Baffle diffraction ~ open baffle

rick57 · « **on:** 15 Dec 2006, 02:08 pm »

I'm in the early stages of designing a 3 way open baffle speaker; likely at this stage in its evolution, to achieve an Fequal about 70-80 Hz constrained by WAF to have a small baffle, supplimented by folding wings.

This is likely to mean some drivers very close to a baffle edge. Most likely the woofer or the tweeter.

per SL www.linkwitzlab.com/diffraction.htm baffle diffraction in an open baffle is not so important as in box speakers.

I understand that the most important thing is that the distance from centre of drivers to the edge of the baffle ~ varies, which I think means that the more off cente the better. Am I right; any other rules?

While I'm itching to get going on this, at present several other things take priority - does either the Edge or MJK's model factor in baffle diffraction, or regardless, do drivers very close to a baffle edge cause a problem?

mcgsxr · « **Reply #1 on:** 15 Dec 2006, 03:41 pm »

I am far from a tech guy, when it comes to these, so take my answer as strictly subjective, experiential opinion.

All I have noticed (ears, not measuring) is that the closer a driver that produces bass is to the edge of the baffle, the less bass there is likely to be - I think this is because the shortest distance "around" the baffle is where the first cancellation begins to occur.

You are correct about offsetting the driver within the baffle, in order to spread the spectrum of frequencies where diffraction will occur. There seem to be multiple schools of thought on how to deal with this phenomenon. Some say offset the driver, spread the spectrum, and there you go. Others favour central driver, tighter spectrum of frequencies, and use eq to limit the ultimate effect.

If I were building a multi way open baffle, I would follow the development that Jeffbehr did, and likely seek to have the mids and highs closer to the edge, with the bass unit relatively central, but offset.

Just my thoughts.

scorpion · « **Reply #2 on:** 16 Dec 2006, 01:02 am »

Rick57,

I would advice in favour of separate bass- and mid-top baffles. Otherwise I think you would have to have
at least 2-3 " thick bass baffles. With regard to placement, you could always correct behind the baffle for
speaker placement.

/Erling

JohninCR · « **Reply #3 on:** 16 Dec 2006, 02:26 am »

Easier than separate baffle is decoupling the woofer from the baffle with a separate structure.
Magnet mount is pretty easy with the woofer near the floor. Since size is of priority, then U-baffle
is the way to go. With the Fequal you want you're looking at a 60" wide flat baffle or a 15" deep
U-baffle, which only needs to be large enough in the other dimensions to fit the driver. I'd choose
multidriver augmentation or U-baffles over monster baffles any day. See JohnK's website at
www.MusicAndDesign.com for U-baffle technical info.

rick57 · « **Reply #4 on:** 17 Dec 2006, 12:51 pm »

Because the distance between the acoustic centre of the drivers should not exceed the (I think it's 1/4) the wavelength of the XO frequency, the tweeter will need to close to the mid, though maybe the mid will not need to be close to the bass unit.

I have read about U baffles; from the comparsion of 11 baffle shapes at www.troelsgravesen.dk/OBS.htm, I might go for posssibly wings that are not symmetrical (troels' option H), back at 60 degress (option J) or maybe a shallow U baffle (option K).

Hinged wings that are adjustable in angle would allow fine tuning.

To return to topic, thinking it through, whatever arrangement of drivers gives the greatest spread (like a lowest Q) of distances from the centre of the drivers to the edges, in both directions, should have the smoothest baffle diffraction pattern. ie asymmetry, which is seen side-side in troelsgravesen's option H.

Cheers

JohninCR · « **Reply #5 on:** 17 Dec 2006, 04:09 pm »

Rick,

Be careful about your reliance on info from that OB study that you linked in your post.
The measurements were all done in the nearfield, most at .5m, which all but eliminates
the rear wave effects from the measured response.

While one might be able to get away with hinged wings with a single fullranger on OB,
I'd caution you against doing so with drivers that make real bass unless you plan to
open them flat or near flat for listening. Despite being open, the forces on folded wings
are significant and you're looking at a lot of vibration with the sizes you need. I've used
quite a few variations of U-baffle hybrid shapes, and they can work very well, but in the
higher frequencies they are more difficult to get right. Using a U only for the bass and
crossing at or below the first dipole peak, as JohnK mentioned, will make life easier.

Rudolf · « **Reply #6 on:** 17 Dec 2006, 07:27 pm »

Putting a driver off axis on an open baffle has assets and drawbacks:
It will give the smoother response - but only On axis.
It will give a better polar response pattern - but only to the Short side of the baffle.

