Hey Guys,
Thanks for the suggested crossover points+slopes and informed perspectives about the box. I've learned a lot and deeply appreciate the sharing.
I'm not yet in a position to speak from experience, but having read almost everything I could find about Karlsons over the past couple years I've noticed there is a tendency to discount the cabinet for it's flaws culminating in the label "party speaker". I think this tendency has been made worse as the cabinet is a "multi-mode resonant system" (which I understand some feel is primarily a BP alignment), but we can probably all agree there are more principles in play affecting Karlson performance and that speakers without predictive models telling us how to scale, tune or match them to drivers relegates said speaker designs to tinkerers, puzzle hounds and kults.
Freddi's points about responses of enjoyed/esteemed alignments being choppy, having worse group delay, longer transients, worse efficiency, less cone control, more distortion, etc, seem valid.
Further Karlson's are just now on the edge of having a working predictive model that allows T/S to guide our implementations.
The status quo years ago was horns "Honk", yet refinements of expansions, materials and flares brought about a new hi-fi horn movement that has advanced old ideas into very refined performers. I can imagine we will someday even see reworkings of classics like the WE15a using composite materials without parallel sidewalls.
The Karlson is a unique combination of alignments with it's own benefits as demonstrated tirelessly by Freddi and others who seem to enjoy them quite a lot. My instinct says the Karlson is going from cantankerous and unpredictable towards reliable, maybe even reasonable thanks to xrk971's simple but seemingly effective (multi-resonant waveguide modeling approach
)
If i was a speaker I would rather start out a party and end up an equation than the other way around.
I can see the arguments for the Karlson and am exploring it in both traditional and newly optimized topologies; and I may quickly find myself in the critics camp as I require any design to hold it's own against other less complicated alignments, but on paper it looks like it does {in specific scenarios} -> PA use being the primary and compact upper bass horn being another. Still, full range looks a bit dodgy yet In fairness I wouldn't be caught using a reflex cab in the studio, but many mixers get along fine with them.
Hopefully some of my sawdust and prototyping will contribute to a refinement that GM and other critical ears could enjoy.
Thanks for the suggested crossover points+slopes and informed perspectives about the box. I've learned a lot and deeply appreciate the sharing.
I'm not yet in a position to speak from experience, but having read almost everything I could find about Karlsons over the past couple years I've noticed there is a tendency to discount the cabinet for it's flaws culminating in the label "party speaker". I think this tendency has been made worse as the cabinet is a "multi-mode resonant system" (which I understand some feel is primarily a BP alignment), but we can probably all agree there are more principles in play affecting Karlson performance and that speakers without predictive models telling us how to scale, tune or match them to drivers relegates said speaker designs to tinkerers, puzzle hounds and kults.
Freddi's points about responses of enjoyed/esteemed alignments being choppy, having worse group delay, longer transients, worse efficiency, less cone control, more distortion, etc, seem valid.
Further Karlson's are just now on the edge of having a working predictive model that allows T/S to guide our implementations.
The status quo years ago was horns "Honk", yet refinements of expansions, materials and flares brought about a new hi-fi horn movement that has advanced old ideas into very refined performers. I can imagine we will someday even see reworkings of classics like the WE15a using composite materials without parallel sidewalls.
The Karlson is a unique combination of alignments with it's own benefits as demonstrated tirelessly by Freddi and others who seem to enjoy them quite a lot. My instinct says the Karlson is going from cantankerous and unpredictable towards reliable, maybe even reasonable thanks to xrk971's simple but seemingly effective (multi-resonant waveguide modeling approach
)If i was a speaker I would rather start out a party and end up an equation than the other way around.
I can see the arguments for the Karlson and am exploring it in both traditional and newly optimized topologies; and I may quickly find myself in the critics camp as I require any design to hold it's own against other less complicated alignments, but on paper it looks like it does {in specific scenarios} -> PA use being the primary and compact upper bass horn being another. Still, full range looks a bit dodgy yet In fairness I wouldn't be caught using a reflex cab in the studio, but many mixers get along fine with them.
Hopefully some of my sawdust and prototyping will contribute to a refinement that GM and other critical ears could enjoy.
