diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Subwoofers (http://www.diyaudio.com/forums/subwoofers/)
-   -   Dipole & Horn sub (http://www.diyaudio.com/forums/subwoofers/95647-dipole-horn-sub.html)

paulspencer 4th February 2007 03:05 AM

Dipole & Horn sub
 
For a long time I've been interested in bass horns and dipole woofers. I've come to the conclusion that the ultimate is a dipole woofer (open baffle) with a bass horn subwoofer crossing from one to the other in the midbass.

Currently I have two AE speakers AV12 subs in a push pull sealed box. For a long time I've wanted to put them in a bass horn, and I now have plans to move into a space that can actually fit them. They can hit about 118 dB at most, but in a bass horn more like 140 dB. Not that I want insane output, I just want them to be more accurate, I think a bass horn will do it.

I've written about it in a past thread:

http://www.diyaudio.com/forums/showt...05#post1030305

Previous experiments with open baffle really impressed me. My existing drivers were a LOT better in this configuration. FYI shown here:

http://www.diyaudio.com/forums/showt...threadid=80908

You can see a H frame in the photo

What I noticed is that no EQ was required to get down to about 30 Hz much to my surprise. I also found that sitting up close to the woofer the output was quite decent. So here is my idea:

Coffee table dipole for 30 - 100 Hz (or something like that)

Something looking like this:

http://www.geocities.com/etude316/Fo...ole_woofer.gif

Perhaps with glass on top!

questions

1. How is D calculated? Is it the depth of the sides (on the right and left hand side as seen on the section view in the middle)

2. Is it a good idea to stuff it to reduce resonances?

3. Will increasing the width reduce the effect of resonances when compared to small and narrow openings?

4. Am I going to need to use a notch filter?

Since this effectively has a transmission line, it seems surprising not to hear talk of using stuffing.

Calvin 4th February 2007 10:28 AM

Hi,

the basic calculations You´ll find -amongst others- on Linkwitz´s website. And there are some simulation progs that give good results.
From my experience with dipolebasses though I prefer a slightly different way of constructing dipole subs than SL.
Turning one of the drivers 180° around so that the front sides of the membranes face each other will have some advantages.
- You can build the sub smaller with considerable less width. That makes it optically much more pleasing and has following effects:
- the free-air-resonance Fs is reduced more in a smaller cabinet
- the Qt rises slightly.
Both effects are positive in that way that You can use drivers with a stiffer supension i.e. higher Fs which is advantageous when stressing the sub with higher power.
Most drivers feature Qts <0.5. So a little help by the smaller cabinet is very useful and may make additional equing/bass boosting obsolete (still though I recommend using a subsonic-filter, even if You don´t use its boosting functions)

Increasing the width of the chambers won´t influence the magnitude of the resonance (between 150 and 400Hz depending on the cabinet´s dimensions) much, but defines the centre frequency of the resonance. For higher bandwidth smaller dimensions are better here too.

As a rule of thumb, I take the dimensions of the drivers shipping cartonage as the dimensions of the 2 outer chambers and app 1/2 to 3/4 of those chambers inner width as the width of the third front chamber (greater value for longer throw drivers). A driver in such a small dipole cabinet should feature Fs values of >25Hz up to 45Hz since You can expect an Fs reduction of ~10Hz.

Stuffing will be of no good in most cases, since it just costs energy in first place. There will be no real damping effect -what should its use be anyway?- since the physical dimensions of the material are way too small to have an effect at this low freq-range. The chambers do work like TLs, but in a freq-range thats way above the crossover freq. So this effect should have only a very small influence on the sound, but stuffing would have a serious effect on the speakers efficiency in its working range.

