Horn Range Question

[Wrinkles]

Is the ability of a horn to reach low frequencies based on the horn itself (length-horn opening-throat size-etc) or is it the driver behind it or both?

Does horn length determine how high the horn will go? Or is it the driver what gives the upper range limitation?

I understand the 511B horn which is rated down to 500 hz is often being substituted for the K400/K401 horn in the LaScala. If I put the the K55M driver will the frequency go down to 400 hertz? I do have the 511B now and am wondering about the 100 hertz difference that by specifications is missing from 400 to 500 hz.

I know these are newbie questions and I am trying to learn.

Is there good reading material on the web that discusses horn design? Or book recommendations?

Thanks,

Wrinkles

[DrWho]

Probably the easiest way I've heard it described is that you have two systems at play: the driver, and the horn.

Every horn has a cutoff frequency (Fc) below which the horn is no longer

loading the driver, which is mostly influenced by the length of the

horn. The size of the mouth mostly determines how flat the response is

over the passband, but can also impact the low frequency limit as well.

This trend is true for all flare rates, but each different kind of

flare rate will differ from the others. For example, exponential needs

to be longer than a tractrix horn.

And then every driver has its own frequency response on something

called a "plane-wave tube", which is essentially a tube that emulates a

perfect horn so that the behavior of a compression driver can be

studied.

So take the plave-wave behavior of the driver and convolve it with the

frequency response of the horn and you'll end up with the system

response.

All that to say, they both play a roll. If the horn's cutoff is lower

than the driver, then the system will be limited by the driver's

response. Basically you'll roll off at the rate the driver rolls off

until you hit the cutoff frequency of the horn where it'll roll off at

a faster combined rate. The same is true if the horn rolls off first.

Basically, the device with less bandwidth is going to determine the

usable bandwidth for the total system.

Often, especially in the pro world, a horn's Fc will be

lower than its useable frequency range because horn designers are

usually trying to optimize the polar response. The horn needs to be

very large to maintain pattern control at Fc, so they generally just

push the Fc a bit lower so that they can control the dispersion better.

For example, the K402 (a newer horn from the pro cinema line) is

actually a 250Hz horn (or is it 350? I forget). It is only used from

400Hz and above because that's the region over which it keeps the desired polar response.

I dunno if the 511 is actually a 500Hz Fc horn or not, but I know Shawn

(sfogg) was once running them with his lascalas with no problem. He has

since upgraded to the Klipsch K510 which is a tractrix constant

directivity style horn versus the exponential 511 with collapsed

polars. I know the K510 is a 500Hz horn and he has had no problems

running the lascala bass bin up to 500Hz. Crossing over higher would

require some changes to the network, but it's probably the route I

would go.

You mentioned some websites...here's a good place to start I think:

http://www.melhuish.org/audio/horn.html

[Guest " "]

Wrinkles

One picture is worth a million DrWho words....

The 511, K400, have a smiular roll off patern as indicated in this picture. The roll off is not cliff like.....is a gradual roll off

The 511 is a better performing horn than the k-400 from a coverage perspective....matched with a ST-350 or some other bi-radial tweeter and your sweet spot becomes wider in your listening area.

400hz would be the first vertical dotted like to the right of the k-400 label on this graph.

There you go...approx 75 words or less...picture included...answer to your question.

[DrWho]

Looking at that plot, the 511 looks to be a 500Hz Fc horn. The K400 is exhibiting that sinc function behavior near the Fc because the mouth is too small. In gets brought up from time to time by I think Maron and djk. Basically, it turns out the K400 can be shortened to yield a very similar frequency response.

[PrestonTom]

Following up on the "driver" part of the equation.

If the driver is forced to reproduce very low frequencies, you will get distortion and possibly damage the driver. This needs to be accounted for when you set up the crossover (both corner frequency and slope) if you are trying to get the mid-range to extend down in frequency.

Good Luck,

-Tom

[Wrinkles]

Thankyou to DrWho, SpeakerFritz and PrestonTom for your responses.

I understand that from the above description of horn design that an Fc of 500 hertz, like the Altec 511B when mated to any driver, will produce very little usable below Fc.

