smith horns and using compression horns for a Karlson party speaker project

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am presently planning things out for my first DIY project, a pair of 15" Karlsons(say what you will about the Karlson). I am hoping to cross the woofer in the K-coupler at something like 1 to 1.5 Khz and hopefully use a compression driver and horn lens like a Selenium D210T to carry things from there on.

Let me add a disclaimer: i may be a novice audiophile and DIYer but i believe i know what i am looking for from my speakers-to-be. because i'm a 19-year-old university student, i have to be pragmatic. i really don't forsee myself having a listening room of any sort until i hopefully complete law school and finish paying off my student loans, etc. as a result, it is also rather unlikely that i am ever going to own anything substantially more "reference class" than a 70's/80's era mid-fi reciever(presently a '76 Marantz/Superscope R-1270) and Technics SL-1600 turntable for a number of years to come. The result is i am not looking to achieve the ultimate reference speaker - rather, my focus is creating a big, bad, and incidentally loud party speaker...but a better party speaker than many/most. at present my objective is to outperform a set of '83 vintage B&W DM220s i borrowed from my friend as far as punch and slam in the bass and low midrange is concerned. i guess you could say i want kickdrums to really "kick". also would like to avoid excessively icepicky highs. in addition these will be my everyday listening speakers when i am not using headphones.

in a party speaker, everyone values high SPLs and pounding bass, but that isn't out of the ordinary for this "genre" of speaker. i would however also like wide dispersion so that in the context of a party, everyone in a room will be enjoying the music reasonably equally. on the Karlson forum, i came across some plans/an article about a Distributed Source Horn, aka "Smith Horn". i did some reading on it and clearly it was conceived in the mono era to radiate in a broad area. somewhere it was suggested that using such a horn would result in smearing of stereo image because it has something like a 120 degree arc...a valid concern but since i will be using my speakers in less than ideal conditions(drunken university keg parties, anyone?), i would actually welcome this. JBL apparently made a smith horn in the 70's...2397 was it? part of the appeal of the DSH is that a guy like me can build it fairly easily because of the lack of vertical expansion. aesthetically speaking, it would look a lot better than me building a box to go on top of the Karlson to accomodate a midrange and tweeter. if possible i would like to stick to a 2-way crossover.

So, after all this rambling, i have a few questions:

1) is it inherently a bad idea to use compression drivers in a hi-fi design? even with proper attentuation, is it likely to result in HF response that is painful to the ear?

2) if i choose to use a pre-made horn lens and build it into an extension of the cabinet top(resembling midrange horn on Klipsch La Scala series I) will this also be a bad idea? a worse idea?

3) the Selenium D210Ti is claimed to have response going down to 800 Hz yet the manufacturer recommends a second-order crossover at 2000 Hz. looking at the specs, it seems the impedance is kind of lumpy below that. can this be reasonably notch-filtered out? is a notch filter even applicable here? i would really like to be able to use it at about 1200 Hz if possible...and i thought i would use a more aggressive slope, probably fourth-order.

4) does anyone have experience with distributed source/Smith horns? what are their characteristics? pitfalls? are they as easy to build as they look?

5) if the Karlson project goes well, i may build a much more conventional-looking sealed tower set of speakers for my brother's birthday present. he is not an audiophile although he knows good sound when he hears it. what if i were to use a compression driver and one of those round Dayton waveguide things on Parts Express? my brother has never experienced high-efficiency speakers before. he has one of those mini-stereo things everyone buys at major retail type places...although when he heard my friend's DM220s with my reciever, he decided he'd have to re-evaluate his audio lifestyle...being 31 and a little more "established" than i am, he's probably not looking for a party speaker.

thanks for the help in advance.
 
Yes JBL made a Smith horn with the number 2397. It was made for the large format drivers with 2" exit. Westlake have made several Smith horns. One that is almost identical in dimensions to the JBL 2397, and a smaller one for 1" drivers.

The JBL 2397 is usable from 800Hz.

I have made a 1200Hz Smith horn for 1" driver. http://audioheritage.org/vbulletin/showthread.php?t=5454

I would look for a JBL PRO driver to run it, maybe a 2425/2426 which is often quite reasonable on eBay.

The look beautiful and play beautifully.
 
thanks a lot...very informative thread!