Last edited:
I actually own and have played with K15 so am reasonably familiar. Changes in aperture flare and gap can have a lot of subjective impact - at least up close and those don't register easily on a graph - probably partly due to the large source, delays, and asymmetrical patterns. K15 may be difficult to improve upon on what it does well. The front shelf does tend to make it somewhat different than a K-coupler which deletes that feature. i need to find someone game to cut either a Karlsonator 8 or 12 for testing and evaluation. Carl likes a curved reflector and says that improved his version of the X15 vs an approximation. i think the curved reflector focus may be critical and as a poor woodworker am leery of them.
If would be nice to get a handle on what front chamber proportions, contour and vent positioins create the smoothest response. If the Acoustic Control 115BK can scale to k15's bulk successfully then that might be pretty cool.
For those without a coaxial, a K-tube treble solution is a natural whether placed within the coupler or mounted on top.
K15 can do things which would leave a reflex gasping plus its "tuneful" - -- I'm not sure how a Kilomax would fare in K15 for "tone".
K12 of course sounds like a Karlson, and has visceral tangibility to tone if required - but K12 is something like a front load horn in extension and could benefit from a subwoofer. K12's ~0.6 cubic foot coupler does a lot for such a small volume.
K8 pushed the design almost past limits and doesn't have anywhere near the power of an 8" in K12. FE206EN is good for K12 size.
If I can feel better, I'll try to record some CD audio of a K12 with soprano and harp plus a recording of a MIDI keyboard emulating a guitar
If would be nice to get a handle on what front chamber proportions, contour and vent positioins create the smoothest response. If the Acoustic Control 115BK can scale to k15's bulk successfully then that might be pretty cool.
For those without a coaxial, a K-tube treble solution is a natural whether placed within the coupler or mounted on top.
K15 can do things which would leave a reflex gasping plus its "tuneful" - -- I'm not sure how a Kilomax would fare in K15 for "tone".
K12 of course sounds like a Karlson, and has visceral tangibility to tone if required - but K12 is something like a front load horn in extension and could benefit from a subwoofer. K12's ~0.6 cubic foot coupler does a lot for such a small volume.
K8 pushed the design almost past limits and doesn't have anywhere near the power of an 8" in K12. FE206EN is good for K12 size.
If I can feel better, I'll try to record some CD audio of a K12 with soprano and harp plus a recording of a MIDI keyboard emulating a guitar
Last edited:
choices to sort through for new K-coupler includes:
pick of baffle angle - that determines coupler volume, on=axis treble lobe. Karlson stated that more tilt gave: lower bass, worse transient response, more "reverb"
pick of front chamber height - what is the optimum height vs potential cavity peaking?
rear chamber volume - what gives the best tradeoffs? - in a regular Karlson. the rear chamber is sometimes picked as ~ (Vas*qts*fs)/50 - the beauty of the original 15, 12 and X15 is a lot of punch and definition with very little cone motion.
pick of front chamber contour - do we want a fairly constant cross-section as with the 1954 K12? - cant the port panel 10 degrees per 1956-onwards K12? - or a 'bass clef" shaped chamber?
pick of upper reflector - in GB's "Karlsonator", the reflector is deleted.
vent position - what's the best place to put the vent? -
in practical K's, de-Q-ing vent techniques can make a speaker system sound "faster" - perhaps at the expense of being able to play pure sine tones.
rear lowpass shelf - this can be used to improve subjective transient response and would be probably adjusted with drums, plucked and bowed bass viola. K15 had a 2" damping pad under the shelf to clean up some stuff before it reached K15's port.
front shelf - this feature disappeared in the smaller couplers when the first K12 made its debut in the fall of 1954. I have had a 0.8 coupler scaled from k15 and it sound and measurements were pretty good. The front shelf tends to segregate some treble from reaching the upper portion of the front coupler.
The K-tube waveguide should almost always be considered before picking a horn or CD waveguide when the crossover point is high enough to allow a K-tube. Carl uses a K-tube tweeter to augment his Fostex Sigma 208. He's a rocker so the 208 is of limited interest as it will poop out on the drumkit tests.
