And now for something Kompletely different!
As an exercise in thinking outside the box, I have decided this will be my next project--a three-way high-efficiency system based on the Karlson Tube broadband-resonator principle--a bit of audio arcana harking back about 60 years.
Midbass: 4" Sch40 PVC pipe ~80" long with a ~40" exponential slot. Driver: JBL 2123h in ~7L sealed. Bandwidth target: 80-350Hz.
Midrange: 2" Sch40 PVC pipe ~20" long with a ~14" exponential slot. Driver: JBL 2485h. Bandwidth target: 350-3kHz.
Tweeter: Transylvania Power Company "The Tube". Or, I may fab up a PVC coupler. Driver: Radian 475PB. Target bandwidth: 3-20kHz.
The system will be angled at ~35 degrees in accordance with John E. Karlson's findings about the optimal listening axis for tube couplers. I may end up spraying the PVC with automotive rocker guard to deaden any resonances. During development, these will be multi-amped through my Rane RPM88 processor. All drivers are in inventory.
This is my first foray into these mystikal Karlson devices, and it promises to be fun. I'm interested to hear the mix of claimed attributes--high efficiency, wide horizontal dispersion and rapid impulse decay (compared to most horns), all in a very slim footprint.
Below is my rough koncept drawing. I've already rough-cut two options for the midrange tube, and the rest of the PVC is waiting in my garage.
As an exercise in thinking outside the box, I have decided this will be my next project--a three-way high-efficiency system based on the Karlson Tube broadband-resonator principle--a bit of audio arcana harking back about 60 years.
Midbass: 4" Sch40 PVC pipe ~80" long with a ~40" exponential slot. Driver: JBL 2123h in ~7L sealed. Bandwidth target: 80-350Hz.
Midrange: 2" Sch40 PVC pipe ~20" long with a ~14" exponential slot. Driver: JBL 2485h. Bandwidth target: 350-3kHz.
Tweeter: Transylvania Power Company "The Tube". Or, I may fab up a PVC coupler. Driver: Radian 475PB. Target bandwidth: 3-20kHz.
The system will be angled at ~35 degrees in accordance with John E. Karlson's findings about the optimal listening axis for tube couplers. I may end up spraying the PVC with automotive rocker guard to deaden any resonances. During development, these will be multi-amped through my Rane RPM88 processor. All drivers are in inventory.
This is my first foray into these mystikal Karlson devices, and it promises to be fun. I'm interested to hear the mix of claimed attributes--high efficiency, wide horizontal dispersion and rapid impulse decay (compared to most horns), all in a very slim footprint.
Below is my rough koncept drawing. I've already rough-cut two options for the midrange tube, and the rest of the PVC is waiting in my garage.
Attachments
Last edited:
Hey, you just reminded me I have pair of 2407s somewhere in my stash, too! Good idea about trying them in K-tubes. I've heard from several who say treble K-tubes sound subjectively clearer and more open than horns/waveguides they've tried...
K15s seem interesting, but I get a sense there's a lot of complicated physics going on there with the reflex loading/etc. modified by the expo. slot.
To me, the K-tube seems like the simplest, most fundamental application of JEK's concept, and that's why I gravitate toward it, even for midbass. I can actually start to wrap my mind around how they might function.
Last edited:
I've wanted to make a pair of tweeter K-tubes for years, but have been to lazy to do it. Your project is way beyond that and very inspiring. Will be following closely to hear what happens.
I'll eat my hat! I'm just in the process of finishing a big waveguide project that I want to be the best within my means for a long time. If something as easy to implement as the tube sounds better there will be gnashing of teeth!
That they sound good at all confuses me. For every speaker I had built I invested great care in reducing diffraction, standing waves and trying to eliminate anything bouncing back into the driver's cone. That the Ks sound good with drivers playing midrange inches off the floor through a wooden curtain is above my head.
While I have heard very good sounding KTubes less than 1/2inch diameter,
all variations I have tried above that size, sounded like carp (yes the fish...)
Neither Selenium midrange driver on 2" PVC, nor Electrovoice 10" cone on
concrete forming tube could be made (by my efforts) to sound good at all.
Anyways: One thing I didn't try with bigger KTubage, but Karlson did...
He coaxed a small end driven Ktube inside the larger tube, driven from
some midpoint where the small tube exits and not driven from the end.
Maybe the synergy of different driving point locations was significant?
all variations I have tried above that size, sounded like carp (yes the fish...)
Neither Selenium midrange driver on 2" PVC, nor Electrovoice 10" cone on
concrete forming tube could be made (by my efforts) to sound good at all.
Anyways: One thing I didn't try with bigger KTubage, but Karlson did...
He coaxed a small end driven Ktube inside the larger tube, driven from
some midpoint where the small tube exits and not driven from the end.
Maybe the synergy of different driving point locations was significant?