The chart shows that the 511 horn goes below Fc and while not a vertical drop, the frequency falls off gradually. It looks like around 310 hertz the response is pretty much garbage. Similarly, garbage above 7000 Hz. Garbage is not the right term, but to me means that the produced frequencies are pretty much unusable due to the lower db. With the chart showing all three horns (K400, 511B and Trachorn) reproducing down to approximately 310 hertz using the same K-55-V driver , then with an unmodified LaScala crossover (my AL-3 network) designed for the 400 hertz horn, it does not seem that there would be any problem utilizing the 511B and the K-55-M Klipsch driver and the crossover at the 400 hertz crossover point together. It would not seem the K-55-M driver would be taxed to anywhere near the damaging point.

The "loading" of the driver is somewhat unclear to me. Is it that when a driver, rated to a lower frequency than the horn, is attempting to develop frequencies below Fc that the driver requires a feedback wave from the horn to develop the "load" for the driver to prevent diaphragm damage from overcurrent or perhaps the waffling or deformation of the diaphragm material?

Is my thinking / explanation above basically newbie BS based on my misunderstandings or is it somewhere close to what is really happening. Would I be heading towards driver damage in my LaScalas if I proceed with the K400/K401 to 511B swapout without crossover network changes and using my K55M drivers? I am presently driving the 511B with Altec drivers (808-8A and 802D) with what appears to be no problems.

Thanks,

Wrinkles

[RFP]

Wrinkles: Check your "In Box"

Rpb

[PrestonTom]

Wrinkles,

Typically, most folks swap in both the 511b horn and the 902 driver. See the threads by SFOGG for a nice description (although he has since gone a step or two further).

The question is: does the 511/902 sound that much better than the 511/k55 combination? I think some have done that, but I don't know if they had a chance to compare to the 511/902 combination.

BTW, if you do go to the 511/902 combination, it is possible to then make the cabinet a 2-way design (some see an advantage in a 2-way design). SFOGG had done this and talked about the results. In a nutshell there may be an issue about 1) boosting the high end (using an equalizer), 2) the dispersion at the very high frequencies and 3) using the proper diaphragm on the 902.

RE: the 802 driver, does that go down to 400 Hz (without distortion, and "can it play loud"). BTW another source of info is to call Great Plains Audio.

Good Luck,

-Tom

[Deang]

I've talked to several who actually prefer the sound of the K-55 to the 902.

[Arkytype]

Wrinkles, et al----

I generated these curves for Al K. during the early ALK Trachorn prototype days. Note that the sensitivity of the three horns is normalized at 1k Hz. The original plot was done in three colors so the individual horn responses could be readily identified. My recollection is that the acoustic output for each horn/K55V combination was very close for a nominal 1 watt input to the K55V driver. All my curves and data are in storage so I can't post the three color version.

Of course, a loudspeaker's "frequency response" curve alone won't reveal all the subtleties of how it sounds.

Before converting my Klipschorns to the Trachorns, I had 511Bs for a few months. The improvement (to my 59 year old ears) was a widening of the soundstage and less of the "horn" sound I had lived with since the horns were purchased new in 1979. The 511B conversion also included adding Beyma CP-25s and changing the crossovers to ALK's extreme slope networks.

Changing to the Trachorns in place of the 511Bs brought the imaging into sharper focus and rendered my center channel Belle Klipsch superfluous.

Daddy Dee and Fini spent several hours listening to my system a couple of years back. At one point, while listening to a k.d. Laing CD, one of us wondered if the Belle was on. It wasn't. The imaging was that good.

Contrary to my Klipschorn modification process, my suggestion for upgrading a heritage model is to make incremental changes. Try modifying one loudspeaker at a time and (if possible) place them side by side. Feed monophonic music to one and then the other using your balance control. If you hear a difference, then determine if it is a "good" difference or if the mod needs tweaking or removing.

For the beginning DIYer, Bob Crites' networks would be a good start. There's been plenty of articles written about the "sound" of capacitors. If you use a good dielectric, the benefits will be immediately audible both in preamps, amplifiers and certainly in dividing networks. Building or purchasing one of ALK's networks is another worthwhile albeit expensive upgrade.

As for tweeter upgrades, the Beymas are the only replacement tweeters I've listened to so I can't comment on other choices available. The best adjective I can use to describe their sound is smooth. Bob Crite's version of the EV might be a better choice if you don't want to change the motorboard.

My goal in making changes to my Klipschorns has been to be able to reproduce music in a lifelike, believeable manner. At this point, the limiting factor is (as it should be) the source material.

Lee

[IB Slammin]

I've talked to several who actually prefer the sound of the K-55 to the 902.