Karlson's X15 tube at 1.875" diameter measured rough but sounded "good" - it was not the only size which he experimented. This tube had a internal scattering device which was 1" diameter and sliced 30 degrees back. For the sake of smoothness, a 1" inside diameter pipe ~5.3" long with half-ellipse slot will measure much better. A one inch K-tube can be extended to 7" or so to fit onto a K's port panel. The old X15 with its internal tube sounds very fast and holographic - just one cabinet sounds pretty convincing on rainstorm recordings.
btw, Carl evaluates several sets of CNC cut wings - will have to see what he's working on.
even an imperfect K is often better overall than some other approaches - I would not be scared at seeing a speaker mounted in a K as have heard worse from drivers flush mounted to a baffle. hahaha
X- do you have space for a small table saw? - there's some cheapies but they would have to have diy extensions to handle a full sheet
pick of baffle angle - that determines coupler volume, on=axis treble lobe. Karlson stated that more tilt gave: lower bass, worse transient response, more "reverb"
pick of front chamber height - what is the optimum height vs potential cavity peaking?
rear chamber volume - what gives the best tradeoffs? - in a regular Karlson. the rear chamber is sometimes picked as ~ (Vas*qts*fs)/50 - the beauty of the original 15, 12 and X15 is a lot of punch and definition with very little cone motion.
pick of front chamber contour - do we want a fairly constant cross-section as with the 1954 K12? - cant the port panel 10 degrees per 1956-onwards K12? - or a 'bass clef" shaped chamber?
pick of upper reflector - in GB's "Karlsonator", the reflector is deleted.
vent position - what's the best place to put the vent? -
in practical K's, de-Q-ing vent techniques can make a speaker system sound "faster" - perhaps at the expense of being able to play pure sine tones.
rear lowpass shelf - this can be used to improve subjective transient response and would be probably adjusted with drums, plucked and bowed bass viola. K15 had a 2" damping pad under the shelf to clean up some stuff before it reached K15's port.
front shelf - this feature disappeared in the smaller couplers when the first K12 made its debut in the fall of 1954. I have had a 0.8 coupler scaled from k15 and it sound and measurements were pretty good. The front shelf tends to segregate some treble from reaching the upper portion of the front coupler.
The K-tube waveguide should almost always be considered before picking a horn or CD waveguide when the crossover point is high enough to allow a K-tube. Carl uses a K-tube tweeter to augment his Fostex Sigma 208. He's a rocker so the 208 is of limited interest as it will poop out on the drumkit tests.
Karlson's X15 tube at 1.875" diameter measured rough but sounded "good" - it was not the only size which he experimented. This tube had a internal scattering device which was 1" diameter and sliced 30 degrees back. For the sake of smoothness, a 1" inside diameter pipe ~5.3" long with half-ellipse slot will measure much better. A one inch K-tube can be extended to 7" or so to fit onto a K's port panel. The old X15 with its internal tube sounds very fast and holographic - just one cabinet sounds pretty convincing on rainstorm recordings.
btw, Carl evaluates several sets of CNC cut wings - will have to see what he's working on.
even an imperfect K is often better overall than some other approaches - I would not be scared at seeing a speaker mounted in a K as have heard worse from drivers flush mounted to a baffle. hahaha
X- do you have space for a small table saw? - there's some cheapies but they would have to have diy extensions to handle a full sheet
IG81, Magnetar, GM and others have expressed that hi-fidelity use of the K15 is best when it's crossed below the notch... The notch I've been trying to figure out how to move lower
Since I have some direct experience in this, I suppose a comment or two is in order.
Like other people, based on the FR, I assumed that a K15 would be better off if crossed over just before that first notch. However, after actually building a pair, the reality was a little different. I tried many different schemes and crossover points.
In practice, for home use, simpler turned out to be better. My favorite was a basic 2 way with a comp driver and K-tube on top, crossed over at 1.2kHz. Though it doesn't look that impressive on paper, the variable slot effect of the front K chamber produces a nice cohesive soundfield. This combination had a smooth sound with wide dispersion and excellent integration. It reminded me a bit of an AR3 but with better dynamics and more detail. It'd be a shame to not use the midrange of this box. In fact, it is one of the strengths of the Karlson, IMO.
GM and I had some lengthy discussions about these at the time. (thanks again for the help GM, BTW) I tried at least some of those suggestions.