Last edited:
I'll eat my hat! I'm just in the process of finishing a big waveguide project that I want to be the best within my means for a long time. If something as easy to implement as the tube sounds better there will be gnashing of teeth!
No worries. Baranek's Law will work strongly in both our favors, and both approaches will sound amazing, I'm sure.
Hi Ken,
I've been thinking about half-inch-diameter treble k-tubes, too. Many people have reported that the 1" dia. Transylvania tube works very well up high, but in the process of trying to apply principles I already know to the mystery of the Karlson Tube, I thought it might be best to use an interior tube diameter that is ideally no more than a quarter wavelength of the highest frequency it will pass, or a practical compromise might be less than half a wavelength. This should keep the majority of the radiation firing @ ~30 degrees out the slot and not down the axis of the tube. Should minimize the formation of standing waves and also result in the most stable directivity.
Of course, wavelengths shrink quickly in the treble region (20kHz wavelength = 0.66"), and I don't know of any compression drivers with throats less than half an inch across, so that's about the best we can do . I ended up buying a used JBL 2404 with a broken "baby cheeks" waveguide, which I discarded. I plan to try it on a half-inch-dia. k-tube at some point. Another option might be the B&C DE5.
I'm interested in your experience with mid and midbass k-tubes. Did you take any FR measurements? Do you recall the tube lengths and the lengths of the slots?
I've seen nice, flat curves published, and I've seen some with deep notches. It seems to me that the notches often appear when you try to get too much bandwidth out of the tube and take it up past the point where half-wavelength<diameter. It might be helpful to think of these more like horns, and not expect more than ~3 octaves out of them.
Last edited:
Yeah, it does look a bit like Dr. Seuss' trombone, doesn't it? My wife and oldest son (5) both think it looks like some kind of futuristic weapon. "Hey," I tell them, "It's just form following function!"
I will certainly be watching for signs of mechanical resonance, but I don't expect it to be a problem. First of all, PVC seems to have a fair amount of self damping and doesn't seem to ring--nowhere near as much as metal, anyway. Also, I believe the exponential slot should greatly spread and reduce the Q of any mechanical resonances in the tube itself. At this point, unless a compelling reason arises, I don't have plans to audition other materials--mostly because it sounds like a lot of work. I'd rather focus my time on trying, measuring and comparing different tube and slot lengths.
Regarding air-column resonances, the tube's throat section before the slot should present a significant load to the driver diaphragm, thus providing gain. If the tube then ended abruptly, the entire output would be dominated by a series of air-column resonances.
But this is where the Karlson magic come in: As a pressure wave travels down the K-tube, an acoustic impedance transformation takes place as it encounters the narrow part of the long exponential slot and is progressively released as the slot widens. This progressive release of the pressure wave means that it does not "see" a distinct end to the tube, so it doesn't reflect back and create standing waves.
This is my understanding of the the theory, anyway.
"Hey," I tell them, "It's just form following function!"I will certainly be watching for signs of mechanical resonance, but I don't expect it to be a problem. First of all, PVC seems to have a fair amount of self damping and doesn't seem to ring--nowhere near as much as metal, anyway. Also, I believe the exponential slot should greatly spread and reduce the Q of any mechanical resonances in the tube itself. At this point, unless a compelling reason arises, I don't have plans to audition other materials--mostly because it sounds like a lot of work.
I'd rather focus my time on trying, measuring and comparing different tube and slot lengths.Regarding air-column resonances, the tube's throat section before the slot should present a significant load to the driver diaphragm, thus providing gain. If the tube then ended abruptly, the entire output would be dominated by a series of air-column resonances.
But this is where the Karlson magic come in: As a pressure wave travels down the K-tube, an acoustic impedance transformation takes place as it encounters the narrow part of the long exponential slot and is progressively released as the slot widens. This progressive release of the pressure wave means that it does not "see" a distinct end to the tube, so it doesn't reflect back and create standing waves.
This is my understanding of the the theory, anyway.
Last edited:
the deep notch I saw in the X15 system tube seemed to be from a diameter transition as the pipe was 1.875" ID. One inch pvc pipe with a 5.25" long slot makes a good sounding and smooth measuring K-tube for one inch compression drivers. There can be a goodly length of unslotted pipe before the 5.25" k-tube to allow more reach (see Faerber Acoustics). Robert Reams used a 12" (probably EV12L) on the "Clothespin" tube which was part of the Karlson-Hypex system.
thinwall 1" pvc pipe vs a 9" long x 1.875" tube made by Carl N. Carl's tube has the slotted section pre-waveguide
used on some examples of X15's tube - -the reason these graphs didn't go high was the nature
of the old Eminence CD8 compression driver.
I have heard some 1" K-tube which I didn't like - if the slot opens too fast and one steel pipe where it seemed too long
Transylvania Power's K-tube is a good one.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
thinwall 1" pvc pipe vs a 9" long x 1.875" tube made by Carl N. Carl's tube has the slotted section pre-waveguide
used on some examples of X15's tube - -the reason these graphs didn't go high was the nature
of the old Eminence CD8 compression driver.
