[JLM]

This design looks similar to a passive version of Linkwitz's Orion speakers with maybe a bit less deep bass.  I'd like to see a bit more efficiency as my experience is that higher efficiency also means greater dynamic response, but maybe with the new breed of digital amps this ends up being less of an issue).

We put horizontal (with about 3/4 inch exposed per piece) bamboo flooring (natural) in our new house (dining, kitchen, foyer, and mud room).  Being a grass its more tolerant of water than wood.  We heard that its as hard as maple (our second choice).  We started with cherry cabinetry (to match my existing furniture) and wanted a nice contrast.  Maple can vary from very blond to almost reddish, but natural bamboo always looks the same (and very good with natural cherry).  And we caught Home Depot selling it for $2.75/sq ft.  I can testify to its weight as we ended up picking it up for the builder.

[PaulHilgeman]

Yes, Bamboo is truly an incredible material.

As to the orion... I have built several full dipole systems, most recently one almost identical, but with dual 12" Peerless XLS woofers instead of dual 10's.  The frame was also a bit deeper to further enhance the ability to play low with minimal strain.  Believe it or not, the single 10 inch woofer in a sealed box, has more output capability than the dual 12" dipole below 57Hz, and more output capability than the orion below 100Hz.  For example, a sealed box, with a typical 10" woofer moving at 10mm P-P, can generate 97dB at 40Hz.  Dual 10" Woofers in a 16" Deep Dipole can generate 92dB.  Take it down to 20Hz, and we are talking about 85dB for the monopole (closed box) vs. 74dB for the dipole.  The monopole system will sound far more 'at ease' than the dipole, since even an 80dB peak at 20Hz would push it past it's 10mm one-way excursion, the closed box would be doing just fine.  This also assumes that the 10" woofer I use has the same motor linearities as the 10" Peerless XLS woofer used in the Orion, but they dont.  

One of the huge problems with dipoles is that the woofers systems resonate at a frequency that is very low, you have to cross over well below this resonance and use a notch filter, otherwise is is clearly audible.  This is one of the reasons why the Linkwitz Orion crosses from midrange to woofer at 110 Hz or so.  110 Might seem like a pretty high frequency, but there still is quite a bit of strain placed on the mirange crossing at that frequency at higher playback levels.  

I found that not only was the overall design much simpler, but sounded better as well with the woofer crossing to the midrange at a higher frequency.  Also, while the theoretical benefits of a dipole system seem to  be in the bass range, I feel that with proper placement and proper speaker design, a closed box can offer everything a dipole bass system can, but with much higher dynamic range, a simpler system design, less moving surface requred, and generally smaller cabinets.

In the midrange, there are the beneifts of reduced reflections, and the lack of cabinet resonances and resonances coming back through the cone.  I feel that this is clearly audible and really would never go back to closed box speakers.  This isnt to say that there arent some very good closed box speakers, they are all just fighting the same battle.

-Paul Hilgeman

[PaulHilgeman]

This shouldnt sound like I was bashing the Orion, it is an excelent speaker.

But for quite a bit less money, and what I consider to be better performance, I belive that Nomad Audio offers a better product.  

-Paul Hilgeman

[Watson]

One thing worth pointing out is that Linkwitz went monopole with his Thor subwoofer, intended to be used with the Orion.  

I think you're probably right, Paul, that as long as the woofer is crossed below baffle step, there are few radiation advantages to using a dipole as a bass driver, and several reasons to go with a monopole.

Any chance of offering your speaker in kit form?

Also, for curiosity's sake, do you have any good pictures showing the color difference between caramelized and natural bamboo?

[Dr. Krull]

One of the huge problems with dipoles is that the woofers systems resonate at a frequency that is very low, you have to cross over well below this resonance and use a notch filter, otherwise is is clearly audible. This is one of the reasons why the Linkwitz Orion crosses from midrange to woofer at 110 Hz or so. 110 Might seem like a pretty high frequency, but there still is quite a bit of strain placed on the mirange crossing at that frequency at higher playback levels.