Of the various ideas proposed, the one that had the most potential was to place an upper midbass horn on above the K15, and then put a wideband midrange horn and a supertweeter horn in between the two. This provides plenty of overlap for ideal crossovers. A jury rigged test pile was quite impressive. The sound was endlessly powerful, effortless, and big. It'd be a good way to go for a crazy good HT system.
Of course club/PA use is a different application. If I was designing this outdoor dance party system, I'd use K15's for a foundation, and fly ~150hz front horns and HF horns above them. That's sort of a simplified version of the above system, and would rock.
Very few high efficiency speakers have super smooth frequency response anyway. It just goes with the territory. That's why there's EQ.
Greg B,
Thanks for sharing your experience with cross-overs on the K15. I want to use a 12 inch coaxial and the manufacturer specs say to use a 2.5 kHz XO, is that too high then?
Do you have a picture of what you describe here:
If you can give more details as to the horns used and the XO points that would be great.
Thanks,
X
Thanks for sharing your experience with cross-overs on the K15. I want to use a 12 inch coaxial and the manufacturer specs say to use a 2.5 kHz XO, is that too high then?
Do you have a picture of what you describe here:
If you can give more details as to the horns used and the XO points that would be great.
Thanks,
X
2k5 or even a bit higher likely will sound better on coaxial than 1K2->1K6 - some compression drivers will start sounding "hard" or "metallic" with lower xover points. For some 10->12cx, its kinda like 1K6 sounds too "hard", 3K5 "lacks substance" and treble power. 12LTA with a high helper tweeter is in the same boat for poor treble power response. Cap choices can be interesting in flavoring a CX. With EM's 2K5 xover values, the highpass can be two 6.8uF caps in series with battery bias to their center junction.
David Dicks recommends just a cap highpass with the P-Audio coaxial line. In that case a swamping resistor across the compression driver can help. Stan Ricker likes that kind of xover but it doesn't work well with everything. I suspect DD likes it as it can emulate the sound of the FR he sells. (and there's plenty of top with no attenuation)
I'm partly basing the 1K6 impressions on using Eminence's 1K6 network whose values are strictly textbook/resistive loads. Their 3rd order highpass comes in around 1K2 - their lowpass probably needs more inductance. I did once use some values suggested by GB which worked well on a 15cx but lost the notes.
I played with series networks and an elaborate elliptical "infinite slope" type network on an Eminence 15cx.
it might be interesting to look around for Beta 12cx crossover solutions and A-B compare to the 2mH/10uF -- 3.3uF/0.33mH scheme
K15 with a 15cx has fooled me into thinking a voice was "in the room" - the same recording on an 8" fullrange in a BK20 horn did not.
David Dicks recommends just a cap highpass with the P-Audio coaxial line. In that case a swamping resistor across the compression driver can help. Stan Ricker likes that kind of xover but it doesn't work well with everything. I suspect DD likes it as it can emulate the sound of the FR he sells. (and there's plenty of top with no attenuation)
I'm partly basing the 1K6 impressions on using Eminence's 1K6 network whose values are strictly textbook/resistive loads. Their 3rd order highpass comes in around 1K2 - their lowpass probably needs more inductance. I did once use some values suggested by GB which worked well on a 15cx but lost the notes.
I played with series networks and an elaborate elliptical "infinite slope" type network on an Eminence 15cx.
it might be interesting to look around for Beta 12cx crossover solutions and A-B compare to the 2mH/10uF -- 3.3uF/0.33mH scheme
K15 with a 15cx has fooled me into thinking a voice was "in the room" - the same recording on an 8" fullrange in a BK20 horn did not.
Last edited:
I was referring to the crossover point at which you would crossover the K15 to a different enclosure, horn, etc. A coaxial is inherently going to be running full range through the K-coupler, regardless of any crossover within its high and low freq components.
AFA the 2.5k crossover, it just depends on what sounds best for that driver. I've heard a few coaxials in K15 boxes, and they all had very different crossover points. The old EV TRX series had an integral whizzer, and a highish 3.5k crossover.
No pictures, sorry. It was a long time ago and I had no digital camera. Besides, it wasn't much to look at, mostly cardboard and duct tape 'test' boxes. It's a bit fuzzy, but I recall a ~300hz conical midrange horn in the middle, and about 100hz midbass on top. An RS 'pod' supertweeter was crossed in at about 9k or so, placed variously. If this sort of approach is taken, it does sort of lose the karlson sound, and effectively becomes a 4 way horn set up.
