3 way open baffle shape - diyAudio
Go Back   Home > Forums > Loudspeakers > Multi-Way

Multi-Way Conventional loudspeakers with crossovers

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 14th April 2010, 02:45 AM   #1
diyAudio Member
 
otto88's Avatar
 
Join Date: Dec 2007
Default 3 way open baffle shape

Hi

I welcome comments on this 3 way open baffle. I’m mostly seeking comments on the baffle shape.

I just received the final set of drivers, so they are set. (After losing my workshop in a marriage split , I’ve just gained access to a good workshop).

I've designed it to both appeal to my sense of aesthetics, but be effective sonically, ie not have any flaws like resonances from large wings/ at the wrong frequencies.

My understanding is that
- The baffle for the mids and treble drivers should be narrow: so that part of baffle - as you can see in the attached pdf, is not much bigger than the 230 mm (c 9 inch) driver.
- The baffle for the bass (actually midbass) should have any “wings” well away from the mid driver, preferably tapered to spread their resonance.

We don’t have birch ply here in Australia, so I will make it out of 18 or 19 mm marine ply. Luckily I can get this at trade prices.

The drivers:
It’s to use with an 845 SET amp of just 16 watts, hence the need for above average sensitivity.
The key part of the speaker is a Supravox wide range driver tailored (in conjunction with Thorsten) for open baffle use, with a QTS of about 0.8, and increased Xmax. The makers specs show it's flat on and off axis to about 8 – 10 kHz.

Above that I’ll use an Eminence supertweeter, and possibly also have one rear firing, on a pivot or gimbal, so it’d be dipole full range. I don’t hear much above 10 kHz so I may not bother with this.

I have a pair of very good active open baffler subwoofers, that can be crossed at up to 150 Hz (when I get them from my ex-wife).

In the meantime I’ll use a Lambda 15 inch midbass. Although the Supravox can go quite deep, Thorsten has said they sound strained at higher levels, so a midbass would take a lot of the load off them, for when I’m having a blast.

Baffle size:
The main baffle is 50 mm wider a side, than the 15 inch midbass, ie 500 mm. Though I don’t believe in sonic benefits of the Golden Ratio, I like the proportions of 800 * 500. The centre of the wide range driver is about 950 mm above floor level, which is about the height of my ears when seated.

The base is 400 mm deep, with a simple triangular wing extending back 350 mm. Maybe I should make the wing higher or longer, for added strength?

As the drivers are so heavy: the Supravox is 9 kg and the midbass 10.5 kg, I intend to have a pair of braces supporting each of their magnets. As it’ll weigh quite a bit, and to easily fine tune location, I’ll have it on castors.

Anyhow, how are the baffle shape and wings?
Attached Files
File Type: pdf 3 way open baffle v1.pdf (38.2 KB, 585 views)
  Reply With Quote
Old 14th April 2010, 03:43 AM   #2
dcathro is offline dcathro  Australia
diyAudio Member
 
Join Date: Mar 2007
Very intereting Otto!

I asume the Supravoxs are the field coils. I am using the same driver in a 3 way semi open baffle system (bass is in a sealed box). Although I am running the Supravox full range and just supplementing the top and bottom.

I have the supravox and tweeter on a hexagonaly shaped baffle that is 360mm at its widest.

With your circular baffle shapes on the mid and tweeter, you are going to need quite a bit of equalization to get a pretty flat response.

How are you going to do the crossover?

Cheers

David
  Reply With Quote
Old 14th April 2010, 04:21 AM   #3
diyAudio Member
 
otto88's Avatar
 
Join Date: Dec 2007
Hi David

Yes they’re the field coils.

The crossover?
Audiology tests confirmed that I cant hear much above 8 kHz(!), so the super-tweeters are possibly not necessary for me.

Both the Supravoxs and the Lambdas are very flat. The Lambdas are about 3 dB more sensitive, so I’ll need L-pads on the Supravoxs.
As there’s enormous flat overlap between these two drivers, I could probabaly cross anywhere from > 200 to evn > 1000 Hz, but will probably cross at about 450 - 500 Hz.

What I need to learn/ get advice and/ or work out: how to keep the phase correct through the crossover. Maybe 2nd order LR is best?

I did not know that circular baffle shapes on the mid and tweeter need quite a bit of equalization, are you sure?

If that is correct, I really like this shape, so I would probably have acrylic wings around these circular baffles, with straight edges. That would allow whatever was the optimal width, for the frequencies to be carried.