Deano,

They must be the die-hard Klipsch purest. The 902-16a is all about the upper mids. As seen in the measurements, and to my ear, the 511/902 combo perform well crossed at 400hz, and can take the beating that I give them, even crossed 1st order. The 902 excells & separates from the K55 pack in the upper end of mid range. Do they have a Heritage sound? Close, but the drift away from that sound begins. Does a K402/Jube have a Heritage sound? Not to me. Does it sound better? Not yes, but hell yes IMHO.

See my point?

tc

[jwc]

I've talked to several who actually prefer the sound of the K-55 to the 902.

I'm not one of those.

As far as heritage sound....very subjective description. Adding the 511 to a heritage sytem offers "improvement" and maintains the "like" sound.

902 driver change out did "move out" of the heritage sound......somewhat......it was very good though.

The two way option with the 902/511 is decent....

Lot of quotes in there.

jc

[DrWho]

With the chart showing all three horns (K400, 511B and Trachorn)

reproducing down to approximately 310 hertz using the same K-55-V

driver , then with an unmodified LaScala crossover (my AL-3 network)

designed for the 400 hertz horn, it does not seem that there would be

any problem utilizing the 511B and the K-55-M Klipsch driver and the

crossover at the 400 hertz crossover point together. It would not seem

the K-55-M driver would be taxed to anywhere near the damaging point.

What

you don't see in these graphs are the impedance and phase responses.

Ultimately, the effective crossover frequency is determined by the

impedance of the driver. If the driver impedance changes, the

electrical crossover

frequency changes. Changing the Fc shifts around the impedance peak

which in turns slides the crossover around and changes the electrical

phase of the system in that region. Additionally, the way the horn

loads the driver also changes the acoutical phase. The combination of

both phase shifts is going to affect the region of overlap between the

bass bin and the squawker. However, the passive crossover forces a

system lacking time-alignment which in turn prevents a perfect

transition. I can't help but wonder if this makes the crossover

compromises less audible.

Overall, these factors are

going to be relatively small since 400Hz and 500Hz is a very small

difference, but for someone striving a higher level of fidelity I think

they are easily audible and very important to address. It would be

fairly trivial to sit down and re-align the crossover, but it would

also requiring some measuring.

I don't think power capability is the issue here.

The "loading" of the driver is somewhat unclear to me. Is it that when

a driver, rated to a lower frequency than the horn, is attempting to

develop frequencies below Fc that the driver requires a feedback wave

from the horn to develop the "load" for the driver to prevent diaphragm

damage from overcurrent or perhaps the waffling or deformation of the

diaphragm material?

Think of the horn as a way of adding mass to the driver, ultimately

making it harder to move the diaphragm. Normally making the diaphragm

harder to move is a bad thing since more output requires more cone

movement. However, the mass being added to the diaphragm is the very

air you're trying to vibrate in the first place. I just think that's so

cool.

The amount of extra mass introduced by the horn is going to be

different for different frequencies (and for different horns too). The

rate at which the area expands will determine how well the air couples

to the driver. If you go too fast, it doesn't take much to visualize

that it starts behaving like a direct radiator. If you go too slow, the

air velocity increases and the potential for reflections off the sides

increase.

Below the Fc, the horn isn't adding any mass to the driver. This means

less air moves for the same cone excursion, which is why you see the

drop in the frequency response. You can also see this in the impedance

response - the driver is easier to move so below the Fc the impedance

is very low. At higher frequencies, the output from the driver tends to

beam so it doesn't see the horn walls. This means it's moving as much

air as if there were no horn present at all and is one of the reasons

you see the rolloff at the higher frequencies. In addition to that,

every driver has a mass and inductance roll off. Basically, the

strength of the motor eventually doesn't have enough power to control

the momentum of the cone (making it harder to switch directions which

is what a high frequency is...switching back and forth really fast).

The motor is also a coil of wire, which effectively is an inductor and

is an additional source of high frequency roll off. For example, a

piezo driver looks like a capacitive load, so the high

frequency behavior is not electrically limited. And since the

diaphragms

are light, piezo tweeters generally exhibit extension into very high

frequencies.

Since air defracts differently at different frequencies, the shape and

flare of the horn's walls will boost different frequencies by different

amounts.

Directivity is also determined by the shape of the effective radiating

area. Since the low frequencies follow the edge of the horn, their

directivity is determined by the size and shape of the mouth of the

horn. Higher frequencies beam off the side walls so their directivity

is determined by the area of the horn where they release. Just

like focusing a flashlight increases its intensity over a smaller area,

you can achieve increased SPL on-axis by making the directivity more

narrow. PWK was very fond of collapsing the vertical polars to achieve

increased on-axis gain at the higher frequencies (to help compensate

for the HF rolloff of the compression driver). So what might be a

nominal 90x40 horn might actually be 90x20 at the higher frequencies.

