns (and he contributes Klipsch with coming to the same equations), he implies and I use the term advisedly because after it is stated his terminology changes and then I wonder if it meant the direct opposite. But in the two horn description, the front horn raises the mid and high with somehow causing the lows to be filtered out, while the back horn raise the lows and filters out the rest. He describes an equation that sets the volume of the pressure chamber behind the driver, the area of the throat entrance to the rear horn and the cut off frequency for the horn. Now he states that the cut off prevents frequencies above its level to enter the rear horn, and I parse the sentence to make sure that is what he means, and I find no alternate meaning. However I have later found gurus who claim the cut off frequency for the horn limits the low frequencies from enter the horn, exactly the opposite, yet quote the same equations.

However their may be other meaning when one uses a front and back horn. Does the front horn create extra pressure with the opening to the horn being the area of the driver, or do they make this front throat by enclosing a portion of the driver? After all, a horn needs pressure to work, or is the flare of the horn enough. Does the front horn then effect the pressure chamber for the back horn? I cannot find anyone who writes clearly and with any analogy other than electric circuits, not volumes, pressure throats inhibiting frequencies to escape the pressure chamber.

But my gut feeling is that if this equilibrium equation ( V=speed of sound x Throat area / 2 pi x Cut off frequency) for the rear pressure sets the volume larger for lower cut off frequencies, perhaps the larger volume prevents the shorter wave length to escape this pressure (aka air or coupling, depending on the romance of modern sages) and if you get a smaller volume for setting a higher cut off frequency, the shorter wave lengths can escape the chamber.

"Drinking deep may clarify the brain" or intoxicate you!

Dwight
 
David,

Would you need to have a front horn if the full range driver has a fairly even response over the 100Hz to 18kHz range?

Where the response falls off from 93dB at 150Hz to 87dB at 50Hz, would not just a rear horn alone raise up the lows to achieve a flatter response across the usable spectrum?

Dwight
 
You can read the theory endlessly and still have questions because it's hard to grasp by just reading text.

You can do a quick sim and 5 minutes and see exactly how it works with pictures (graphs).

Basically you can just consider the two separate horns. The front horn is small and it provides acoustic gain to the higher frequencies. The back horn is large and provides gain to the lower frequencies. If you balance it out right you can provide a smooth gain bandwidth of up to 6 octaves. Think of a 2 way speaker with a horn tweeter and a horn loaded woofer. It's essentially the same thing except you are using two horns on one driver to cover the different frequency ranges.

The theory articles are usually not wrong but it can be very difficult to understand what they mean without seeing a practical example and some graphs. The theory articles are going to talk about filter chambers and stuff, this is all good theory but you can simulate the effects yourself in 5 minutes and see what actually happens.
 
David,

Would you need to have a front horn if the full range driver has a fairly even response over the 100Hz to 18kHz range?

Where the response falls off from 93dB at 150Hz to 87dB at 50Hz, would not just a rear horn alone raise up the lows to achieve a flatter response across the usable spectrum?

Dwight

If you have a full range driver and just use a back horn to load the lower frequencies that's a back loaded horn. No you don't need a front horn.

But in that scenario you can only load the low frequencies so much, you can only provide enough acoustic gain to lift the level of the lower bass up to the sensitivity of the driver's midband sensitivity.

If you use both a front and back horn you can raise the acoustic gain of the low AND higher frequencies up to a level HIGHER than the midband sensitivity of the driver, and you can do that for a bandwidth of up to about 6 octaves (3 octaves per horn).
 
For ease of use for this type of compound horn TL.app and Hornresp are probably tied. They are both pretty easy to use.

Transmission Line | Leonard Audio
Hornresp

Both have instructions that should be pretty easy to follow but if you need help I know how to use both.

There'a also Akabak but it has a much steeper learning curve.

If you are talking about quarter-wave.com, MJK's site, I don't think he has a worksheet that can simulate a compound horn.
 