I found that not only was the overall design much simpler, but sounded better as well with the woofer crossing to the midrange at a higher frequency.

Actually, the "huge" problem is due to the enclosure, not the dipole arrangement itself.  The Linkwitz Orion uses a 4th order LR electrical crossover at 120Hz and the excursion of the driver isn't all that large.  In fact, I'm told that the woofers run out of steam well before the midrange shows any sign of stress.

I'm curious if you are using an XO point of 250Hz in this speaker.  What order is the crossover?

"Also, keep all of your XO's out of that delicate 300-3000 Reigon if you can.  Especially the 300Hz reigon."

http://www.diyaudio.com/forums/showthread/t-22078.html

-Krull

[PaulHilgeman]

Actually, the "huge" problem is due to the enclosure, not the dipole arrangement itself.

Correct, but in order to separate the front wave from the rear, there must be some separation baffle, either wide or deep.  Through my own testing, I have found significant resonances in these open chambers of air resonate.  Both the Orion and the Phoenix Dipole woofer use a variety of notch filters to deal with these resonances.

And with proper filtering and crossing about an octave away from these frequencies it is not a huge problem at all.  In my design, I took the route of a sealed box woofer over a dipole woofer system for several reasons:

1.  Allows for a higher frequency crossover from midrange to woofer.  This places less strain on the midrange.

2.  Much greater output capability (dynamic range) in frequencies below 100Hz. than a similar dipole arangement with two woofers per side.

Simpler overall system design, with better overall sound quality, and greater dynamic range really leaves no other option.

The Linkwitz Orion uses a 4th order LR electrical crossover at 120Hz and the excursion of the driver isn't all that large.

Correct, but the midrange dipole losses are also compensated for, so remove the 4th order High-Pass, and the midrange will play flat down to around 30Hz, where the natural roll-off of the driver occurs and the Low-Shelving filters run out.  So, the 4th order electrical filter in that case is a 4th order acoustical filter since driver impedence dosnt matter with active filtering.

In fact, I'm told that the woofers run out of steam well before the midrange shows any sign of stress.

This all depends on the music you are listneing to really.

The midrange is most limited in the range right above the crossover frequency.  At 140Hz, the midrange would be limited to 108dB at 5mm of excursion.  The woofer system is at or below that point from 82dB on down.  So, again, it really depends on spectrum.  Technicaly, the woofers are more limited, but in my experience, placing the Seas W22 on an open baffle of about 11-12 inches wide, crossing 4th order at 120Hz, there is significant movement required by the midrange.  At these frequencies, I would prefer slightly more dynamic range.

I'm curious if you are using an XO point of 250Hz in this speaker. What order is the crossover?

Acousticaly, 3rd order.  Even power response... goes with the whole theme.

As to my quote, That was 2 years and 7 months ago, and I had just started playing with dipole speakers about 6 months before.  I had recomended that as a 'rule of thumb'.  At this point, and back then, I feel/felt that if properly done, any crossover point can be done to sound good.  Actually I prefer the sound of a tweeter crossed pretty low, cutting that 300-3000 range rigth down the middle  .  If you have heard the Orions, or any of SPTech's speaekrs, I am sure you would agree.  There are very few woofers that dont exhibit very bad properties above 1500Hz or so, such as stored energy and harmonicaly exciting higher frequency resonances.  Tweeters do a much better job in this range.

For advanced speaker designers that have designed 20 different (good) speakers using passive, active and DSP crossovers, I dont think that it holds water.  For a novice that had posted on the DIY Audio boards, I still think it is a good general guideline, the project that he was about to embark on was nearly impossible, I have no idea how it turned out.

[Davey]

Paul,

I'm sorry.....I haven't been keeping up with the forums.  The coaxial driver is not an off-the-shelf unit is it?  This is a W22 driver modified by you to incorporate the tweeter?

