On the Subject of Edges

[sunbeamgls]

So whatis it with speaker baffle edges?

Squared off? Rounded? If rounded, is it a small rad or a large rad? 

Should your baffle be wide or narrow? Which is better for imaging? Width determines where a crossover needs baffle step to balance edge dispersion effects, but which part of the frequency range is it best to allow this to be dealt with?  If you have a rounded edge baffle, which baffle width measurement is used to calculate baffle step frequency or does is just need to be in the right ballpark?

Why don't more speakers use pods like B&W D8xx series where there is less edge to worry about, even in the bass cabinet of the latest iteration?

What else should be baffling me about these things?

 

[awkwardbydesign]

http://www.linkwitzlab.com/diffraction.htm  Have a read.

 

[rv295]

Baffle step relates far more the dimensions of the baffle than the edges themselves.  Rounded edges are better (providing a wider dispersion of the wave pattern as it reaches the corner) but a round edge will not fix baffle step.

Are you thinking about building some open baffle speakers? 

If you want to simulate and learn more about baffle step I'd recommend having a play around with these:

http://www.tolvan.com/index.php?page=/edge/edge.php

http://www.tolvan.com/index.php?page=/basta/basta.php

 

[rv295]

Intelligent design will incorporate the baffle size and driver response together as one.

 

[awkwardbydesign]

Intelligent design will incorporate the baffle size and driver response together as one.

When I had an active system, using Jordan drivers, I was able to cross over at the BSC frequency, and being active simply adjust the gain to level the drivers' responses. Less simple with a passive crossover.

 

[sunbeamgls]

The crossover will be Exakt, so really easy to design (at least the basics will be, getting it tuned right won't be quite so simple).

It won't be open baffle as such, but I was thinking the upper mid, tweeter Ns super tweeter might work as small sealed boxes but standing proud (as insprouting out of the top of the baffle) of the box I'll put the lower mid and woofer into (with the lower mid having its own sub enclosure inside the main box.

I will read the link and see where that takes me. Thanks.

 

[awkwardbydesign]

Would you physically time align the drivers?  With that system you easily can.

 

[sunbeamgls]

Would you physically time align the drivers?  With that system you easily can.

Yes, of course

 

[rv295]

I find visualising sound waves differently in different situations helps me understand things that otherwise made no sense.

In terms of baffle step; I found the easiest way to understand what is going on - is to imagine the sound waves as simply a high pressure area.  As the sound wave (or high pressure area) leaves the speaker it naturally wants to equalise and seeks lower pressure areas to move in to.  

As the pressure (or sound wave) reaches the edge of the baffle (cabinet or open baffle) it discovers low pressure areas to the side and behind the speaker and starts to move in to these areas to equalise out (high pressure naturally seeks low pressure areas to equal out).

If the baffle is wide enough for a full wave to travel across it without reaching an edge, this pressure is dispersed forward to the listener.  If a wavelength is longer than the distance to the edge of the baffle, its pressure will move around the sides and rear of the baffle (cabinet or open baffle).  As the whole wavelength pressure is connected, some of the wave energy is lost to the sides and rear and less pressure is moved forward to the listener.  

So frequencies with wavelengths that are longer than the distance to the edge of the baffle will lose energy to the sides and rear of the speaker.  This causes the roll-off in volume in the longer wavelengths.

...

I don't mean to teach you guys how to suck eggs, I just wanted to share that for anyone interested and struggling to understand why it happens.

When I had an active system, using Jordan drivers, I was able to cross over at the BSC frequency, and being active simply adjust the gain to level the drivers' responses. Less simple with a passive crossover.

That's the joy of digital active crossovers, you can almost throw the rule book out the window.  I know I did.

 

[awkwardbydesign]

That's the joy of digital active crossovers, you can almost throw the rule book out the window.  I know I did.

That's the usual assumption, however it is an analogue crossover!  A Ben Duncan design, published in HiFi News in 1981.  I built it then, have updated several times since, and it is still very good.  I prefer to keep the signal in the analogue domain, rather than run it through an ADC/DAC process. 

 

[orangeart]

So rounded corners are only useful if thier radius approaches that of the wavelength affected. So on a wide baffle speaker where the width dictates that shorter wavelengths have already radiated in a 2pi fashion (that is forward) those wavelengths won't really be helped by a round over. The longer wavelengths will however an so having a large round over here is of great benefit. Some of you will remember my wide baffle 4 ways with vacuum formed sides from Scalford last year and the year before.