I am intrigued by your shrunken 1200 Hz horn. did you ever get satisfactory measurement accuracy from your computer? how high would you say they begin to roll off at? also, what frequency did you cross them over at, and what slope?

another question, out of curiosity, on the speakers to which you mounted your Smiths, there seems to be a tweeter lens with downward-pointing vanes...what do you call those things? i've seen them in other places before - i particularly remember my uncle used to own a pair of 3-way Technics floorstanders with such a lens on the tweeters. i'm sure he still has them but he lives in Virginia and since him and my dad no longer "get along" it's unlikely i'll see those speakers again.

back to Smiths.

I have a few scans of the original Smith Horn article which appeared in Audio Engineering Magazine in January 1951. I'm sure you've come across it before...the article details an 800 Hz 1" horn...i'd consider designing my own using the information given in the article but it happens that the "example" horn is just what i needed - at least i think. I am worried though, that it will turn into just a mid horn and i will need another solution for HF. i was really hoping to cover the range with one horn per Karlson. Ken Pachkowsky's mini westlake smiths on his speakers seem really cool and a possible solution if i can't get one horn to do it all. is it likely i will need to go with two horns? do you use have to use a supertweeter with yours?

I also don't think i can accomodate anything made by JBL into my budget for this speaker project...hence, i wanted to use a 1" compression driver by Selenium or Eminence which i could score for under $60 US new per unit.

finally, i noticed that on all the smith horns detailed on Lansing Heritage, including yours, the vanes do not extend all the way to the mouth of the horn whereas the original Smith article clearly shows the tips of the vanes at the edge of the horn mouth. what is the philosophy behind this?

thanks.
 

GM

Member
Joined 2003
>1) is it inherently a bad idea to use compression drivers in a hi-fi design? even with proper attentuation, is it likely to result in HF response that is painful to the ear?
====
Greets!

Absolutely not! Depends on the driver, diaphragm material, but mostly how you apply it WRT XO point slope and shelving ckt.
====
>2) if i choose to use a pre-made horn lens and build it into an extension of the cabinet top(resembling midrange horn on Klipsch La Scala series I) will this also be a bad idea? a worse idea?
====
It depends on the XO point/slope, which dictates the horn size, ergo the physical offset that affects the acoustic phase match at this point. If digital delay is used, it's a moot point.
====
>3) the Selenium D210Ti is claimed to have response going down to 800 Hz yet the manufacturer recommends a second-order crossover at 2000 Hz. looking at the specs, it seems the impedance is kind of lumpy below that. can this be reasonably notch-filtered out? is a notch filter even applicable here? i would really like to be able to use it at about 1200 Hz if possible...and i thought i would use a more aggressive slope, probably fourth-order.
====
Hmm, the Selenium site says 1.5 kHz: http://seleniumpro.com/www/Shopping/index.html

Regardless, this is based on a 160 W power handling capability. If you can get by with less then you can use a lower XO point, but you want the lens to load down to at least an octave below the XO point, so at 800 Hz you'll need at least a 400 Hz lens and even then, power handling will have fallen to just a fraction and efficiency will be down around where it needs to be to blend with the K. A broadband notch filter with a bypass cap will be required to flatten its response. Whether it will play loud enough at reasonably low distortion at a frat party remains to be seen since I have no experience with this particular driver.
====
>4) does anyone have experience with distributed source/Smith horns? what are their characteristics? pitfalls? are they as easy to build as they look?
====
Distributed source, yes, Smith, no. They're fine until the frequency reaches ~a WL across. Above this point the comb filtering is unacceptable to me, but some folks really like them without a super tweeter, so as always YMMV.
====
5) if the Karlson project goes well, i may build a much more conventional-looking sealed tower set of speakers for my brother's birthday present. he is not an audiophile although he knows good sound when he hears it. what if i were to use a compression driver and one of those round Dayton waveguide things on Parts Express? my brother has never experienced high-efficiency speakers before. he has one of those mini-stereo things everyone buys at major retail type places...although when he heard my friend's DM220s with my reciever, he decided he'd have to re-evaluate his audio lifestyle...being 31 and a little more "established" than i am, he's probably not looking for a party speaker.
====
If going this route, I recommend just buying one of WP's kits and save yourself some hassle: http://seleniumpro.com/www/Shopping/index.html