The polar behavior is

important because the sound we hear in our rooms is a combination of

direct and reflected sounds. You need to make sure that the LF and MF

have similar polars in the crossover region so that there aren't sudden

changes in the tonality of the reflections.

I know I'm being a little abstract and not targeting what seems to be

your specific application, but you expressed an interest in the theory

and I thought I would try to expand on your questions to the best of my

knowledge (which is nowhere near complete). However, specific to your

current application I would encourage someone to determine what

shortcomings in the K400 they would like to improve upon and then find a solution that does just that. In my humble

opinion, I find the 511b to be more similar to the K400 than different,

which is to say that I feel they both suffer from the same types of problems.

Would it not make sense to persue a solution that doesn't suffer from

similar problems without introducing new ones? It's easy to say, but I

think it's also a practical and realistic approach to consider (unless of course you already own the altec horns).

[DrWho]

They must be the die-hard Klipsch purest. The 902-16a is all about the upper mids....

The 902 excells & separates from the K55 pack in the upper end of mid range.

I have found that those would also be my impressions when listening at louder SPL's. I dunno if everyone feels the same, but at lower levels I think it's harder to tell.

Has anyone ever tried a 902 on the K400?

[Guest " "]

With the chart showing all three horns (K400, 511B and Trachorn)

reproducing down to approximately 310 hertz using the same K-55-V

driver , then with an unmodified LaScala crossover (my AL-3 network)

designed for the 400 hertz horn, it does not seem that there would be

any problem utilizing the 511B and the K-55-M Klipsch driver and the

crossover at the 400 hertz crossover point together. It would not seem

the K-55-M driver would be taxed to anywhere near the damaging point.

What

you don't see in these graphs are the impedance and phase responses.

Ultimately, the effective crossover frequency is determined by the

impedance of the driver. If the driver impedance changes, the

electrical crossover

frequency changes. Changing the Fc shifts around the impedance peak

which in turns slides the crossover around and changes the electrical

phase of the system in that region. Additionally, the way the horn

loads the driver also changes the acoutical phase. The combination of

both phase shifts is going to affect the region of overlap between the

bass bin and the squawker. However, the passive crossover forces a

system lacking time-alignment which in turn prevents a perfect

transition. I can't help but wonder if this makes the crossover

compromises less audible.

Overall, these factors are

going to be relatively small since 400Hz and 500Hz is a very small

difference, but for someone striving a higher level of fidelity I think

they are easily audible and very important to address. It would be

fairly trivial to sit down and re-align the crossover, but it would

also requiring some measuring.

I don't think power capability is the issue here.

The "loading" of the driver is somewhat unclear to me. Is it that when

a driver, rated to a lower frequency than the horn, is attempting to

develop frequencies below Fc that the driver requires a feedback wave

from the horn to develop the "load" for the driver to prevent diaphragm

damage from overcurrent or perhaps the waffling or deformation of the

diaphragm material?

Think of the horn as a way of adding mass to the driver, ultimately

making it harder to move the diaphragm. Normally making the diaphragm

harder to move is a bad thing since more output requires more cone

movement. However, the mass being added to the diaphragm is the very

air you're trying to vibrate in the first place. I just think that's so

cool.

The amount of extra mass introduced by the horn is going to be

different for different frequencies (and for different horns too). The

rate at which the area expands will determine how well the air couples

to the driver. If you go too fast, it doesn't take much to visualize

that it starts behaving like a direct radiator. If you go too slow, the

air velocity increases and the potential for reflections off the sides

increase.

Below the Fc, the horn isn't adding any mass to the driver. This means

less air moves for the same cone excursion, which is why you see the

drop in the frequency response. You can also see this in the impedance

response - the driver is easier to move so below the Fc the impedance

is very low. At higher frequencies, the output from the driver tends to

beam so it doesn't see the horn walls. This means it's moving as much

air as if there were no horn present at all and is one of the reasons

you see the rolloff at the higher frequencies. In addition to that,

every driver has a mass and inductance roll off. Basically, the

strength of the motor eventually doesn't have enough power to control

the momentum of the cone (making it harder to switch directions which

is what a high frequency is...switching back and forth really fast).