I always recommend to use a passive notch, consisting of a series inductance (ohmic value ~0.4...0.8) and a parallel notch (RCL-series connected, whereby in many cases the ohmic value of the inductance can be the complete R-value, so that the RCL reduces to an CL-circuit).
The passive notch lowers the Fs a bit, raises the Q and even efficiency will in most cases remain the same (the losses by the series inductance are countered by the higher Qt-value)
Adding another parallel C such that the filter looks like a 12dB+par. notch will reduce the Fs further by a couple of Hz.


jauu
Calvin

paulspencer 5th February 2007 10:50 AM

Hi Calvin and thanks for your input. If I recall from previous threads, you are a bit of a Ripole enthusiast.

I've been to Linkwitz site a gazillion times, but last I checked couldn't find this one.

If I'm not mistaken, you are talking about a ripole? They do seem interesting but aren't my preference. Size isn't the issue here since this is a coffee table, and I'm thinking of making it larger than the dipole would actually require. I also like to see the drivers, and am also considering a H frame.

Since my first post I have put my existing AV12 subs in a H frame placed in the coffee table position, and used EQ to get it flat down to ~23 Hz. It's quite surprising that no boost is required. Power handling is reduced only because less power is required for a given excursion - no air mass in a box.

Subjectively I do tend to think dynamics are more pronounced at times.

I also suspect that distortion should in fact increase since it now is more dependent on the linearity of the driver suspension system and motor. In a box the air mass behaves in a more linear fashion. This has been mentioned by Danley in regards to horns and the sealed chamber.

Quote:

Stuffing will be of no good in most cases, since it just costs energy in first place. There will be no real damping effect -what should its use be anyway?- since the physical dimensions of the material are way too small to have an effect at this low freq-range. The chambers do work like TLs, but in a freq-range thats way above the crossover freq. So this effect should have only a very small influence on the sound, but stuffing would have a serious effect on the speakers efficiency in its working range.
If there is no effect, how will it cost energy?

You say that it will reduce efficiency without having any real desired effect?

To me this sounds a little back the front. Below 100 Hz any stuffing could be considered acoustically transparent, having no impact on the efficiency here. The desired effect is that it would damp resonances occurring higher in the range in which it can have attenuation. This is the logic behind the idea. This would stuff up being able to see the drivers and pose a few practical challenges.

Looking at the in-room response with my RTA, I see no resonances, but they may be too narrow to show up. I'm not excited about having to make a notch filter, as it means wrestling with speaker workshop and finally building one of those mic preamps! I'm sure I'll do it eventually though.

Calvin 5th February 2007 03:00 PM

Hi Paul,

having built quite a few subwoofers (amongst were profs with up to 12 15"er drivers built up from scratch), I was convinced that the classical " large membrane area+Qt=0.5+active equing" was the best and most precise sounding bass -apart from some active feedback things. This holds probabely still true for PA or open air usage.
But when I first listened to a good dipole sub in a normal listening room (it was the BMC400 by Ridthaler), I was flattened!
Not only was this baby tiny (just 68L for 2 15"er!!) but it played the most precise and deep bass I´d thought could be possible.
Over the years I gained experience with these ´folded baffle´-dipoles and loved them more and more. Building electrostats, you ´ll probabely always search for a bass that can accompany an ESL harmonically.
The dipole is imo the only (subwoofer-)solution that is actually able to rise an ESLs quality. Its sonic footprint is similar to an ESL -very precise, no boom or boxiness and ´speed´.
Besides that, it often integrates in the room better, reducing problems with room modes without the need of voluminous room treatment. Their dipolar distribution character is ideal to reduce microphonic effects, when You have to place Your system between/besides the speakers (think of Tube gear or vinyl).
The setting up of the right parameters simply by ear is nearly impossible with other subs, because if You change one parameter -eg. the crossover-freq- not just this one parameter but two or more parameters will change at once -eg the volume changes too. The high linearity of the dipole subs makes it much easier to find the right parameters, because shifting one parameter just changes this one parameter.
What might be one of the most valuable features -though it seems, that hardly anyone knows about- is the ´annoyance-factor´. Subs can be very troublesome if You have noise sensitive neighbours/family. A dipole plays just loud in the listening room! Outside the room it is distinctively lower in level, than any other sub I know! It doesn´t exite room boundaries as much as normal subs do.
With all these features together in one principle, I couldn´t help but becoming a fan ;)