You’re welcome! Glad you followed through with some testing that I still haven’t been able to make time for and still help others a bit, so glad whoever ‘x’ is and others have ‘picked up the torch’ to demystify them.
Different strokes..... lower mids, yes, higher up…….neither my young teen nor the K15 owner's war damaged hearing could tolerate it loaded with a variety of inexpensive mid 50s 'FR' drivers with/without whizzer and/or mids diffusor driven with his various DIY tube amps, though it made then incredible [mid] bass, especially when corner loaded, so ultimately became just a 'sub' in a bi-amped system.
Freddy told me it was designed for the Altec 604 duplex or similar, though never seen any docs to suggest this, but in retrospect it makes sense and may have been a success if it had been marketed as such instead of a generic ‘one size fits all’ solution.
It was the K tube he showed me that makes the most sense to complement the various K alignments though, so seems like optimizing this combo should have the highest priority.
That, and adding more coupling chambers to both increase the K’s usable BW and increase its resonant structure to a higher Q, ergo sound more vibrant, yet smoother than the originals due to the way we perceive sound. It should still be smaller than a mid-bass horn or ~three octave tapped pipe/horn with the ~same lower corner frequency.
Anyway, this is where I was headed when I planned to help the late GregB some years ago, but ‘life’ happens, too often ‘paving my way to Hell with good intentions’.
GM
GM,
I am glad to have picked up the torch on this one.
I did not set out to do that, but after simulating the Karlson and seeing what it can do from an efficiency standpoint, I see why Freddi has been such a vocal proponent all these years. Now with a tool that lets me optimize and try out different drivers or T/S params, it has been a lot of fun to see what can be done to demystify it. I have not seen the mid-bass and mid-range performance of the K15 matched by the many different speaker alignments that I have modeled - it truly is a timeless design made in the day before any simulation. Thanks for all your help and tips on acoustic modeling along the way.Regards,
X
K15 (scaled down) works well with TB W8-1772 and Lowther PM6C
To get this established back in the proper thread and also just in case you may have missed this because it was sort of OT and posted in this thread: http://www.diyaudio.com/forums/full-range/162253-build-thread-k-slot-rear-loaded-horn-tangband-w8-1772-a-12.html
It turns out that the K15 can work really well (K15 cabinet scaled down to 12/15 x length and 8/15 x width) with the W8-1772 and PM6C. The sims show excellent bass extension down to mid 40's Hz and ideal impulse response meaning the dynamics will be excellent. You may do a double take seeing a Karlson and Lowther paired together - but they seems to be well-matched.
Here are the freq responses:
Tang Band W8-1772
Lowther PM6C
Similarly, the Audio Nirvana 10 works perfectly in a full-size K15
All of these sims are for placement very far away from walls and now include damping in the back chamber (different than the K15 plans which call for pads, I actually have stuffing in the upper chamber and lower chamber but none around the vent or rear shelf). It seems to help smooth out the response without impacting efficiency.
To get this established back in the proper thread and also just in case you may have missed this because it was sort of OT and posted in this thread: http://www.diyaudio.com/forums/full-range/162253-build-thread-k-slot-rear-loaded-horn-tangband-w8-1772-a-12.html
It turns out that the K15 can work really well (K15 cabinet scaled down to 12/15 x length and 8/15 x width) with the W8-1772 and PM6C. The sims show excellent bass extension down to mid 40's Hz and ideal impulse response meaning the dynamics will be excellent. You may do a double take seeing a Karlson and Lowther paired together - but they seems to be well-matched.
Here are the freq responses:
Tang Band W8-1772
Lowther PM6C
Similarly, the Audio Nirvana 10 works perfectly in a full-size K15
All of these sims are for placement very far away from walls and now include damping in the back chamber (different than the K15 plans which call for pads, I actually have stuffing in the upper chamber and lower chamber but none around the vent or rear shelf). It seems to help smooth out the response without impacting efficiency.