Example from the mid 1930's, developed by Olson and Massa.

I have been reading, thinking, designing, reading more, trying to get a grip on a rear loaded horn for a full range driver.

For me, the more I digest, the more I need to find some facts to hang my hat upon, so to speak, but it is difficult to actually get answers off of the internet, answers seem so vague, sentences defy parsing for clarity!

I am reasonably educated. A few years in engineering, a degree in environmental studies, a degree in architecture, ran a construction company, and do a bit of industrial design. A hands on fellow who can build most things. But the information on horn speaker cabinets is less on the dimensions of the space, but on the electrical circuitry used for analysis, not how I design. Information online does not show design but abstract concepts, and I find it hard to gain clarification on accepted horn cabinet design principles.

Can you clarify these items:

1. as the source of the frequency is the driver's cone movement, is the beginning point of measurement for the length of a quarter wave length horn, the driver cone, or the throat opening at the rear air coupling box (air compression volume behind the driver)?

I notice that the length of the rear horns for the early cabinets, Lowther, Voight, et all, are not a quarter wave length of the drivers bass frequency at peak resistance, but much less. Which brings me to ask just what bass frequencies are these highly exclaimed speaker cabinets producing, the upper cut-off frequency, and how does a shortened horn then produce the maximum (lowest) bass frequency that the driver is capable of?

So, as I examine rear loaded horns shorter than the 1/4 wavelength supposedly producing bass extension, is the length beginning at the back side of the cone, and not at the throat opening, to make the horn longer than shown?

With this condundrum of the rear loaded horn increasing the bass response to create a flat, uniform sound pressure across the entire driver's frequency range, what size of horn, throat size, do I need to make to raise the maximum design bass to flatten out the range of response for say a 10P Alpair driver, a 5 dB difference? Some drivers used with horn cabinets need 10-20 dB of bass lift to flatten their frequency curves, and yet the authors claim such wonderful results, is that possible?

2. most design drawings of horn are rectangular expanding tunnels for a horn, logical because wood is being used and curves are too difficult to construct. But sound is spherical, expanding from a point source, the driver, and the horn is best being exponential in expansion over its length, so all rectangular designs are rough approximations of linear, not exponential, expansion, or even worse, uniform cross sectional lengths for short distances then abrupt changes in direction and then an expand again for a distance. This jerky motions seems then to need adjustments to remove distortions with stuffings. Such art of correction makes the science of sound redundant, just hit and miss on avoiding distortion.

I read that bass sound magnification through horns need not have smooth gently expansions, but that does not make sense unless it is the rectangular wood sheets of construction that creates this edict. For example the so called stepping at the mouth of the Fostex horn cabinets, why not just use a smooth plate that smooths out the bumps? Is it all just art?

So with 3D printing, creating smooth exponential expansion, in a circular cross section, would this not be the solution for better distortion free amplification? After all, the science and testing of theories originated with circular cross section horns, and our flat constructions are just trying to be economic with time and space.

3. Use of a Tractrix Curve for rear loaded horns, compared to Exponential Curve, and Dinsdale's observations:

Dinsdale states that the Tractrix Curve produces the best horn sound. But was he discussing ALL horns, or just mid-bass horns? Dinsdale's equations for the coupling chamber, throat area, exponential expansion of horn, length and area of mouth, when put together are impossible to achieve for a 40 Hz full range driver, and, his cabinets did not even come close to producing such a long rear loaded horn, yet, many exclaim at the sound produced. So I assume that his use of a tractrix horn was merely adding it, with a bit of curve smoothing, to the end of the exponential horn. When you plot out a tractrix horn using a .75 of driver cone area for the throat size, a tractrix horn mouth is mammoth compared to the exponential mouth at the same length. So is this adding a tractrix mouth flare just a better way of reducing turbulance at the horn mouth?

