Ultra-compact dipoles

[gainphile]

So I managed to have sometimes to drive the speakers to Xmax today . Great to see that the 10" woofers are the limiting factor. The 5" held its composure with 280hz xo.

An 'issue' with slot loaded woofer is storage energy. Not unlike a bandpass sub. Here we see the storage energy issue at 100hz. Whether it's audible or not remains to be heard. I'm sure it colours the sound.

[Jonasz]

Sounds good there's enough oomph down low.

One final thought, what about trying a lower crossover between woofer and mid just to get away from the woofers resonanses a little? I'm sure 175-200hz would be doable?

Now, where's my minidsp...? :P

[gainphile]

At some stage I might try lowering the crossover frequency. But really the storage energy is anywhere above 100hz. I guess that's why SL lowered the Orion 4 frequency to 92hz?

Here is a sonogram of the same data.

[JonMarsh]

At some stage I might try lowering the crossover frequency. But really the storage energy is anywhere above 100hz. I guess that's why SL lowered the Orion 4 frequency to 92hz?

Here is a sonogram of the same data.

I think that has a lot to do with it. Any raised Q due to the baffle/slot loading does change the audible characteristics, and the sense of articulation, in my experience. I'm still concerned about that possibility with the Arvo Pärt Mk IV, but we'll see. Thanks for sharing this much information on your measurements.

[charlesd]

Diagonal Baffle?

I was wondering if it would improve the stored energy issue if the baffle for the woofers was placed at a diagonal so the opening would be larger. I really like the concept of a dipole with a narrow tower. The design is looking great!

[gainphile]

It would need to be built and measured to be sure. But that's a great idea.

Here MFK executed nicely that idea,but with high SPL 12":

http://mfk-projects.com/dipole_sub-woofer_titanic_mkii.htm

[owdi]

Dipole *adds* to the loss you already get from the 2pi>4pi transition, so starting with full BSC means he's only adding dipole correction to the mix.

Oh I know there is less active correction that has to happen with BSC in the passive XO, but I'm puzzled why do any correction in the passive XO if you have the active filter.

[cjd]

Well technically speaking there is no bafflestep loss in going from 2pi to 4pi with an open baffle loudspeaker. Bafflestep and OB losses do start at the same point, as governed by the baffles width, but with an OB you don't have to compensate for open baffle losses and then on top of that compensate for baffle step losses, you only need to compensate for the OB losses, which are 6dB/octave, rather then the 6dB bafflestep transition which occurs over the range of a couple of octaves.

Much better said. Thanks.

[Juhazi]

Looking very good! And thank you cjd and others fo very informative and educative discussion.

I've been thinking of a quite similar solution but with a plate-amp driven bass section (Hypex http://www.hypex.nl/index.php?option...=80&Itemid=88). I just don't have measuring gear (and know-how) to really make an effort. Hypex also makes programmable amp-boards that are popular among European DIYers...

[sfdoddsy]

Oh I know there is less active correction that has to happen with BSC in the passive XO, but I'm puzzled why do any correction in the passive XO if you have the active filter.

I assume it is because the passive xover corrects the drivers and the active corrects the speaker.

With my 3 way active dipoles I EQ each driver, then do the crossover, then EQ the overall response. A hybrid system like Andy's would use the passive crossover to get the treble and mid OK before adding the active dipole EQ.

Saves having an extra amp channel. If I could build crossovers I would have tried it with mine.

[gainphile]

If designing hybrid system from scratch, then the passive XO correction is not necessary.

But in this case I am taking a passive xo product from box speakers, which already has BSC. This is needed as part of the original box design.

As Steve said, passive xo design is hard, so all the work had been done by Zaph for Woofer-Tweeter crossover and we leverage this.

[matssvensson]

So I managed to have sometimes to drive the speakers to Xmax today . Great to see that the 10" woofers are the limiting factor. The 5" held its composure with 280hz xo.

An 'issue' with slot loaded woofer is storage energy. Not unlike a bandpass sub. Here we see the storage energy issue at 100hz. Whether it's audible or not remains to be heard. I'm sure it colours the sound.

Would you please share how you performed the measurement for this burst decay? As far as I try interpret the burst decay graph, then a curved diagonal decay like the one visible in this graph would indicate a reflection rather than resonance/strored energy (which should show up more like a straight line parallell with the periods axes in the diagram). But I'm no expert. Sharing the measurement set up might help further explore the reasons behind the plot decay.