Cheers

Last edited by otto88; 14th April 2010 at 04:23 AM.
  Reply With Quote
Old 14th April 2010, 05:10 AM   #4
diyAudio Member
 
Join Date: Jun 2009
If you do some simulations in Edge you'll see the height of the dipole peak is an inverse function of the baffle's extent beyond the radiating area of the driver. So consider dispensing with the baffle on the upper part of the Supravox.

The primary criteria for the U baffle sizing is the effective path length needed to hit the desired SPL. Doing the maths on that should help you pin it down. You'll likely want some reinforcement between the fullrange and sub to better manage baffle vibration, so take that into account as well.

Any odd order Butterworth or even order Linkwitz-Riley is arguably phase correct, at least in the warped phase sense. Personally with an analog crossover I'd cross this LR4 as low as the Supravox could comfortably handle. Linear phase aside, I've found digital is pretty handy for two way dipoles as dialing in the equalization and correcting for driver limitations is much quicker than with analog.

With respect to sonic flaws are you aware this approach will have a fair amount of directivity variation? Check in Edge, but off the top of my head I'd guess the dipole peak on the Supravox with the baffle shown will be around 1kHz. I'd expect front to back wave symmetry won't be particularly good either due to magnet and spider obstruction.
  Reply With Quote
Old 14th April 2010, 06:51 AM   #5
diyAudio Member
 
otto88's Avatar
 
Join Date: Dec 2007
Thanks for your good points and questions

“If you do some simulations in Edge you'll see the height of the dipole peak is an inverse function of the baffle's extent beyond the radiating area of the driver. So consider dispensing with the baffle on the upper part of the Supravox”

The baffle on the upper part of the Supravox is intended to be not much more than wide enough for strength, the driver weighing 9 kg (and to support the baffle for the super tweeter). Though if I brace the magnet at the rear, as shown on the right hand side, the upper baffle can be very narrow, say 10 or 12 mm.

“The primary criteria for the U baffle sizing is the effective path length needed to hit the desired SPL. Doing the maths on that should help you pin it down. You'll likely want some reinforcement between the fullrange and sub to better manage baffle vibration, so take that into account as well”.

With the midbass’ Xmax a moderate 6 mm, Sd of 855 cm2, baffle 500 mm wide and wings with an average depth of 180 mm, I didn’t think the desired SPL would be an issue, down to about 40 Hz. Later I will supplement the deep bass with subs.
Qes is only 0.35, but offsetting this, but the tube amp will have Rs of about 2.5 ohms. (I might also try the 16 ohm taps, power will be limited, but bass should be extended: maybe for listening at lower levels).

But thanks, I will estimate the path length. That’s I believe from the centre of the driver cone (in the Lambda, where there’s a phase plug) - the shortest path to the back of the wing?

“Any odd order Butterworth or even order Linkwitz-Riley is arguably phase correct, at least in the warped phase sense”

What is *warped phase?


“Linear phase aside, I've found digital is pretty handy for two way dipoles as dialling in the equalization and correcting for driver limitations is much quicker than with analog.”

By digital do you mean an active XO to the bass? I’d almost forgotten, I have a Behringer DCX-2496 I could use . .

“Personally with an analog crossover I'd cross this LR4 as low as the Supravox could comfortably handle.”

The DCX would allow easy experimenting with XO point and slope. Then implement in analog. I’m not sure though, how closely the results would translate

“I'd expect front to back wave symmetry won't be particularly good either due to magnet and spider obstruction”

. . of these drivers?
You’re probably right about the Supravox. I chose it for its good on and off response, lack of peaks and high sensitivity (the ability to tweak the Q may be useful). Can’t have everything.
The Lambda is fairly open at the back

“With respect to sonic flaws are you aware this approach will have a fair amount of directivity variation? Check in Edge, but off the top of my head I'd guess the dipole peak on the Supravox with the baffle shown will be around 1 kHz”

I thought the Edge modelled baffle step, not the peak of a dipole?
Whether the peak is 800 or 1600, it would need to be balanced in crossover, so is their a benefit in varying that point by a different baffle width?
  Reply With Quote
Old 14th April 2010, 08:42 AM   #6
diyAudio Member
 
gainphile's Avatar
 
Join Date: Dec 2007
Location: Melbourne the sunny city!
Your design looks correct acoustically. The all-important midrange would have smooth on and off-axis response to low-2khz. But this would look like a peak, which is expected with any dipole speakers. EQ this peak using notch filter ... try f=650hz Q=1.4 depth=-6db and work from there (if you have measurement tools then it's easy to find the exact notch). Then add shelving lowpass to 1 oct. below W-M xo frequency.

