Board logo

Homegrown Speakers
T66 - 8/2/10 at 05:50 PM

Cool


http://englishrussia.com/?p=2133#more-2133

smart51 - 8/2/10 at 05:54 PM

Cool in an industrial loft conversion but I wouldn't expect them to sound good. Cylindrical is the second worst shape after spherical for accoustics and steel would ring too much I'd have thought. Still cool though.

Confused but excited. - 10/2/10 at 08:17 PM

Cylindrical is good as there are no opposing flat surfaces to reflect sound internally. Should not ring if the wadding shown is fitted correctly.
They are not exactly high end or high power after all and look quite good if you are into the 'themed' look.

MikeRJ - 10/2/10 at 08:50 PM

quote:
Originally posted by Confused but excited.
Cylindrical is good as there are no opposing flat surfaces to reflect sound internally.


I was thinking the exact same thing; anything with parallel sides (i.e. most speaker designs) is sub-optimal since reflections and standing waves can be formed. The upside is that a box design is far, far easier to build which is why the vast majority of speakers are designed like this. Unless you want to pay big money of course.

Thin steel is certainly not a great construction, but being a cylinder it will at least resist deforming under pressure variations and with sufficient damping material it could work reasonably well. It doesn't look like any proper design work has gone into this particular project however, just a couple of random car driver units cobbled into a fire extinguisher. Just noticed the amplifier is running off an old PC power supply as well

I've seen a some high end speaker designs based on big concrete pipes which work well. Concrete is a good material for this as it has plenty of internal damping.


[Edited on 10/2/10 by MikeRJ]

smart51 - 10/2/10 at 10:03 PM

But every point has a parallel surface and they're all the same distance apart. The standing waves you get in the enclosure will all be at the same frequencies. The effect will be a bit like a comb filter. I'd expect it to ring badly.

Simon - 10/2/10 at 11:51 PM

I think as a homemade fashion statement/work of art/working decoration, they're quite effective.

As for sound quality, I don't suppose most people would notice the difference or for that matter, be bothered.

ATB

Simon

MikeRJ - 11/2/10 at 12:11 AM

quote:
Originally posted by smart51
But every point has a parallel surface


By definition there are no parallel surfaces in section.

Have a read of this, and you can see that having a speaker mounted in the side of a cylinder is somewhat better than a cuboid, but not as good as a sphere in terms of flatness. A cylinder with a speaker mounted at one end is (unsurprisingly) quite poor.

smart51 - 11/2/10 at 08:09 AM

quote:
Originally posted by MikeRJ
quote:
Originally posted by smart51
But every point has a parallel surface


there are no parallel surfaces


A cylinder with a speaker mounted at one end


At any point on a cylinder, the opposite wall is, for a very short distance, parallel. Move round by 1° and there is a short parallel opposite wall there too - and all the way round. The point is that no matter what angle you look at it from, all the points on a circle have a parallel point that is exactly the same distance away as every other point. A rectangular cross section speaker, or better still a trapezoidal cross section does not.

Can you say that there are no parallels on a circle, then claim that a sphere has an end?

MikeRJ - 11/2/10 at 10:52 AM

quote:
Originally posted by smart51

Can you say that there are no parallels on a circle, then claim that a sphere has an end?


I didn't claim a sphere had an end, I mentioned having the speaker mounted at the end of a cylinder didn't give good results.

The point with a curved surface is that it doesn't provide a reflection in a single direction, which is a cause of standing waves. If you consider a circle to be made up of many short straights then yes, you could consider that there are parallel sections, but the energy reflected from any particular point will be far smaller than a wide flat plate.

Not sure why I'm arguing to be honest, Mr Olson has done all the work and proved at the cylinder gives a flatter response than a cuboid enclosure. In particular he shows the baffle that the speaker is mounted in is one of the more important factors, and flat was the worst by quite a margin.