Also, My take on Mr Linkwitz adjustment of the woofer to mid transfer function in Orion 3.3/3.4 has more to do with the delicate balance of equalizing room gain and transition from 2 pi to 4 pi radiation for the units rather than stored energy issues.

Then, thanks for sharing your explorations in designing controlled directivity loudspeakers. I follow your blog with great interest since I'm into these type of speakers and dipoles myself. :T

[gainphile]

It is a nearfield measurement, as per Post At The Movies and on TV

For Orions, you could be right. I definitely see no problems even at 120hz with my larger W-Frame dipoles.

I happily run them at that frequency.

[sfdoddsy]

What impresses me, Andy, is how good all of your various iterations look. I've had mine for 10 years and they still look like test mules, whereas every one of your 16 or so versions looks great.

Wanna build some cabinets for me?

[gainphile]

Sure..... if I have a CNC that is :lol:

At the end of that last build I was able to cut within 0.5mm precision using circular saw and guide. Gives you some ideas much I have breath MDF dust! 8O. Also a Makita is man's best friend :T

I feel this S20z is lacking in aesthetics. Will try to think of something.

[matssvensson]

Thanks for clarifying the measurement set up and nice you had a picture of it as well. I could have double checked that myself. ops:

I've made similar measurements on my W-baffle speakers with a pair of Acoustic Elegance Dipole 10 woofers and can see the same type of expected quarter wave resonance at around 300 Hz. Similar to the one you display in post #28. I've manage to reduce the resonance by lightly stuffing the cavities with a pair of wool stockings filled with some lambs wool. It improves the frequency response as well as the burst decay, and the sound improves. Just make sure the stuffing doesn't touch the cones.

Another idea might be to divide your cavities in half by a horizontal piece of MDF between the speaker units. Making the cavity each unit sees smaller should push the cavity resonance higher up in frequency.

I used to run my W-baffle woofers to 300 Hz in a passive system and thought it worked well. Now I run them actively to around 130 Hz and am completely satisfied. :B

[davewantsmoore]

I feel this S20z is lacking in aesthetics. Will try to think of something.

Hah. Yeah. I've been on the verge of building my own "compact dipoles" ... in a sideways woofer format like you ... can't really come up with anything that will get WAF approval as they'll be for the living room (somehow boxy 'normal' speakers are ok... but anything that looks different is thumbs down).... also mounting waveguide in an aesthetically pleasing way is troubling me.

I do love the wood on yours + black + the ZA drivers ... great.

Finally some decent days (I'm in Hobart) are popping up to get outside for midrange measurements too... I should get prepared :-)

[charlesd]

I really like your original design drawing of the S20Z. The rails on the front baffle that went all the way from the top to the floor enhanced the thinness of the design.

Here is a Compact W-Frame Dipole design that I am thinking of building that incorporates some elements of the narrow baffle S20Z and the W-Frame woofer of the S19.




The narrow top baffle helps make this speaker appear smaller although it's not as small as the S20Z. Using a w-frame should eliminate stored energy issues and also reduce vibration. In order to keep the w-frame as narrow as possible I plan on routing slots to make the sides only 0.6cm (1/4 in) thick where the woofer is the widest. Since it's an open baffle strength shouldn't be an issue.

[davewantsmoore]

Genius.

I've been thinking on how to build w-frames but keep them as narrow as possible. I have difficulties with the "Minister for Aesthetics" at my house. (SWMBO)

Luckily I will soon have somewhere of my own where I can keep all the 'ugly' speakers :-)

[gainphile]

Great design there.

I've built 24cm dipoles and they definitely looks very narrow.

Another idea that I had was using push-pull woofer in slanted position. Like W-Frame but the drivers are angled to reduce space even more. Let's call it a "Z-Frame" dipole sub.



All the angles would make construction more challenging though. Acoustically they should be similar.

[JonMarsh]

Z-frame.... that almost sounds French!

And while I've done my share of angles in the past, (and coming in the near future), this would be an "interesting" build to accomplish- especially for the heavier 10" woofers on the market. Would it make the force cancellation a bit trickier? I think it might introduce a rocking couple...

[gainphile]

He..he .. having no access to table saw it will be nearly impossible for me. Unless someone can show me how to do it with circular saw??

In any case we've had few good runs on great weather so I hope to conclude and post this project soon!



A bit tricky as it seems everyone and his dog went out and fly their planes on such weather. And I meant stop taking measurement every minutes or so 8O. I live about 20km from a small airport and Bonbeach seems to be their favourite run. Must be nice from up there but I also get to understand what frequencies these planes drone at.