Would using a much smaller throat to keep the tractrix mouth to a reasonable size for a quarter wave length 40 Hz horn, give a significant bass boost, or a negligable one when trying to raise the Fo of the 10P Alpair?

4. The design of my cabinet has a vertical firing driver with its forward voice traveling through a bent horn with a 360 degree horizontal output. From a smaller test driver, 6P Alpair, and wooden cabinet firing the rear loaded horn downwards onto the floor, with a fibreglass forward "radial" horn , this changing of driver aspect from vertical to horizontal is not just a wave guide, a deflector as some have assumed by my description, the sound is much larger than the straight on driver normally produces -- the sound intensity is the same anywhere in the room as compared to a test with a quality larger Totem speaker sitting on top of my first radial horn model which was not even louder in its forward broadcast. So there is an amplifying horn producing a stronger sound.

This brings up Dinsdale"s other dictum stating that the length of the rear loaded horn must be adjusted, lengthened? when the cabinet has a short horn at the front of the driver. Why does the rear horn need to be adjusted?

Does this have to do with a time delay for the front horn being delayed by its horn length with the rear horn emitting first at the face of the cabinet, putting the front and rear sound production out of phase? Timing for front and rear sound all begins at the driver cone, with the distances front and rear being a multiple of 1/4 wavelengths, being the measure of time differences with correct phase. The whole point of the rear loaded horn is to transfer the negative, cancelling, sound phase from the rear of the driver through a 1/4 wavelength putting its sound into a positive phase.

But if the rear horn is lengthened by the distance of the front horn, say 12 inches, then is the rear horn still a 1/4 wavelength horn suitable for rear pressure phase correction? Or does increasing the rear horn make it miss both the cut-off and Fo boost frequency? How accurate does a rear horn then have to be for it to function as desired?

I thank you for reading this, and I hope you will have the time to clarify the concepts of rear loaded horn cabinets for me.

Sincerely,

Dwight Chizen
Architect retired.
 
Really two issues: (1) how to match a heavy piece of cardboard and copper to thin air and (2) how to deal with rear wave.

Various benefits to horns. But for woofers, the efficiency gain is no great advantage these days and not much else that's special about the sound, bass or treble, that can't be achieved by other means.

Yes, you can design a horn to address the rear wave. But for obvious reasons, horns (like the wonderful Klipschorn) often just use a sealed box. That has various advantage and since the cone is heavily loaded by the weight of the horn-effect, it can be quite small (like the Klipschorn box).

BTW, easy to make things shaped liked horns but they may not do real impedance matching and really are just labyrinths (like the old posts in this thread). I sure love my 17-foot labyrinth. A horn will fail dramatically as it goes below it's design point (sadly, can't usually be much below 35 Hz) but not a labyrinth.

B.
 
Last edited:
Dear Ben,

Thank you for your reply. You are correct that it is two questions! Perhaps it is four answers that I was looking for to give me confidence in my build.

I had hoped for specific answers to each question, no? I am stumped as to knowing which is correct, my own evaluation, or an improved version of reality. In the end, if I don't get a new way of thinking about how RLHs work, I will just go with my own intelligence. Maybe I was wrong, but I had thought that there might be someone out there that would have an answer and not just wanting to see their name in print. I may have been to optimistic.

Still it it good that you had the time to reply, although I cannot do anything with your observation.

Thanks,

Dwight Chizen,
Toronto
 
Member
Joined 2021
Paid Member
Very interesting thread, in response to your comment about folded horns and curves being difficult to construct, check this out: Folded Horn Acoustic Guitar. Patent # 10,777,172, can find on Google Patents of course. I have two five foot long folded horns inside an acoustic guitar. The horn is exponential and has a 1:8 ratio from throat to horn, with a speaker mounted to each compression chamber. This is long enough to capture the lowest frequency on an acoustic guitar at 1/4 wavelength. The true acoustic sound is a captured by a mic, not a pickup. These horns have zero sharp corners, of course CAD design and CNC equipment needed to do this. Not very difficult if you have the equipment, or maybe can be sourced out. What do you think, and thanks in advance for comments. I like positive and not so positive as well, how else do we learn?
https://patents.google.com/patent/US10777172B1/en?q=folded+horn+acoustic+guitar&oq=folded+horn+acoustic+guitar"]https://patents.google.com/patent/US10777172B1/en?q=folded+horn+acoustic+guitar&oq=folded+horn+acoustic+guitar
 
Can you clarify these items:

Dunno, let's find out..........