An externally hosted image should be here but it was not working when we last tested it.
I have heard some 1" K-tube which I didn't like - if the slot opens too fast and one steel pipe where it seemed too long
Transylvania Power's K-tube is a good one.
Last edited:
Do you know the dimensions of the "Clothespin" coupler? I've always wondered about it, because its design doesn't make much sense to me. Mostly, I wondered about the dual slot, which seems to me like it would be less effective than a single-slot tube. Have you ever seen any acoustic measurements of it?
So you were driving that 2" K-tube with a 1" driver?? I guess that would be a recipe for notchy chaos!
That's really interesting that you can run a long stub section before the slot. Sort of like a plane-wave tube, I guess; the waves don't care how far they go as long as there's no sharp reflections from the terminus. Could even stuff it with some open-cell foam and see if there's any benefit...
Does a long stub section significantly lower the system's Fs?
So you were driving that 2" K-tube with a 1" driver?? I guess that would be a recipe for notchy chaos!
That's really interesting that you can run a long stub section before the slot. Sort of like a plane-wave tube, I guess; the waves don't care how far they go as long as there's no sharp reflections from the terminus. Could even stuff it with some open-cell foam and see if there's any benefit...
Does a long stub section significantly lower the system's Fs?
Last edited:
hi Bill F.
clone of X15 tube - the original University 1" format driver screwed into a phenolic block - there was a phenolic wafer which held a "pre-wave-guide" (I have run a 1" to 2" conical horn adaptor to that tube and IIRC, it played smoother on the bottom but didn't sound as good as the weird transition)
here's the impedance of a BMD440 driver on "The Tube" vs 11" of pipe taped to a 5.5" K-tube-there's a posiibility the ripple in high frequency Z of the long pipe is from it being jammed part-way into a 2-bolt to 1/3/8" horn adaptor - - note the low frequency peak
I've got one Clothespin - its 9.75"ID, 10.25"OD and 27" long to the two "tips" -only did a near-field of it sitting on a 10" speaker - to give it a chance, it should be mounted in a plywood baffle and run with a low qts 12" with a sealed back chamber. Years ago I made a couple of traced posterboard templates - don't have them anymore and rarely get rides anywhere to pick up the simplest of materials. A regular klam-projector might make sense for mid-bass use - there might be some combination of angle and dimensions to get one smooth. Carl has built some "Rocket" style mini-klams - don't know it he has made larger (will ask).
to me, regular Karlson's make the most sense in the least amount of space/cabinet depth. Drivers should have low mass and good strength.
clone of X15 tube - the original University 1" format driver screwed into a phenolic block - there was a phenolic wafer which held a "pre-wave-guide" (I have run a 1" to 2" conical horn adaptor to that tube and IIRC, it played smoother on the bottom but didn't sound as good as the weird transition)
An externally hosted image should be here but it was not working when we last tested it.
here's the impedance of a BMD440 driver on "The Tube" vs 11" of pipe taped to a 5.5" K-tube-there's a posiibility the ripple in high frequency Z of the long pipe is from it being jammed part-way into a 2-bolt to 1/3/8" horn adaptor - - note the low frequency peak
An externally hosted image should be here but it was not working when we last tested it.
I've got one Clothespin - its 9.75"ID, 10.25"OD and 27" long to the two "tips" -only did a near-field of it sitting on a 10" speaker - to give it a chance, it should be mounted in a plywood baffle and run with a low qts 12" with a sealed back chamber. Years ago I made a couple of traced posterboard templates - don't have them anymore and rarely get rides anywhere to pick up the simplest of materials. A regular klam-projector might make sense for mid-bass use - there might be some combination of angle and dimensions to get one smooth. Carl has built some "Rocket" style mini-klams - don't know it he has made larger (will ask).
to me, regular Karlson's make the most sense in the least amount of space/cabinet depth. Drivers should have low mass and good strength.
Yeah, I suspect the extra ripple is from the non-optimal connection again...
That low peak at ~200Hz is very interesting. Guess what the quarter wavelength of 200Hz is? That's right, 16.5"--the exact length of your stub pipe + coupler! I love it when things make sense.
It seems to me that, aside from all the cool things K-tubes do higher in their passband, down at the bottom of their response, they're pretty much just quarter-wave transmission lines with a "fuzzy" terminus. So my instinct is to start my exploration using the following guidelines:
1) K-Tube length =/> quarter wavelength of lowest frequency.
2) K-Tube diameter =/< quarter wavelength of highest frequency. (Or compromise at a half wavelength.)
3) Smooth transition from driver throat to tube--it seems to make a real difference.
4) Basically think of them as horns and don't expect extremely wide bandwidth.
5) Drive them with horn-appropriate transducers.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.