[5th element]

I have also taken to repurposing ironing boards for such endeavors, I highly approve! Nice spot for measuring in though, the planes aside.

[gainphile]

Yes, they are locally available woofers here, and I already have 4 of them.

The Dayton Audio SD270A-88 10" DVC Subwoofer would make a great low-cost choice with Fs of 26hz and Qts 0.43. But shipping cost is prohibitive to Australia!

Sven from DIYAudio purchased these SD270A woofers and used them for the similar sub , although not as floorstander but separate one. He sent me measurements which shows it's a perfect woofer for the job (Fs, cost, etc.)

I took his .frd and it requires minimal EQ to hit F3=20hz.




Compare this with my Jaycars:

[gainphile]

I have also taken to repurposing ironing boards for such endeavors, I highly approve! Nice spot for measuring in though, the planes aside.

We can only accept the highest possible quality tools for the job ! :lol:

[charlesd]

The Z-Frame looks like a very cool configuration to make a narrow baffle. I also really like the the "Z" look goes with the "Zaph Audio" theme. It seems like the cavities are wider so you would have less energy storage issues. You would also get some force cancelation with the mostly opposing drivers. It would be a little challenging to build with the extra angles and small confines.

Thanks for posting the measurements of the Dayton SD270A from Sven. They do look good.

Is there a reason that most W-frames have drivers that both point up instead of symmetrical like in my drawing? Is this so there is the same amount of cone droop from the weight of the cone? Or, it it just ease of mounting? I kind of like the aesthetics of a symmetrical design (one pointed up and one pointed down) and not seeing the basket (unless it a nice cast one).

[cjd]

Is the a reason that many W-frame configurations? I have seen have drivers that both point up instead of symmetrical like in my drawing? Is it so there is the same amount of cone droop from the weight of the cone? I kind of like the aesthetics of a symmetrical design and not seeing the basket (unless it a nice cast one).

Basically, it balances out any audible effects of the frame/magnet on a dipole.

You do have to consider sag any time you mount drivers like this. I'm on the fence whether I want to try, or instead have a much wider base than I was previously considering.

[5th element]

We can only accept the highest possible quality tools for the job ! :lol:

^>^ I used an ironing board as a keyboard stand too for a while too and that worked quite well.

[5th element]

Basically, it balances out any audible effects of the frame/magnet on a dipole.

You do have to consider sag any time you mount drivers like this. I'm on the fence whether I want to try, or instead have a much wider base than I was previously considering.

It also cancels out even order distortion.

Second order harmonic distortion occurs when the forward motion of the cone differs from the backwards motion. In other words the peak going side of a sinewave is different from the trough going side of the sine wave. If you connect up two drivers push pull, then the peak going side of the sinewave is created by the forwards going motion of one cone and the backwards going motion of another. Likewise the trough side of the sinewave is also created by the forward going motion of one cone and the backwards going motion of the other. Like this you've got an identical situation for both the peak and the trough and this therefore reduces distortion.

This is also why balanced circuits have lower levels of even order distortion as the principle is the same, you run two together, but one has it's phase inverted, so a + and - are used for each half period.

[wmcarpenter]

Hello all. First post. I really appreciate the sharing of your knowledge and experience with loudspeaker design and construction. Because of forums such as this, I'm going to take this hobby to the next level. I've been tinkering with the idea of a dipole with some TB raw drivers left over from another project but I really like this nearly turnkey dipole kit from Zaph/Gainphile. Thank you for these contributions.

Here are my design goals: music primarily, low distortion, minimize room nodes, constant directivity and amplitude across frequency range (to the extent possible), excellent transient response (however one defines that), time aligned and phase coherent (possibly redundant), as aesthetically pleasing as possible.

I'm intrigued with the idea of building this Zaph/Gainphile dipole however as suggested somewhere, I'm thinking of incorporating the MTM version of the kit from Madisound. I might taper the frame inward around the tweeter and back out for the lower midrange. I'm also probably going to construct the subwoofer assembly in the W frame from the previously mentioned post(s).

Questions:

I'm not that astute when predicting impact on response from making changes in dipole designs. I've been trapped in the "box" until just recently, so I apologize if these questions are naive.

How will the MTM configuration impact the overall response and how should I compensate when adjusting the active crossover values?

Any pros or cons to taper the tweeter baffle width in narrower versus leaving it the same width for the entire MTM length?