The motor is also a coil of wire, which effectively is an inductor and

is an additional source of high frequency roll off. For example, a

piezo driver looks like a capacitive load, so the high

frequency behavior is not electrically limited. And since the

diaphragms

are light, piezo tweeters generally exhibit extension into very high

frequencies.

Since air defracts differently at different frequencies, the shape and

flare of the horn's walls will boost different frequencies by different

amounts.

Directivity is also determined by the shape of the effective radiating

area. Since the low frequencies follow the edge of the horn, their

directivity is determined by the size and shape of the mouth of the

horn. Higher frequencies beam off the side walls so their directivity

is determined by the area of the horn where they release. Just

like focusing a flashlight increases its intensity over a smaller area,

you can achieve increased SPL on-axis by making the directivity more

narrow. PWK was very fond of collapsing the vertical polars to achieve

increased on-axis gain at the higher frequencies (to help compensate

for the HF rolloff of the compression driver). So what might be a

nominal 90x40 horn might actually be 90x20 at the higher frequencies.

The polar behavior is

important because the sound we hear in our rooms is a combination of

direct and reflected sounds. You need to make sure that the LF and MF

have similar polars in the crossover region so that there aren't sudden

changes in the tonality of the reflections.

I know I'm being a little abstract and not targeting what seems to be

your specific application, but you expressed an interest in the theory

and I thought I would try to expand on your questions to the best of my

knowledge (which is nowhere near complete). However, specific to your

current application I would encourage someone to determine what

shortcomings in the K400 they would like to improve upon and then find a solution that does just that. In my humble

opinion, I find the 511b to be more similar to the K400 than different,

which is to say that I feel they both suffer from the same types of problems.

Would it not make sense to persue a solution that doesn't suffer from

similar problems without introducing new ones? It's easy to say, but I

think it's also a practical and realistic approach to consider (unless of course you already own the altec horns).

DrWho....need some pictures....I think your handle should be "word smith"

[JohnA]

I've talked to several who actually prefer the sound of the K-55 to the 902.

Deano,

They must be the die-hard Klipsch purest. The 902-16a is all about the upper mids.

As seen in the measurements, and to my ear, the 511/902 combo perform

well crossed at 400hz, and can take the beating that I give them,

even crossed 1st order. The 902 excells &

separates from the K55 pack in the upper end of mid range. Do they

have a Heritage sound? Close, but the drift away from that sound

begins. Does a K402/Jube have a Heritage sound? Not to me. Does it

sound better? Not yes, but hell yes IMHO.

See my point?

tc

My 902/511B/FH1 2-way combo

sounds smooth and makes a good live singer sound very real. I

like them a lot and don't hear a "horn" sound even on piano. I

would think they would sound excellent on a cleaner La Scala bass horn.

I would not be surprised if it sounds better than the Klipsch

treble horns.

[Wrinkles]

Before converting my Klipschorns to the Trachorns, I had 511Bs for a few months. The improvement (to my 59 year old ears) was a widening of the soundstage and less of the "horn" sound I had lived with since the horns were purchased new in 1979. The 511B conversion also included adding Beyma CP-25s and changing the crossovers to ALK's extreme slope networks.

Thankyou Lee,

Did changing the 511B out to the Trachorn remove the rest of the "horn" sound?

Thankyou,

Wrinkles

[Wrinkles]

Dr. Who,

Thankyou. You understood exactly what I was looking for in your explanation. Theory, that is not over my head in super technical terms. Thanks again.

I do have the Altec 511b horns presently and plan on installing the K55M drivers on them once the adapters come in. I'll see how it sounds and go from there.

Wrinkles

[mas]

DrWho....need some pictures....I think your handle should be "word smith"

Fritz, Who may use words in the process of explaining things, but at least he contributes to the discussion.

On the other hand, every one of you posts can be reduced to a picture... the Same picture:

[DrWho]

lol Mark - maybe we can put together a transformer coloring book?

Wrinkles, if you want to get a bit more into the numbers you might try checking out HornResp:

http://www.users.bigpond.com/dmcbean/

It's a free program that predicts the acoustical output for all sorts

of horns (and other things too). You will be able to see for yourself

how each variable influences the total system output. Being a

simplified physical model it does not perfectly correlate to real life,

but it illustrates very well how different design aspects influence the total system.

Again on the technical side, you might find the writings of D. B. Keele rather helpful:

http://www.xlrtechs.com/dbkeele.com/papers.htm

In fact, I think he writes specifically about mouth size too (which is the issue with the K400 I was referring to).