1. as the source of the frequency is the driver's cone movement, is the beginning point of measurement for the length of a quarter wave length horn, the driver cone, or the throat opening at the rear air coupling box (air compression volume behind the driver)?

Throat opening of the low pass filter [back box].

I notice that the length of the rear horns for the early cabinets, Lowther, Voight, et all, are not a quarter wave length of the drivers bass frequency at peak resistance, but much less. Which brings me to ask just what bass frequencies are these highly exclaimed speaker cabinets producing, the upper cut-off frequency, and how does a shortened horn then produce the maximum (lowest) bass frequency that the driver is capable of?

One is interested in the horn's acoustic path length that is somewhat longer than the horn length plus need to add in any floor, wall, corner boundaries.

Historically they were tuned [Fc] to actual Fs to get highest impedance matching to the amp for most power efficiency, but actual horn loading was somewhat higher up in the [mid] bass/lower mids since there was basically no signal down low to load.


With this condundrum of the rear loaded horn increasing the bass response to create a flat, uniform sound pressure across the entire driver's frequency range, what size of horn, throat size, do I need to make to raise the maximum design bass to flatten out the range of response for say a 10P Alpair driver, a 5 dB difference? Some drivers used with horn cabinets need 10-20 dB of bass lift to flatten their frequency curves, and yet the authors claim such wonderful results, is that possible?

No clue and TTBOMK never seen any info for designing horns to some specific dB gain over 'x' bandwidth [BW].

Absolutely, been there, done that. ;)

2. most design drawings of horn are rectangular expanding tunnels for a horn, logical because wood is being used and curves are too difficult to construct. But sound is spherical, expanding from a point source, the driver, and the horn is best being exponential in expansion over its length, so all rectangular designs are rough approximations of linear, not exponential, expansion, or even worse, uniform cross sectional lengths for short distances then abrupt changes in direction and then an expand again for a distance. This jerky motions seems then to need adjustments to remove distortions with stuffings. Such art of correction makes the science of sound redundant, just hit and miss on avoiding distortion.

IME, most folded horns are parabolic, i.e. parallel sides with slanted board dividers.

I read that bass soundt that does not make sense unless it is the rectangular wood sheets of construction that creates this edict. For example the so called stepping at the mouth of the Fostex horn cabinets, why not just use a smooth plate that smooths out the bumps? Is it all just art?

So with 3D printing, creating smooth exponential expansion, in a circular cross section, would this not be the solution for better distortion free amplification? After all, the science and testing of theories originated with circular cross section horns, and our flat constructions are just trying to be economic with time and space.

LF is of course huge and our hearing acuity sucks down low, so such construction is for blocking higher frequencies of whatever type from comb filtering with the driver's output.

3. Use of a Tractrix Curve for rear loaded horns, compared to Exponential Curve, and Dinsdale's observations:

Dinsdale states that the Tractrix Curve produces the best horn sound. But was he discussing ALL horns, or just mid-bass horns?

As stated, he's sadly mistaken. Conical wave guides [WG] are and all horns can be improved further with proper driver/horn throat terminus [mouth] transition. For low tuned BLHs, pure hyperbolic is best, though usually a hypex is done as a compromise.

I presume ALL horns since he seemed Tractrix centric.