GM
 
thanks for the reply!

hmmmm, so it seems there was after all a reason i was having a hard time finding a horn driver that went as low as i wanted it to....the lens would be too big! thanks. i didn't know how much "space" i'd have to leave in the FR with a horn. i guess i am resigned to crossing somewhere around 1500. if i cross an Eminence PSD2002 or similar Selenium at 1600 Hz with a DSH, that would leave just barely over one octave as the horn in question loads down to 750 Hz. I don't want to cross the woofer too high as i hear the Karlson can sound funny as you go higher in the midrange....and i already know that "funny" is an understatement in some people's opinions of the Karlson coupler. to the average joe non-audiophile ear(you know, raised on the gospel of Bose) i believe it should smoke regardless.

as objectively as possible, just how "funny" does a Karlson get beyond 1000 anyway?

also, are there any issues associated with heavy attentuation? i'm not worried about overall efficiency loss because my woof is likely to be in the 96 dB neighbourhood, plus a zobel if it will affect efficiency much. i was thinking if i have to drop it by about 9 dB, i'd bring it down by 7 dB with a fixed L-pad, then have a variable pad(assuming +/- 2 dB? parts express doesn't specify the range) so that all the way up it would have 2 dB more than the woofer, at halfway it would be in line with the woofer, and when turned down it would be attentuated below "balance".
 

GM

Member
Joined 2003
>thanks for the reply!
====
Greets!

You're welcome!
====
>as objectively as possible, just how "funny" does a Karlson get beyond 1000 anyway?
====
My experience with these was back in the mid-late '50s, so my memory's dim now, but it mirrors the physics of the situation in that these are basicly bandpass (BP) alignments with four resonant cavities, so even with the K-slot to increase/'smooth' its HF output it still is highly resonant, so it's just a series of decreasing amplitude peaks/nulls with increasing frequency. Above ~500 Hz our hearing acuity becomes good enough to notice it in a big way, so pretty much mine and some other's limit WRT a XO point, though the hardcore 'Karlsonites' love the sound XO'd as high as 1.6 kHz when loaded with vintage Altec duplex (coincident) studio monitor drivers, so as always YMMV.

Then there's the issue of dispersion. Due to the K-slot, it holds a wide dispersion centered around ~120 deg over much of its HF BW and doesn't collapse to 90 deg until ~700-800 Hz according to some plots I've seen, so in this respect the nominal 120 deg Smith style multi-cell horn is a pretty good match if made large enough to work all the way down to ~250 Hz, or a mouth around 58.2"/41.8"/33.3" wide for a 1"/1.4"/2" exit driver when measured straight across. Depth would need to be at least ~13.56" plus driver/lens transition, so figure ~17" plus the driver's length.

At higher XO points then, ideally a narrower, but longer lens is required to get a good blend at the XO point, but may be too 'tight' to get the desired coverage pattern.

Bottom line, you're going to have to build at least one K and try various horn lens/XO point/slope combinations to find out what works for you in your app(s), something I gather you have neither the time/$$ to do.
====
>also, are there any issues associated with heavy attentuation?
====
Sure! It won't play as loud as the woofer due to such disparate power handling capability, so typically either only highly compressed source material is used or you use a compressor to do it. For example, the D210Ti's 160 W/1.5 kHz rating drops to ~22.4 W/500 Hz, so can handle only ~13.5 dB peaks above its attenuated efficiency, ergo at 96 dB it can hit ~109.5 dB/m. Rock music is typically compressed to ~15-20 dB, so the average would be as low as ~89.5 dB/m plus another ~3 dB for a stereo pair. As frat party levels go, this is pretty low, or at least it was back before you were even a gleam in your daddy's eye. ;) Altec A7 'Voice of the Theater' combo horn systems was the norm then (at least in my area), so worst case was ~103 dB/m/stereo avg. with a relatively high power tube amp (40 W). :)

Note though that between the fact that the LF output typically rolled off below 60-80 Hz and the A7's shelved response below ~200 Hz, we weren't overly concerned about running out of the 15" woofer's excursion even with some midbass boost to get those 'live' sounding kick drum/bass guitar 'hits'. :D

GM
 
I use the Smiths from 1200Hz to 7KHz.

