Driver combinations - An Scanspeak Design

[NurEinTier]

Interesting.
Does your wide baffle remark include OB or just BR and sealed?

[HiFiNutNut]

It applies to both.

When we see the drivers on a baffle, we see the direct sound source coming from the drivers, then the indirect sound sources coming from the edges (though at a lower amplitude). This is how we should see diffraction. Wide baffle results in sound sources coming from a wide area. Narrow baffle results in sound sources coming from a narrow area, closer to a point source. We also need to understand that the diffracted sound is delayed sound. So a wide baffle creates more time smearing.

Practically, you can find wide baffle speakers usually give a large sound stage. This is the artifact of diffraction from a wide baffle. The sound is artificially enlarged. Some time smearing would probably mask a lot of hard edges in the music and make the music a bit more smooth. Some people may prefer this type of sound. Narrow speakers, if designed properly, give you good image, pin-point location of musical instruments, the finest grain of texture in human vocals, etc. This is more accurate sound reproduction.

Of course, when we worry about the baffle width we are mainly concerned with frequencies from probably 1k up to about 10k, in a range our ears happen to be most sensitive. Above 10k, the tweeter starts to beam so the effect of diffraction is of no concern.

With OB, narrow baffle is not possible because of the loss in sensitivity. What I do is to use angled side wings to achieve the goal of effective wide baffle, and make the front baffle as narrow as possible with a round edge. Note that the ends of the side wings are also sources of diffraction, but they are secondary so I expect them to be lower in amplitude thus not causing problems. In this case, another set of problem to deal with is how to eliminate the possible resonance of the U-frame. I don't think in a typical small room the back waves of a dipole would give much advantage at all. It probably does more harm than good. So I use damping material to damp my shallow, anged U-frame. Resonance is definitely not a concern. So I can still achieve a narrow baffle with round edges with OB.

Regards,
Bill

[HiFiNutNut]

Hi, Rabbitz,

Reading your enlightening web site I recall you use 3mm thick felt in your speaker, if I have not lost my memory. May I ask where you found them?

Regards,
Bill

[HiFiNutNut]

Nicholas,

Hi, if you indeed proceed with your project, contact me for 4 x 26W8861T00. They are the top of the range, 10" ScanSpeak Revelators. I bought them earlier this year but never come to a chance to use them yet. I now want to get rid of them. You will get them cheaper than you could get elsewhere. They have never been used (actually they have never been unboxed from the boxes supplied by ScanSpeak).

If you use the Revelator tweeter and midwoofer, you may as well use the Revelator woofers for the ultimate sound. Surely they are better than the 25W. Send me a PM if you are interested.

Regards,
Bill

[CeramicMan]

Originally posted by HiFiNutNut
It applies to both.

When we see the drivers on a baffle, we see the direct sound source coming from the drivers, then the indirect sound sources coming from the edges (though at a lower amplitude). This is how we should see diffraction. Wide baffle results in sound sources coming from a wide area. Narrow baffle results in sound sources coming from a narrow area, closer to a point source. We also need to understand that the diffracted sound is delayed sound. So a wide baffle creates more time smearing.

Hi Bill,
Sorry, I think that's a bit of an oversimplification and that you're mistaken with some of your claims...

What you seem to be describing is reflection, re-radiation and possible resonances caused by sharp corners at the edge of the baffle, because they produce sharp discontinuities in the expansion rate of the sound wave emanating from the speaker. The diffraction (allowing the sound to curve behind the speaker) would be the relatively benign part.

One option would be to make the baffle so wide that the sound never reaches the edge, or that it has decayed so much that the effect is negligible when it does. Another option is to make the baffle so narrow that the speaker's whole operating range is below the so-called Baffle Step.

Generally speaking, neither of those two approaches are practical, so I'd say that a reasonable middle-ground would be to employ other techniques such as: curving the baffle to give a smoother transition between the front and sides (a la: horns with non-linear expansion rates), and/or use a thick layer of soft felt/wool/etc around the 'sharpest' areas of the baffle. This reduces the speed of sound near the surface, assisting the sound waves in diffracting around sharp corners.