I try to demonstrate that with an OB simulation for the Visaton B 200 mounted on a baffle of 125x50 cm with the driver mounted at 100 cm height - centrally and 15 cm offset to the left:

Look at the polar radiation diagram. Its crude since it accounts only for 0, 30, 60 and 90° left and right. But it´s values have been taken from the actual frontal radiation pattern of the B 200 at those angles (measured in the standard box at Visatons anechoic chamber). Links is left, rechts is right, vorn is front, hinten is back. The red pattern is for the driver in the central position, black is for the offset position (as near to the edge as the driver allows). Actually only the radiation to the front has been measured. The radiation to the back is simulated with the same values, but 180° out of phase.

What do we see:
1. With the driver moved to the left edge of the baffle the dipole 8 is bent to the right (long) side of the baffle. When the driver doesn´t illuminate the baffle edge any longer (from 2500 Hz up), the dipole 8 returns to its normal position.

2. For low frequencies there is a SPL loss to the left side which is not fully compensated by the gain to the right. So for the offset driver you will loose some SPL in the room.

3. While for most frequencies the off axis response to the left stays below the on axis response, the response to the right is more erratic. So you have to listen to the short side of the baffle.

If this is of any interest, I could give the complete 0, 30 and 60° response diagrams for both mounting positions in the usual representation too.

Rudolf
www.dipolplus.de

scorpion · « **Reply #7 on:** 18 Dec 2006, 10:55 am »

Rudolf,

Very good, new knowledge generated.
To me, at least, the full frequency diagrams are interesting.

/Erling

rick57 · « **Reply #8 on:** 18 Dec 2006, 12:16 pm »

JohninCR

> one might be able to get away with hinged wings with a single fullranger on OB

I'm intending a high efficiency three way with wider range divers, with these specs, rather than a fullranger

     Bass   "Midbass" Tweeter
max Hz        110 1300
Min Hz              110 1300
XO     Active Passive   Passive
     two of      Ribbon
dB in box      97     99
Fs     18      35
Re     3.5      6.6
Qms     3.7      4.5
Qes     0.21      0.35
Sd     466      855
Vas     139      405
Xmax peak 12.5 3.3
(Le)     4.2      0.2

Piano hinges could be used, and opened up for listening . .

> I'd caution you against doing so with drivers that make real bass unless you plan to
open them flat or near flat for listening.

The wings could be opened for listening.

> Using a U only for the bass and crossing at or below the first dipole peak, as JohnK mentioned, will make life easier.
I missed this reference - can you link please

Thanks

rick57 · « **Reply #9 on:** 18 Dec 2006, 12:44 pm »

Hi Rudolf

To make sure interpretation of this is correct:
"it´s values have been taken from the actual frontal radiation pattern of the B 200 at those angles (measured in the standard box at Visatons anechoic chamber)"
. . these values are from an open baffle, not a box?

(assuming I have it correct) interesting analysis . .

An SPL loss to one side, if that side is to the side wall ~ should be an advantage.

To see closely, could you post the *static open baffle patterns at eg 40, 100, 500, 2000 (maybe 4000) Hz?

Cheers

Rudolf · « **Reply #10 on:** 18 Dec 2006, 02:42 pm »

Quote from: rick57 on 18 Dec 2006, 12:44 pm

. . these values are from an open baffle, not a box?

Sorry no, these are NOT actual measurements on an open baffle!
The frontal radiation pattern of the B 200 was measured with the driver on a small baffle in a 500 l box under anechoic conditions at 1 m distance:

This ensures that we do not deal with a theoretical driver having the TSP of the B200, but with the actual cone break-up-modes and radiation pattern variations of the B 200.

This pattern was brought into the simulation for an open baffle. The backward radiation of the B 200 was assumed to be the same as the front radiation. Up to 1000 Hz this seems to be justifiable. At 2500 Hz the 90° SPL of the boxed B 200 is 10 dB down from the on-axis value. So the driver does not see very much of the radiation from the back. I took the simulation to 3200 Hz only to demonstrate the return to the axial radiation pattern. It would have been nearly the same for a B 200 without any baffle or even in an IB.
So the results in the 1000-2500 Hz range may be questionable in the details.

Quote

An SPL loss to one side, if that side is to the side wall ~ should be an advantage.

If you put the side with the SPL loss to the side wall, you will listen to the other side - the one with the mixed up frequency response

I believe the better consequence would be to place the OBs with the drivers to the inside and toe them in some way.

Rudolf

I will post the wanted patterns later this day when I am back home.