A quick correction on the Orion woofer system.  It does not use a variety of notch filters.....it actually doesn't use any.  The Phoenix system uses a single notch filter to attenuate the cavity resonance formed by the "W"-baffle, but the Orion system uses an "H"-baffle and no notch filter is required with the 120Hz crossover point.

Your effort looks intriguing and well thought out.

Good luck.

Davey.

[PaulHilgeman]

Interesting, thanks for the correction.  What is the "200 NF" on the ASP page of the Orion's weibsite?  I assume it is just part of the dipole correction.

Let me clarify what I meant by "variety", is that simply the Orion used one notch filter, and that the Phoenix used another, that different filters were required for the different driver layouts / cabinet construction / depth.

Also, I think I may have mispoke a bit before.  I have done so much research with H-Frames, W-Frames and U-Frames that I think I verbaly exagerated the 'sound' of the resonances found in dipole woofer systems.

I should have said that typicaly U-Frame woofer systems have pretty strong resonances that must be filtered.  Typical H-Frame systems (like used in the orion) only have the fairly simple and fairly easy to control dipole roll-off and its associated 3-6dB peak just above roll-off.

No, it is not an off-the shelf unit.  I actually machine down copper cylinders to re-create the standoff for the copper phase plug in the W22, this retains as much copper in the motor as possible.  The change in impedence is very very small.

Thanks for the compliments,
Have a great weekend Davey.
-Paul[/b]

[opnly bafld]

Paul,
By U- frame, do you mean basically an open back box?
If so, does a lower cutoff freq. or the depth of the box make a difference?
Thanks,
Lin

[Davey]

Paul,

The "200 NF" is an unused component location in the Orion configuration.  SL included it so folks who might use the ASP circuit boards for other speaker projects can have the option of a notch filter in the woofer portion if they so choose.  It can also come in handy for a low-frequency room peak notch filter.

The woofer drivers in the Orion are corrected for dipole roll-off and Q increase (both 6db/octave shelving filters) and then a generic LR24 crossover is applied.  There are no notch filters of any kind.  The "peak" you refer to is significantly higher (and away from the roll-off) in an "H" configuration and does not need any attention if the crossover frequency is fairly low and the crossover slope fairly steep....as they both are in the Orion.

The "U"-frame configuration is quite different, and not one used by Linkwitz in any of his designs that I know of.

Lin,

Some good reading here:

http://www.linkwitzlab.com/frontiers.htm#L

http://www.musicanddesign.com/u_frame.html

Cheers,

Davey.

[opnly bafld]

Davey,
Thanks for the links, they answered my   s.
Lin

[johnk...]

Davey is correct in that the Orion does not use as notch filter in the woofer circuit, nor is/should a notch filter (be) used with a U-frame, such as the NaO woofer. The 1/4 wave resonance of the U-frame must be acoustically damped as it affects the radiation form the front of the driver and the rear of the U-frame differently.

I do have a question for Paul though. I am a little concerned about the variation of the tweeter response and possible distortion (IM) with the W22 cone motion. Have you looked at that? Or perhaps just looking at the tweeter response with the W22 cone positioned at +Xmax,-Xmax and centered?

I'd be interested in seeing what that looks like.

It’s a nice design and I agree that you won't be sacrificing much with regard to bass response by goinf with a monopole woofer. It is a necessary trade off for a single amp design.

[PaulHilgeman]

Hey John,

Yes, I have looked at it.  The midrange really dosnt see too much cone motion, maybe 1mm peak to peak at pretty average listening levels.  The frequency response variation below 10K is non-existent at cone in vs. cone out levels.  I have also looked at IM levels as well, and will post some things to the website within the next week of looking at 300/3000 and 300/10000 tones.  

Yeah, I agree with the bass, I have two wildly different rooms to listen in and I still think it is a stretch to justify dipole bass systems.  I was a long-time proponent of it, but after careful research and listening, I found that it seems great on paper, but after measuring and listneing and measuring and listneing and building way to many active EQ and crossover circuits that it was all for pretty much nothing.  