I was trying out an Eminence Kappa 15A driver as it seemed to be very high efficiency (100.5 dB) to see what you can do in a K15. I scaled the K15 in length by 18/15 to get a little more bass extension in this case. For grins, I only applied 300 mW of excitation (1.55 volts) and the SPL vs Freq is still very respectable at 1 m. We are in headphone amplifier territory here
Attachments
Last edited:
I don't think the Beta 8A is a good fit for the Karlson K15 (or K12) - there is a peaky bass punch. See simulation below. I think a better performance can be had in a Karlsonator scaled to 12/15x. Use the CD to fill in the top end above 2.5kHz with the standard Eminence crossover. Use K-tube horn or a wide angle 90 deg horn.
K15:
Karlsonator scaled 12/15x:
Attachments
Last edited:
I don't think the Beta 8A is a good fit for the Karlson K15 (or K12) - there is a peaky bass punch. See simulation below. I think a better performance can be had in a Karlsonator scaled to 12/15x. Use the CD to fill in the top end above 2.5kHz with the standard Eminence crossover. Use K-tube horn or a wide angle 90 deg horn.
K15:
Karlsonator scaled 12/15x:
I just started this thread today, so I'm not familiar with "scaled to 12/15X." Can you clarify that for me?
Another question I have concerns the cabinet. Is it open-back behind the driver(s)? I have 4 of the Vifa TC9FD drivers, and am giving serious thought to your Mini-Karlsonator. Can you post a picture of the back of one?
Thanks,
Mike
for a regular Karlson, you could go with about a "K10" size - leave out the front shelf or keep it very short for experimentation
Beta 8cx vs Beta 10cx in a “K10” size coupler – no front shelf – upper port board cant 12 degrees, baffle cant, 30 degrees - - rear volume 1 cubic foot - front volume ½ cubic foot – I think the rear chamber was a bit large for best power handing but the rear lowpass gap wasn't as tight as might be – it reminded me of my Klipschorn or some K15 setups in subjective balance.
we want the passband reasonably flat without too much cavity peaking - I think to a mild degree that cavity shape can affect that outcome. Vent placement can matter - tune by ear or instrumentation.
a K-tube can be used on top - or inside the coupler - Imageshack has lost my image with 8cs but here's the cabinet pix with BOFU
the stock Karlsonator 8 might be a good match - I don't have the wits to use akabak
the vent panel in the cabinet below was setup with two sliding panels which allowed either one central placed vent or when butted, two vents against the sidewalls - - in general for a cabinet this size, the center vent which faced the open part of the aperture, had more "fill" in the 300 to 400Hz region.
distributed horizontal slot vents as with the "Dutch K12" might be a good option
Beta 8cx vs Beta 10cx in a “K10” size coupler – no front shelf – upper port board cant 12 degrees, baffle cant, 30 degrees - - rear volume 1 cubic foot - front volume ½ cubic foot – I think the rear chamber was a bit large for best power handing but the rear lowpass gap wasn't as tight as might be – it reminded me of my Klipschorn or some K15 setups in subjective balance.
we want the passband reasonably flat without too much cavity peaking - I think to a mild degree that cavity shape can affect that outcome. Vent placement can matter - tune by ear or instrumentation.
a K-tube can be used on top - or inside the coupler - Imageshack has lost my image with 8cs but here's the cabinet pix with BOFU
the stock Karlsonator 8 might be a good match - I don't have the wits to use akabak
the vent panel in the cabinet below was setup with two sliding panels which allowed either one central placed vent or when butted, two vents against the sidewalls - - in general for a cabinet this size, the center vent which faced the open part of the aperture, had more "fill" in the 300 to 400Hz region.
distributed horizontal slot vents as with the "Dutch K12" might be a good option
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.
Last edited:
Mike,
Sorry I missed this post - you had some questions here.
12/15x literally means 12 divide by 15 scale. The 15 is there because the cabinets were designed for nominal 15 in drivers. A close first cut of the size of a Karlsonator can be achieved by dividing the driver size by nominal size - but usually it is a little bigger to improve bass extension.
The Karlsonator is not open back. I you look at the plans it resembles a tapped horn with the driver faces pointing out and a Karlson slot on the mouth. The back of the drivers couple into the TL or horn.
Greg B's Karlsonator Six might be about right - he can link the plan - its the middle cabinet loaded with a FE166EN in the picture below
An externally hosted image should be here but it was not working when we last tested it.