With the W frame subwoofer enclosure, is "sag" something that inherently impacts response or just overall life of the driver? How does one compensate for this sag?

I'll probably use the suggested Dayton 10" subs in push/pull with phase inverted. What is the best wiring configuration for lowest distortion and best transient response? This is primarily for music low-end--for window shattering movie explosions I'm building a pair of closed box subs.

I may consider a 10" sub with a lower overall coil inductance to improve transients along with the LT circuit through my miniDSP. However I don't want to compromise the budget too much, otherwise I'll start an upgrading spree on everything and find myself starting over from scratch.

What are the opinions on voice coil inductance and transient response?

Any thoughts or tips would be welcomed on my first dipole project related to this existing build thread!!

[gainphile]

How will the MTM configuration impact the overall response and how should I compensate when adjusting the active crossover values?

Horizontally it should be similar. SPL wise the MTM would be more sensitive so you would need to adjust the levels accordingly. Or alternatively, push the xo lower as you'd have more SPL capability.


Any pros or cons to taper the tweeter baffle width in narrower versus leaving it the same width for the entire MTM length?

You really need to take it outside for polar measurements to see the effect. For sure the uniformity of response would be pushed higher for the woofers (good). Have a read to this too for dipoles with changing width:
linkwitzlab.com/frontiers_7.htm#D2


With the W frame subwoofer enclosure, is "sag" something that inherently impacts response or just overall life of the driver? How does one compensate for this sag?

There is a specific requirements for the woofers suitable for down firing due to gravity and mass. Basically Sag = Cms x mms x 9.8 and it should be < 5% of xmax.
gainphile.blogspot.com/2012/09/upside-down-down-firing-woofer.html

[wmcarpenter]

Excellent and articulate. Thank you--I'll get to work and keep everyone informed.

[charlesd]

Thanks for the explanation of why the drivers are mounted the same direction in W-Frames.

Here is another Z-Frame type configuration that I mocked up:



PROS
- Slim profile
- May have less stored energy issues
- No driver sag
- Doesn't seem too difficult to build/mount drivers
- A little more traditional look (higher WAF)

CONS
- Only has partial driver force cancelation
- Harder to construct than the straight slot design
- A little taller box than a W-Frame

[gainphile]

That's an excellent design!

H-frame should be OK considering the SPL which we are talking about. The Orions, for example had H-frame for a decade with no one complaning

[oneplustwo]

My only question would be if you can get the drivers in and out or if you would have to place them during assembly? Or you can make the diagonal pieces removable but that probably doesn't make you feel good about strength and being vibration free.

[charlesd]

Thanks gainphile! I really like your original slot design for it's simplicity and your Z-frame for large separation (higher SPL) in a slim baffle. I'm really looking forward to seeing your measurement and listening impressions. I'm not sure if the angled design is worth the extra complexity but I'm just playing with different designs to see what I would like to build. I really like your concept of full dipole with a reasonable size and cost but still have great performance.

My only question would be if you can get the drivers in and out or if you would have to place them during assembly? Or you can make the diagonal pieces removable but that probably doesn't make you feel good about strength and being vibration free.

I did the mockups in the PhotoShop (close to scale) and tried to simulate inserting the driver by rotating and moving it. When I put the basket in the cut-out first and then slid/rotated the driver into it's final place it seemed to just barely fit. I think I will need to use machine screws with hex/allen heads to fasten the drivers in the tight corners. If I do go with this deign I will definitely cut out the baffle first and do a test fit to make sure I can mount the driver. I would prefer to have all the pieces securely glued if possible.

[charlesd]

Here is a rendering I did of gainphile's Z-frame design for those that are interested.




The woofer placement would be tight but it seems like they would fit. There are a lot of things to like about this design (e.g. slim, low stored energy, higher SPL,...)

[gainphile]

Beauty !! :P

[jkrutke]

I haven't stopped in here in a while, but I have to say there's some cool ideas brewing.

Gainphile: nice work on those dipoles, glad you put them to good use.

[wmcarpenter]

I've decided on the MTM version of the Zaph kit for this slim dipole build, with a W frame design for subs but with drivers both firing upwards--similar to Gainphile's S19 concept. Still considering 10" sub options to get the most per $ and energy/time invested.

May need to widen bottom half a few inches for W frame but willing to compromise on aesthetics. Some of these Z frame ideas definitely have potential.

Making a complete inventory list so I can purchase, begin assembly, and testing next weekend.

[gainphile]

I haven't stopped in here in a while, but I have to say there's some cool ideas brewing.