In order to achieve similar dynamic range below 80Hz with this speaker in a dipole, it really wouldnt be able to be shipped around the country.  Cardioid is another story, but still requires the low-end eq and a lower frequency crossover to the midrange.

-Paul Hilgeman

[Dan Driscoll]

Hi Paul,

That is a very interesting design, I'd love to hear them some time.

I do have a question about the mid-range/tweeter mounting. Is there a reason (beyond esthetics) for having side and top 'rails', rather than mounting them on a flat panel? I assume it is not structural, bamboo plywood is very strong and rigid, it should be easily capable of handling the weight of the coax mid-tweeter arrangement without bowing. It seems to me they would cause some reflections within the volume immediately behind the baffle. Since they block the direct 90* angle, do they interfer with the 90* off-axis cancellation? Perhaps the rails help dampen vibrations in the panel?

FTR, I'm not a speaker designer or an accoustical engineer, but I do know a little bit about accoustics, perhaps just enough to make myself look silly in this instance.  

[johnk...]

Hi Paul,

Thanks for responding. I think my concerns may have been over stated. I was thinking of the more extreme midrange excursions that are present in designs like the NaO II and Orion with a 120-125 Hz x-o to woofer (a necessary evil with a dipole or U-frame woofer). I probably should have been thinking more about the NaO III which is similar in concept to your system except for the important coaxial nature of the mid and tweeter. The crossover in my III design is at 250 Hz and that means the max excursion goes down by a factor of 8 compared to the II for the same SPL. I presume your woofer is crossed to the mids in the same frequency range.

 As you know, I also spent a lot of time looking into low frequency reproduction myself with monopole, cardioid and dipole woofer systems where the on axis anechoic was carefully matched. And I also looked at the theory too. I just didn't hear or see any really advantage of one form over the other. Yes, you can make all the claims about excitation of room modes and power response you like but “globally better” just doesn't follow from "different". I think is also important to recognize that systems aren’t always set up in rectangular shaped rooms. There are usually openings to other rooms and what ever and with today’s home designs with open floor plans..... Additionally, as you are probably aware, when looking at power response I recently realized that a monopole woofer in close proximity to the floor is actually a better power match to a free standing dipole mid.

The biggest issue is really that of selling the dipole mid. That is where the major difference in perceived sound is derived and it's either what someone wants or doesn't want. A correctly designed conventional front radiator speaker can sound very good too with the right setup.

[PaulHilgeman]

Yes, same frequency range

I agree with everything with regard to the different bass systems.

Yes, a conventional front firing, non-coincident speaker can sound very good, but I have yet to hear one that sounds this good.

Have a good one John,
Paul Hilgeman

[Occam]

Paul,

I like your design very much. While coincident origin would be a prerequiste of linear phase behavior in the mid/tweet region and above, its obviously not all that is required. The crossover implementation is also an additional factor. Could you provide a miked waveform picture of a squarewave with a periodicity of an octave above (or thereabouts) the mid's highpass? (the curved shoulders would be expected). On axis would do for a start, but off axis, in addition would be great.

TIA,
Paul

[Davey]

Paul,

Maybe I missed it, but I don't think Paul claimed linear phase behavior of this speaker system.  Third-order acoustic slopes, phase coherent, linear power response, etc.  However, those don't imply linear phase operation and a reproduced square wave would look mighty distorted me thinks....as expected.

Cheers,

Davey.

[PaulHilgeman]

We will see what it looks like.  The actually 'Look' of it might be pretty bad because of the +/- 2dB peaks in the 10Khz - 20Khz reigon, this tends to make square waves look really bad.

Keep in mind that square waves tend to look really really bad on most speakers.

What did you have in mind of comparing it to?  I would like to keep the measurement as similar as possible.  If they measured at 6Khz, I would like to measure at 6Khz. etc.

Thanks,
Paul

[sts9fan]

Whats the point of a square wave? The speaker will never see one with music so what does it tell you?