Woofer EQ is easy, just 6db shelving lowpass.

Let us know how you go!
__________________
http://gainphile.blogspot.com
  Reply With Quote
Old 14th April 2010, 09:48 AM   #7
CLS is offline CLS  Taiwan
diyAudio Member
 
Join Date: Jun 2005
Location: Taiwan
Since the mid-high section is very narrow already, you might want to try naked driver without baffle here

See My open baffle dipole with Beyma TPL-150
(that system evolved quite a lot later... )


or narrow baffle with suspended driver: Beautiful Swingin' Speaker
  Reply With Quote
Old 14th April 2010, 03:26 PM   #8
diyAudio Member
 
Join Date: Jun 2009
Quote:
Originally Posted by otto88 View Post
The baffle on the upper part of the Supravox is intended to be not much more than wide enough for strength
0mm and attaching the super tweeters to the Supravox is probably fine. Take a look at the end of Stig Erik's thread CLS linked. I didn't suggest hanging the Supravox nude in my previous post since then you'd lose the baffle shape you like so much. But if you're flexible on that hanging is a better solution for sonics.

Quote:
Originally Posted by otto88 View Post
Thatís I believe from the centre of the driver cone (in the Lambda, where thereís a phase plug) - the shortest path to the back of the wing?
With this baffle probably closer to the average path; what you're trying to approximate is the power weighted integral over all paths. That's what Edge calculates, just not for folded baffles.

Quote:
Originally Posted by otto88 View Post
The DCX would allow easy experimenting with XO point and slope. Then implement in analog. Iím not sure though, how closely the results would translate
For warped phase crossovers you can translate pretty much anything you want to a Sallen-Key, biquad, or state variable. Not sure what the DCX supports in the linear phase---some folks have mentioned FIR but last I checked the manual it didn't mention FIR or whether the DCX does linear phase FIR synthesis---but transient perfect/filters with constant group delay (as opposed to warped phase filters with varying group delay) are much harder to implement in analog. See cuibono's Violet DSP thread in this forum for a fair amount of discussion on this.

Quote:
Originally Posted by otto88 View Post
I thought the Edge modelled baffle step, not the peak of a dipole?
Depends on whether or not the open baffle check box is checked. The general rule of thumb is if you want decent directivity don't cross drivers above their dipole peak. If you want good directivity a more stringent metric is to cross the driver so there's no significant SPL above the dipole peak---personally I define significant as 35 to 40dB down, but vary to taste based on design goals and the eveness of the back wave. It's best to measure the back wave (IIRC look for Rudolph's threads here; he found radiation patterns are rather less intuitive than you might think). While you can get a loose estimation of how even a pattern the driver might support from how much cone is exposed past the spider and magnet, whether the pole piece is vented, and so on, the actual pattern's hard to predict.

I'd likewise expect the Supravox to be the main issue here. In the case of the AE the U baffle will shift the pattern away from dipole. I don't work with U baffles but my guess would be probably not by too much in the current form---I recall coming across some measurements that found a ~2cm deep U baffle actually improved pattern evenness slightly. (Gainphile, did you do those?) If you cross low I'd bet room effects will be more significant. If you cross high the limitation's probably beaming on the 15; not sure how much the U baffle lowers it but the dipole peak with a 500mm flat baffle's probably around 500Hz. If you're thinking analog and LR4 and want good directivity transitioning off the bass I'd try to cross 200Hz or below.
  Reply With Quote
Old 14th April 2010, 09:57 PM   #9
dcathro is offline dcathro  Australia
diyAudio Member
 
Join Date: Mar 2007
Hi Otto,

Quote:
Originally Posted by otto88 View Post
Hi David

Yes they’re the field coils.
Nice choice! It gives great sound and tremendous tunability.

Quote:
The crossover?
Audiology tests confirmed that I cant hear much above 8 kHz(!), so the super-tweeters are possibly not necessary for me.
You might not be able to hear sign waves above 10K, but I bet you will hear the result of the missing top frequencies on transient response and music harmonics.

I found the addition of a tweeter was a big help for the Supravox full range. I started trying to bring it in above 6K with a 1st order filter, but found it got better as I took it higher and ended up at 18K!. Interestingly, this is also what Thorsten found.