I'd also keep in mind the expansion rate of the backwave, to make sure that it's not "horn-loaded" so to speak, and won't adversely load the cone with strong box resonances.

Cheers,

[rabbitz]

Originally posted by HiFiNutNut
Hi, Rabbitz,

Reading your enlightening web site I recall you use 3mm thick felt in your speaker, if I have not lost my memory. May I ask where you found them?

Regards,
Bill

Hi Bill

There's the use of 2 different materials plus 80mm damping material in the U frame.

On the baffle proper I'm using a sound deadener that's barium mineral based (WES Code MAT675). The 45° sections from the baffle and the sides of the U frame have felt (about 5mm thick) which is a liner they use under ironing board covers. Most supermarkets or hardware shops usually have it.

I have not tested it without the lining but it can't hurt and anything that stops bouncing waves has to be good. The design was more about stopping the backwave coming back through the driver than an OB, so it's a bit of each. The damping is an important part of the design and very necessary as they sit around 500mm from a back wall. You just get enough of the reflected sound from the back wall to give them a bit of magic.

Cheers

[dlr]

Originally posted by CeramicMan

Hi Bill,
Sorry, I think that's a bit of an oversimplification and that you're mistaken with some of your claims...

I think that he may interpret some of it differently than I might because I believe the wide/narrow debates such as this to be over-simplifications as well, but the basics are correct. The wider baffle would indeed distribute the geometric origin of the diffracted wave and hence it's time delay, frequency impact and I suppose possibly the perception and impact on the imaging.

That said, making a sweeping statement without considering how that can be used to advantage by choice of crossover type, Fc and slope as well as the impact of the dispersion pattern of the chosen drivers to use each driver maximally makes it unclear whether one or the other has an absolute advantage. I don't think that either one does in general, there are too many possible design choices that affect it. Most of the current rage for narrow baffles is more the WAF factor and/or fitting into a home theater setup due to space considerations.

What you seem to be describing is reflection, re-radiation and possible resonances caused by sharp corners at the edge of the baffle, because they produce sharp discontinuities in the expansion rate of the sound wave emanating from the speaker. The diffraction (allowing the sound to curve behind the speaker) would be the relatively benign part.

The term diffraction encompasses the full set of the interactions at a discontinuity, whether baffle edge, adjacent driver cones or anything else. In that latter sentence you're thinking of the step, but that is just one part of diffraction.

One option would be to make the baffle so wide that the sound never reaches the edge, or that it has decayed so much that the effect is negligible when it does. Another option is to make the baffle so narrow that the speaker's whole operating range is below the so-called Baffle Step.

The latter really can't occur for all but true woofers/subwoofers because just the faceplate on a tweeter, a midrange or a midwoofer mounted on a cylinder end will still have diffraction, only a different signature. They will exhibit step. Woofers also have little to no impact on imaging due to diffraction as well.

The other aspect too often overlooked is that diffraction and it's step component rely on the height as well as the width, so even a 10" woofer on a narrow, but tall baffle might still have some step effects. It will also not be operating into 4-pi space as well, since at a woofers frequencies sitting in a room, the floor alone dictates that it's operting into 2-pi space up to some transition frequency.

Generally speaking, neither of those two approaches are practical, so I'd say that a reasonable middle-ground would be to employ other techniques such as: curving the baffle to give a smoother transition between the front and sides (a la: horns with non-linear expansion rates), and/or use a thick layer of soft felt/wool/etc around the 'sharpest' areas of the baffle. This reduces the speed of sound near the surface, assisting the sound waves in diffracting around sharp corners.

I'd be interested to know how you've determined that the diffraction amelioration of felt has to do with speed of sound change alone and not significant absorption and/or changes in the laminar flow in the region of the felt. Do you have references to studies that determined this to be the sole effect? I've not seen any studies that indicate the precise mechanism.

Dave