Gainphile: nice work on those dipoles, glad you put them to good use.

Thanks... They do keep me entertained :T

Finally I have documented and posted the build.

http://gainphile.blogspot.com.au/2012/06/s20-z-compact-open-baffle-using-zaph.html

Now I will ride my motorbike !

[avddreamr]

@gainphile
Wow, what a great design, incredibly high on value I would imagine.
One question, on placement, is that 80 cm /50 cm from the front of the mid/tweet baffle or from rear of the bass bin, or minimum from either.

[wmcarpenter]

Some pics

So, I deviated from the original some. Upper portion is same equalized with miniDSP for dipole roll off. Lower bass portion I tried the dipole bass array in W frame but I wasn't particularly satisfied with the SPL and extension, so, I tried an experiment for kicks. Built 4 cabinets and put two 10" Dayton DVC subs on front and rear baffle of each in push-pull configuration. Wired in parallel/series for each cabinet (1x1) so each sub is wired in push-pull parallel (4ohm), and the total cabinet for two subs is wired series at 8 ohms. Each stack is wired in parallel to get 4ohm total for each side, with a Crown XLS amp running in stereo mode and equalized through MiniDSP with Linkwitz transform down to 20hz with .5 QTC, and 2nd order LR low pass filter with 180HZ roll off, its a decent transition with plenty of tight, distortion-reduced bass. Used software to calculate LT filter after having two 10" DVC subs in a very small box and keeping xmax just within limits of power handling to lowest frequencies. Wiring out of phase in push-pull creates a dipolar pattern but overall, it feels like more constant punchy amplitude down to low frequencies. Very controlled bass here, nothing flapping around with heavily reinforced cabinets. Note: these have not been measured but my ears are pleased so far. I could be delusional with this bass array but it sounds great and gave me ample practice on my new table saw. Measurements and further equalization to follow.

What I might do? Put these MTM's in a box for rear surrounds, and try the Zaph MMTMM kit in an open baffle design with existing bass array for better vertical controlled directivity and to minimize room nodes across the entire range. That would make a big speaker with big sound.

[wmcarpenter]

Well, I've learned a lot since first starting this build. A few failed experiments, a few important realizations when trying to connect theory with real-world (measurable) results, but most of all, I've learned that there is so much to learn, and this is a hobby I'll probably keep doing for a lifetime. Thank you gainphile (and everyone else) for sharing your ideas and designs with a most grateful community of "sound" architects; both novice and experienced.

"With the drawing of this Love and the voice of this Calling,
We shall not cease from exploration,
And the end of all our exploring,
Will be to arrive where we started
And know the place for the first time." T.S. Eliot

[JonMarsh]

So, I deviated from the original some. Upper portion is same equalized with miniDSP for dipole roll off. Lower bass portion I tried the dipole bass array in W frame but I wasn't particularly satisfied with the SPL and extension, so, I tried an experiment for kicks. Built 4 cabinets and put two 10" Dayton DVC subs on front and rear baffle of each in push-pull configuration. Wired in parallel/series for each cabinet (1x1) so each sub is wired in push-pull parallel (4ohm), and the total cabinet for two subs is wired series at 8 ohms. Each stack is wired in parallel to get 4ohm total for each side, with a Crown XLS amp running in stereo mode and equalized through MiniDSP with Linkwitz transform down to 20hz with .5 QTC, and 2nd order LR low pass filter with 180HZ roll off, its a decent transition with plenty of tight, distortion-reduced bass. Used software to calculate LT filter after having two 10" DVC subs in a very small box and keeping xmax just within limits of power handling to lowest frequencies. Wiring out of phase in push-pull creates a dipolar pattern but overall, it feels like more constant punchy amplitude down to low frequencies. Very controlled bass here, nothing flapping around with heavily reinforced cabinets. Note: these have not been measured but my ears are pleased so far. I could be delusional with this bass array but it sounds great and gave me ample practice on my new table saw. Measurements and further equalization to follow.

What I might do? Put these MTM's in a box for rear surrounds, and try the Zaph MMTMM kit in an open baffle design with existing bass array for better vertical controlled directivity and to minimize room nodes across the entire range. That would make a big speaker with big sound.

You're clearly learning a lot of new stuff and having fun- the journey is just as important as the destination!

I like this for the money quote:

I could be delusional with this bass array but it sounds great and gave me ample practice on my new table saw.

Try to get some measurements you can share with the rest of us! If it sounds great, that's likely a good hint-