Size of dipole subs:
While a simple open baffle woofer shows better results the bigger it is, things change when You use an folded baffle (name it H, N, W....pipapo). In this case building smaller leads imo to better results. Btw. with the proposed arrangement, You´d be able to see both membranes when looking from the front side. And there is symmetry, while in most dipoles built in the Linkwitz-fashion there is not.

Damping effects:
At no place have I said no effects! I just said, that the negative effect within the working range will probabely be larger than any positive effect in the filtered freq-range.


jauu
Calvin

paulspencer 5th February 2007 10:47 PM

At this stage, the biggest question in my mind is whether the coffee table concept is viable or not. I'd be crossing at 80 - 100 Hz over to some OB dipole mains. I may have to put the woofers with the main panels. What appeals about the coffee table concept is the output advantage, as well as being able to see the drivers.

noodle_snacks 5th February 2007 11:14 PM

Assuming you are using your AV12s + ultracurve. What is to stop you building a cheap, MDF prototype then equalising to see how it goes? This is similar to what I did with my dual 15in H-Frame and my BFD. The problem in my case was output capability (the cheap drivers i used had some noise via the vented polepiece once you got to any reasonable output level, but you do have a more powerful amplifier than I do, and i believe drivers with more displacement, and the sub you build may have a longer pathlength.

If you were to make a final version, black melamine MDF (not the particle board crap). Edged, then with perspex on top would look pretty good, assuming it was well made. You do either have to cut the melamine with a scribing panel saw or a router to avoid tearout. But imho its far less work than painting if you don't mind the matt finish and cutting a few mitres (get your saw setup accurately beforehand). Which reminds me, did i ever send a photo of the desks i made to you?

paulspencer 6th February 2007 12:30 AM

Nope, you didn't.

I have a h frame coffee table set up right now. I think the Rythmik servo kits will be better though.

When horn loaded, I expect my AV12s will improve a lot, not only because of the gain in efficiency (hence much lower excursion), but also as it causes the driver to be more linear by loading both sides of the cone.

To work best in OB, a driver needs a very linear motor as the excursion is higher for a given output. This is why I think the servo system is going to be better suited.

noodle_snacks 6th February 2007 07:58 AM

What sort of device are you going to use to crossover? Wouldn't a steep enough filter negate the need for a notch? The H frame for the phoenix on the linkwitz site only has a relatively broad 10db peak or so, if your coffee table ends up similar, things shouldn't be too hard to notch out. Measurements would help but i think it is really best left to experimentation.

noodle_snacks 6th February 2007 08:05 AM

PM'ed you a photo of desks. What sort of device are you going to use to crossover? Would a steep enough filter negate the need for a notch? The H frame for the phoenix on the linkwitz site only has a relatively broad 10db peak or so, if your coffee table ends up similar, things shouldn't be too hard to notch out, however ideally i'd do this after the crossover (because the subs might have higher distortion than your mains would, and just correcting the net response at your listening position might lead to greater variation as you move around hte room. Measurements would help but i think it is really best left to experimentation.

Calvin 6th February 2007 08:27 AM

Hi,

the problem of the folded dipoles is simply that the chambers which are formed through the folding show a resonance, thats peak frequency depends on the physical dimensions oft he chambers and that lies the higher in frequency the smaller the dimensions are.
A sub with 15"ers can´t be build smaller than this one:
http://www.diyaudio.com/forums/showt...threadid=95647
As You might see from the measurements the peak is at ~250Hz. If You want to stay away from this peak roughly one octave below, You´ll end up with a useful bandwidth maximum of ~150Hz.

jauu
Calvin


All times are GMT. The time now is 05:44 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2