To visualise how sound is affected by the edges it is easier to talk about them as reflections or more accurately illuminations. Picture a flat pond into which you drop a stone, the wavefront of the ripple travels out equally in all directions. If that wavefront encounters a twig dangling into the water a new ripple will start and travel out in all directions from that point as well. Those two wavefronts will interact now, sometimes reinforcing and sometimes canceling each other. The pebble is your driver and the twig is your speaker edge.

Wider baffle designs sometimes get a bad rap for not imaging so well, this has a lot to do with the potential for the edge diffraction interacting with the lower mods rather than higher up like in narrower speakers, very large round overs like in my speakers will sort this.

I always try and work my crossover points so as to coincide with the baffle step. Works a treat!

Stefan

 

[rv295]

That's the usual assumption, however it is an analogue crossover!  A Ben Duncan design, published in HiFi News in 1981.  I built it then, have updated several times since, and it is still very good.  I prefer to keep the signal in the analogue domain, rather than run it through an ADC/DAC process.

Very true, I did assume you meant digital active as it seems to be the norm in active crossovers these days.  Analogue active crossovers were simply the stuff of dreams for me, for many years.  I didn't (and still don't) have the knowledge required to build my own and I didn't have the funds to purchase any.   I'd have started DIY speaker building years ago if I had some active crossovers.

So rounded corners are only useful if thier radius approaches that of the wavelength affected. So on a wide baffle speaker where the width dictates that shorter wavelengths have already radiated in a 2pi fashion (that is forward) those wavelengths won't really be helped by a round over. The longer wavelengths will however an so having a large round over here is of great benefit. Some of you will remember my wide baffle 4 ways with vacuum formed sides from Scalford last year and the year before.

To visualise how sound is affected by the edges it is easier to talk about them as reflections or more accurately illuminations. Picture a flat pond into which you drop a stone, the wavefront of the ripple travels out equally in all directions. If that wavefront encounters a twig dangling into the water a new ripple will start and travel out in all directions from that point as well. Those two wavefronts will interact now, sometimes reinforcing and sometimes canceling each other. The pebble is your driver and the twig is your speaker edge.

Wider baffle designs sometimes get a bad rap for not imaging so well, this has a lot to do with the potential for the edge diffraction interacting with the lower mods rather than higher up like in narrower speakers, very large round overs like in my speakers will sort this.

I always try and work my crossover points so as to coincide with the baffle step. Works a treat!

Stefan

Nice explanation Stefan, I find it really helps to visualise what is going on in order to get a better understanding. These kind of descriptive ways of explain things are really valuable. :^

 

[awkwardbydesign]

Very true, I did assume you meant digital active as it seems to be the norm in active crossovers these days.  Analogue active crossovers were simply the stuff of dreams for me, for many years.  I didn't (and still don't) have the knowledge required to build my own and I didn't have the funds to purchase any.   I'd have started DIY speaker building years ago if I had some active crossovers.

Since I went passive 3 way I have been struggling to recapture what I had!  I only changed because my wife couldn't handle the complexity, especially if "something happened to me"!  I have a habit of crashing my bike on trackdays. 

Next one is next Friday!

 

[danilo]

IMO it's simple Do what pleases you aesthetically. Seriously unlikely you will hear differences.

I too once thought old Linkowitz was a genius... he reads (writes?) like one.

Then I heard his setup. Mehhh is a charitable descriptor.   Avoid imo.

 Spherical speakers  surface periodically. V low  WAF though.

 Some stuff  wee single cone/full range drivers  (no Xover ;-)  into a glued together pair of IKEA Bamboo salad bowls.

 Works better than it sounds actually.

 G'luck

 

[Speedskater]

Off Topic

Are any of the 30 year old Ben Duncan articles available on line?

I still have some of the Hi-Fi News magazines but have lost others.

 

[awkwardbydesign]

Off Topic

Are any of the 30 year old Ben Duncan articles available on line?

I still have some of the Hi-Fi News magazines but have lost others.

Here you go- http://www.saturn-sound.com/history/hi-fi, diy projects.htm  Scroll down to 1981.  He is still active (sic) and you can get updates from him, but he is a bit crusty if you get into a disagreement with him!  I updated the ICs, caps and slopes .http://s1134.photobucket.com/user/awkwardbydesign/library/Active crossover