JohninCR · « **Reply #11 on:** 18 Dec 2006, 03:34 pm »

Rudolph,

I needed a fresh Monday am reread to see past your cool graphic.
I interpret your post that for flat baffles it's probably best to go with
center mounting and leaving the vertical dimension to create the driver
to edge differences to smooth the dipole ripples.

To me it also explains why I've always had a preference in left/right placement
of my hybrid partial U-baffles, and why they sound different. I typically use
angled wings with one shorter in depth than the other. I've always preferred
the sound with the short sides to the center. After toe-in the left/right travel
distances are close to equal in front, but possibly more important, I hear the
rear wave reflections from the "short side" of the baffle.

Rudolf · « **Reply #12 on:** 18 Dec 2006, 11:19 pm »

I follow with the full frequency simulation diagrams for both driver positions on the baffle.
First for the symmetrically mounted B200:

And for the offset mounted B200 (15 cm to the left):

Below 500 Hz and above 2000 Hz everything looks beautiful - but what a mess in between. For the symmetrically mounted driver the distance to the baffle top is the same as to the baffle sides. It certainly would have been better to change the distance to the top. I´m quite shure a shorter distance would have helped.
Obviously the radiation pattern to the short side of the asymmetric baffle looks best - but it´s only one side.

rick57: Those static polar patterns will have to wait til tomorrow. They are not forgotten!

Rudolf

wikin · « **Reply #13 on:** 19 Dec 2006, 02:37 am »

Quote from: Rudolf on 18 Dec 2006, 11:19 pm

I follow with the full frequency simulation diagrams for both driver positions on the baffle.
First for the symmetrically mounted B200:

And for the offset mounted B200 (15 cm to the left):

Below 500 Hz and above 2000 Hz everything looks beautiful - but what a mess in between. For the symmetrically mounted driver the distance to the baffle top is the same as to the baffle sides. It certainly would have been better to change the distance to the top. I´m quite shure a shorter distance would have helped.
Obviously the radiation pattern to the short side of the asymmetric baffle looks best - but it´s only one side.

rick57: Those static polar patterns will have to wait til tomorrow. They are not forgotten!

Rudolf

What happens to the response if you add a BSC circuit consisting a 1.5mH+10ohm in series with the B200 ? Can you please measure it ? Thanks a lot.

rick57 · « **Reply #14 on:** 19 Dec 2006, 09:57 am »

Rudolph,
Do you agree with JohninCR's interpretation that it's probably best to go with
center mounting, leaving the vertical dimension to create driver to edge differences to smooth the ripples?
Thanks
Rick

Rudolf · « **Reply #15 on:** 19 Dec 2006, 11:03 pm »

JohninCR · « **Reply #16 on:** 20 Dec 2006, 12:19 am »

Quote from: Rudolf on 19 Dec 2006, 11:03 pm

And yes, I would agree with JohninCR quite some way....

It's nice we approaches from opposite ends of the spectrum meet. Good things will result.

Regarding The Wall, great album, but I guarantee smoother response would have resulted
by addressing edge diffraction at least at the side and top near the driver. I have a couple
of very narrow cabs for a pair of my FE108's and changing the edge geometry makes a HUGE
difference in the sound with the driver very near the edge of a cab, at least for boxes. For
OBs that is something on my experimental measurement list, because I'm with Rudolph that
much of the edge diffraction with OB's doesn't net to zero and is directional in nature. JohnK
may be right regarding an infinitely thin baffle with a driver with equal front and rear radiation,
but it's also theoretically possible to cross a continent on a gallon of gas. I want to find out
just how different the real world is.

rick57 · « **Reply #17 on:** 20 Dec 2006, 02:52 am »

mcgsxr

When you suggested “building a multi way open baffle, I would follow the development that Jeffbehr did” – were you referring to http://www.audiocircle.com/index.php?topic=34567.0?

thanks

rick57 · « **Reply #18 on:** 20 Dec 2006, 02:58 am »

Rudolf

Thanks for the further information.

> The remaining on-axis response ripple could then be straightened by moving the driver in the vertical direction.

I wonder if MJK’s worksheet models this?

> After putting my FR driver as near to the upper edge as mechanically possible, things became worse

As I’m likely to have a “top to toe” driver length (15” woofer, 6.5” mod & 8” ribbon in a row) of 30” ~ 75 cm, how much extra should be allowed, say 5” ~ 12.5 cm either end, for a minimum height of 1 metre?

Cheers

rick57 · « **Reply #19 on:** 20 Dec 2006, 03:01 am »

JohninCR

> changing the edge geometry makes a HUGE difference in the sound with the driver very near the edge of a cab, at least for boxes. . For OBs that is something on my experimental measurement list

a case for collecting large pieces of cardboard for mock-up baffles

Cheers