It is usable to 13-14KHz, so you will need a UHF driver.

I got a JBL 2425J a few months ago for $58 on eBay...

The Horn/Lens combo is JBL HL91.

Here's the last measurement I did. Smith horn on JBL LE85 with JBL D16R2425 dia.
 

Attachments

  • smith1200_le85_horn2.jpg
    smith1200_le85_horn2.jpg
    76.8 KB · Views: 1,073
i guess my philosophy has to change some more if i can't achieve full HF extension with a Smith. this leaves me with a few choices...all of them indicating that a 3-way topology is inevitable:

1) scrap smith altogether and go back to conventional tweeter + mid - don't want to do this because i don't want to lose the dispersion so i am essentially putting this idea on the back burner altogether.

2) switch smith to being a mid unit. add slot tweeter like Selenium ST324(i always wanted to use this thing because it "looked cool"...might have a justification now since it has a dispersion of 120 degrees to match the Smith) to cover the range and accomplish original goal of crossing woof below 1 Khz. this does, however mean that i have to design my own "big smith" which means i'm going to be going to uncharted waters with it. like this is my first DIY project and first major woodworking project in a while since rebuilding a long section of fence 2 years back!

3) just get a damn coaxial.

i'm leaning towards number two because while it's a risk the payoff could be huge...but the first one seems so easy...

i guess i'll just have to build the Karlsons first to find out how it blends...
 
yggdrasil or GM,

I'm curious about the HF rolloff with the Smiths. Is it a function of the gap between the upper and lower sections of the horn?

Also, is the concept of the Smith usable as a simple waveguide without necessarily loading the driver for low end boost? I've been tossing around ideas for a way to get >150 degrees of horizontal dispersion without going to a Geddes style wave guide and without needing to load the driver. I want to use a Fountek ribbon above 2.5-3khz which is well within it's XO range, but I'd like to get as much horizontal dispersion as possible (180 degrees would be great :)) and I'd like to keep the natural HF extension of the ribbon.

I'm probably asking too much. :smash:
 
I would say a 2-way with a 15" at the bottom is very difficult to achieve. You would probably want to cross it over below 1KHz.

JBL have made a studio monitor 4430 which has a 15" low end and a 1" compression driver on a constant directivity horn for the HF. In order to achieve flat response equalization is built into the passive crossover network. You can find lots of info here: http://audioheritage.org/vbulletin/showthread.php?t=3838 You will also find alternative horns / wave-guides.

I have tried a 2-way Altec drivers about a year ago. I ended up adding a UHF driver. The details was lacking above 10KHz.

Have you considered the Altec 811 horn? Usable from 800Hz on a 1" driver.

AJ:
I will have to guess here, but my Smith horns consist of 6 different sections distributed horizontally. The rolloff might come from phase cancellations between the sections. The smaller Westlake horn does not have sections and is flat to 16KHz.

My Smith horns have 130 degrees horizontal disperson, however at 60 degrees off center they start rolling off at 8KHz.

I am not familiar with ribbons, so I cannot really offer any advice regarding to do or not to do.

I have been playing with smaller Smith horns for JBL UHF drivers. It seems to have potential, but there are still some work to do.....
 
hmm, well, i think after a lot of rambling and idea-tossing, i've had to cut it down to two possible 3-way solutions:

If i am to salvage the Smith idea, i will have to use it as a midrange with a more suitable driver. the best one for the job looks like the Selenium D250, sold in pairs on parts express. $70 for the pair means each driver is only $35. since the D250 works down to 400 Hz and even recommends a 500 Hz crossover, i could design a bigger Smith then kill it at 8 Khz and use the Selenium ST322 or 324 from 8 Khz on up. this is about the only super-tweeter that looks viable to match with the wide dispersion of the horn because being a slot-loaded design, it also claims 120 degree dispersion. i think it would otherwise be a little weird if only the midrange was broadly distributed.

pros: original goal of wide dispersion can still be accomplished, high efficiency, unconventional solution which "looks cool"

cons: have to design and build my own horn, which will be large and result in large and somewhat ungainly looking horn parked atop my Karlsons, also will nead to heavily attentuate tweeter and mid-horn to bring in line with comparatively inefficient(96 dB!) woofer

otherwise, i could go back to my original original plans which specified a cupped-back midrange and Dayton PT2B planar tweeter be mounted opposite each other in the wings of the Karlson.