Quote:
Both the Supravoxs and the Lambdas are very flat. The Lambdas are about 3 dB more sensitive, so I’ll need L-pads on the Supravoxs.
Sorry, if the lambdas are more sensitive then you won't need to pad the Supravoxs. Whatever you need, the field coil allows you to vary the sensitivity of the Supravox quite drastically without the need for padding.

Quote:
As there’s enormous flat overlap between these two drivers, I could probabaly cross anywhere from > 200 to evn > 1000 Hz, but will probably cross at about 450 - 500 Hz.

What I need to learn/ get advice and/ or work out: how to keep the phase correct through the crossover. Maybe 2nd order LR is best?
With the Supravox capable of going down quite low, I would try and cross it over as low as I could - 250Hz. That way you get out of the vocal range.

Also, with a driver like the Supravox, I would only use 1st order filters. It just seems to me that using higher order filters or digital filters on the Supravoxs is like taking a good bordeaux and making a spritzer .

Quote:
I did not know that circular baffle shapes on the mid and tweeter need quite a bit of equalization, are you sure?

If that is correct, I really like this shape, so I would probably have acrylic wings around these circular baffles, with straight edges. That would allow whatever was the optimal width, for the frequencies to be carried.
As others have said, have a play around in edge, and make sure you tick open baffle. Having a large part of the baffle at the same distance from the driver center will give you quite a high baffle peak.

My baffle peak is centered arount 600Hz, ranging from 250Hz to 1300Hz with a 5db peak. I use a passive paralled LRC in series to notch this out.

Cheers

David

Last edited by dcathro; 14th April 2010 at 10:04 PM.
  Reply With Quote
Old 15th April 2010, 05:27 AM   #10
diyAudio Member
 
otto88's Avatar
 
Join Date: Dec 2007
CLS:
From the thread “My open baffle dipole with Beyma TPL-150”

“I've tried baffle-less "midrange" of 12 inch driver, which could play down to about 300 Hz with little EQ.
For an 8 incher, I guess you'll need some more EQ, or the lower range may suffer somewhat”

The Supravox is about 2 dB less sensitive than the Lambda. I could happily live with the look of baffle-less diver, but don’t want to lose sensitivity by using EQ.

One Linkwitz Orion owner suspends his mid panel from the ceiling. I’d rather the more flexible Beautiful Swingin' Speaker approach. It should give some benefit isolating the mid from the 15”, though I don’t know just how audible it would be.

I thought that when the active OB subs are later implemented, I might take the Supravox down to the subs' maximum XO of 160 Hz.
Then, if the baffle is fixed (ie not swinging), clear acrylic extensions could be added either side, as in Nuuk’s and others, to allow them to play that low. But probably I should plan to always use the midbass, to avoid sound “strain” at higher levels - for plenty of headroom . .

. . if the midbass is now likely to continue in this system when subs are in, I need to get that lower crossover right.


TW

Attaching the super tweeters to the Supravox - don’t know how I'd do it: a metal frame extending up and forwards off the Supravox coil, attaching to the tweeter’s magnet = like an alien’s eye on an arm??. It’s probably better to time align those two physically, though I don’t know how accurate it will be in practise.

I was socialising last night, but tonight will play with the Edge.

I’d just use the DCX (which has 3 XO types, has Butterworth, Bessel or LR) to experiment with frequencies, rather than ongoing. I have some of Bob Ellis’ group buy boards from a few years ago if I decide to go active.

I do want “good” directivity, so need to find out the dipole peak. Looks like the Lambda will be crossed lower than I first thought. They are the probably the flattest on and off axis above the range of a typical 15”, of any 15, so that’s a bit of a waste . . I might go down to


BTW before I went out last night I used Linkwitz’ spreadsheet Max SPL.xls, and learnt that the midbass Vd on the intended baffle allows it to reach 45 Hz at 99 dB, into half space. I’d think that if the driver is at the very bottom of the baffle, and that’s on castors, it’d be pretty close to half space.

Last edited by otto88; 15th April 2010 at 05:35 AM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Most unusual baffle shape? madisonears Multi-Way 26 1st March 2009 07:05 AM
open baffle shape smak Multi-Way 21 27th October 2005 07:18 PM
Size and shape of infinite baffle j.burtt Multi-Way 22 27th September 2005 01:26 AM
ideal baffle/enclosure shape for OB? akunec Multi-Way 5 30th May 2005 01:14 AM
Australians- what solid timbers for baffle? (open baffle loudspeaker) tktran Multi-Way 13 29th November 2004 11:09 PM


New To Site? Need Help?

All times are GMT. The time now is 06:51 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