Dinsdale's equations for the coupling chamber, throat area, exponential expansion of horn, length and area of mouth, when put together are impossible to achieve for a 40 Hz full range driver, and, his cabinets did not even come close to producing such a long rear loaded horn, yet, many exclaim at the sound produced. So I assume that his use of a tractrix horn was merely adding it, with a bit of curve smoothing, to the end of the exponential horn. When you plot out a tractrix horn using a .75 of driver cone area for the throat size, a tractrix horn mouth is mammoth compared to the exponential mouth at the same length. So is this adding a tractrix mouth flare just a better way of reducing turbulance at the horn mouth?

No clue, the pioneers [W.E. ,etc.] + DIY experience taught me how to design horns without T/S specs or using higher math, so by the time I was exposed to it had already concluded that Tractrix was of no interest to me.

That said, a rapid or flared expansion at the mouth does reduce mouth induced distortions even if a simple flat one like used in most modern CD horns or IME just a flat baffle with some damping if made large enough.


Would using a much smaller throat to keep the tractrix mouth to a reasonable size for a quarter wave length 40 Hz horn, give a significant bass boost, or a negligable one when trying to raise the Fo of the 10P Alpair?

The throat area [St] sets the horn's HF response, which is a function of the driver's upper mass corner [fhm]*, so no. You want to raise [Fc] relative to [fhm], then change its flare frequency.

*http://www.xlrtechs.com/dbkeele.com...Preprint) - LF Horn Design Using TS Paras.pdf


4. The design of my cabinet has a vertical firing driver with its forward voice traveling through a bent horn with a 360 degree horizontal output. From a smaller test driver, 6P Alpair, and wooden cabinet firing the rear loaded horn downwards onto the floor, with a fibreglass forward "radial" horn , this changing of driver aspect from vertical to horizontal is not just a wave guide, a deflector as some have assumed by my description, the sound is much larger than the straight on driver normally produces -- the sound intensity is the same anywhere in the room as compared to a test with a quality larger Totem speaker sitting on top of my first radial horn model which was not even louder in its forward broadcast. So there is an amplifying horn producing a stronger sound.

This brings up Dinsdale"s other dictum stating that the length of the rear loaded horn must be adjusted, lengthened? when the cabinet has a short horn at the front of the driver. Why does the rear horn need to be adjusted?

The driver's upper mass corner [fhm] sets the point where the BLH 'hands off' to the driver/FLH, so yes, as the [fhm] rises higher, the BLH gets longer for a given cut-off [Fc] and vice versa.

Does this have to do with a time delay for the front horn being delayed by its horn length with the rear horn emitting first at the face of the cabinet, putting the front and rear sound production out of phase? Timing for front and rear sound all begins at the driver cone, with the distances front and rear being a multiple of 1/4 wavelengths, being the measure of time differences with correct phase. The whole point of the rear loaded horn is to transfer the negative, cancelling, sound phase from the rear of the driver through a 1/4 wavelength putting its sound into a positive phase.

But if the rear horn is lengthened by the distance of the front horn, say 12 inches, then is the rear horn still a 1/4 wavelength horn suitable for rear pressure phase correction? Or does increasing the rear horn make it miss both the cut-off and Fo boost frequency? How accurate does a rear horn then have to be for it to function as desired?

Yes in that it will need to be designed in, i.e. design for 'x' [fhm] and drop in a driver with a sufficiently different one and now it's audibly misaligned, though with smaller BLHs and/or low tuned, the room dominates to the point where one usually winds up with a very chopped up response that thankfully our poor hearing acuity + very powerful processor yields a pleasing musical performance.

The main thing is getting the octaves around the BLH >>>> FLH transition [acoustic XO] as seamless as practical same as with any electrical one.

Last, but not least is using a driver with the 'right' specs for the app, though with today's inexpensive DSP and higher power drivers, not the big deal it once was unless using vintage low power tube amps that require very efficient speakers.


I thank you for reading this, and I hope you will have the time to clarify the concepts of rear loaded horn cabinets for me.

You're welcome!