pros: easy, no major attentuation issues, get to use planar tweeters(i like planars!), pretty cheap, retains original Karlson box shape without any extraneous things poking out

cons: ultra-wide dispersion sacrificed, will have to potentially attentuate woof by 2 dB to bring in line with mid and tweet thus resulting in overall system efficiency of ~94 dB, was hoping to achieve ~97 dB but may have to let that go.

am i just getting ahead of myself? i think i am likely to simply drop the smith project altogether.
 
i think i am likely to simply drop the smith project altogether.
Having built DSH horns, and listened to JBL "Smith" horns, I'd agree dropping them is a good idea.

The wide dispersion can be appealing, but the diffraction, comb filtering, and uneven vertical dispersion vs. frequency are all an annoyance.

A simple square or rectangular conical horn with the throat smoothed out can be built easier than a DSH, sound better, and provide constant directivity.

I won't say much about Karlson other than it was the first cabinet I ever built, and the first cabinet I ever sold.
I was glad I recovered the cost of the speaker in it when it sold.

Also was glad when Freddiy recently bought the 8" pair I inherited from my dad's early 1960's build.
 
hmm, well, i think after a lot of rambling and idea-tossing, i've had to cut it down to two possible 3-way solutions:

If i am to salvage the Smith idea, i will have to use it as a midrange with a more suitable driver. the best one for the job looks like the Selenium D250, sold in pairs on parts express. $70 for the pair means each driver is only $35. since the D250 works down to 400 Hz and even recommends a 500 Hz crossover, i could design a bigger Smith then kill it at 8 Khz and use the Selenium ST322 or 324 from 8 Khz on up. this is about the only super-tweeter that looks viable to match with the wide dispersion of the horn because being a slot-loaded design, it also claims 120 degree dispersion. i think it would otherwise be a little weird if only the midrange was broadly distributed.

pros: original goal of wide dispersion can still be accomplished, high efficiency, unconventional solution which "looks cool"

cons: have to design and build my own horn, which will be large and result in large and somewhat ungainly looking horn parked atop my Karlsons, also will nead to heavily attentuate tweeter and mid-horn to bring in line with comparatively inefficient(96 dB!) woofer

otherwise, i could go back to my original original plans which specified a cupped-back midrange and Dayton PT2B planar tweeter be mounted opposite each other in the wings of the Karlson.

pros: easy, no major attentuation issues, get to use planar tweeters(i like planars!), pretty cheap, retains original Karlson box shape without any extraneous things poking out

cons: ultra-wide dispersion sacrificed, will have to potentially attentuate woof by 2 dB to bring in line with mid and tweet thus resulting in overall system efficiency of ~94 dB, was hoping to achieve ~97 dB but may have to let that go.

am i just getting ahead of myself? i think i am likely to simply drop the smith project altogether.

I achieved surprisingly good results when I took a Paraline and basically sawed it in half. It works a lot like a Smith horn when you do this:

1) The width of the horn is dramatically larger than the height.
2) Due to the above, the vertical directivity is very wide and the horizontal is narrow. (Yes, that's counter-intuitive, but that's how directivity works.)
3) High frequency extension was surprisingly good. I was getting output all the way to 20khz, and it was pretty darn smooth.


Arguably the main reason I gave up on the project was sheer ADHD. It worked pretty well. If I had to complain about any aspect of the project, it was that the cabinet wouldn't 'disappear' the way that my Summas do. I believe this is because the Summas have much lower diffraction at the mouth.

As Art noted, if you have the space, a plain ol' conical would likely be superior. (Most of my projects are designed for the car, where space is tightly constrained.)

I posted pics and measurements of my half-a-paraline on this forum. This was about 2 months ago IIRC, just check out my posting history.
 
On the topic of ribbons- I have the fountek neoC3.5H and im very happy with it. It has a good sensitivity so they should fit your 3way.
Since your mids could reach to 7k, id cross the ribbons in a little higher, say 6k?

This is how i intend to use them in a compromised 3way im planning using drivers i have collecting dust for me.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.