karatestu wrote: ↑Fri Jun 25, 2021 9:09 am
How to make an accurate wooden sphere with a router. No need to buy a wood lathe.
Yes, but, the inside won't be spherical and you thus retain some of the apparent complexities of flat internal surfaces.
You found the flaw in this Geoff but being an engineer you were bound to.
The only way to do it is to make two hemispheres and hollow out the centre of each like when making a bowl. Then glue them together and fit in the above contraption
Ant would say just use a lathe. For a start I don't have one and secondly I haven't used one since high school and I don't remember being that good at it then.
I'm not sure the internal profile would be all that important: it's the reflections on the outside you're worried about right?
Just think how gnarly the insides of a Doc-Modded speaker are after random plates of steel have been slapped all over the place!
These users thanked the author CycleCoach for the post:
I think I might be in the HiFi trade.
I am not currently seeking treatment for this.
Current System: Linn LP12 Lingo Ekos Troika, NVA PHONO2 with BBPSU, NVA AP20 (White), Quad ESL 57s &/or NVA Cube3s (White)
CycleCoach wrote: ↑Fri Jun 25, 2021 11:41 am
I'm not sure the internal profile would be all that important: it's the reflections on the outside you're worried about right?
Just think how gnarly the insides of a Doc-Modded speaker are after random plates of steel have been slapped all over the place!
The inside isn't as important as the outside however I would like to have it round if possible but it's not a deal breaker. It won't be a sphere inside anyway with the driver poking in to it at either end.
These users thanked the author karatestu for the post:
Some quote I found about diffraction, imaging blah blah
"Imo the Precedence Effect explains many observations made about correlations between imaging precision and baffle width, in particular those posted by @shadorne on the previous page.
The Precedence Effect kicks in at about .68 milliseconds, which corresponds to the time it takes for a sound wave to travel about 9 inches, which in turn correlates with the distance around the head from one ear to the other. We get our primary directional cues from that first .68 milliseconds, after which the ear/brain system suppresses directional cues from reflections.
So any reflections occurring within those first .68 milliseconds will degrade imaging, and in general the closer to that .68 millisecond threshold, the worse the degradation. This is because as we approach .68 milliseconds, we are approaching the interaural time delay that would correspond to a sound coming from one side or the other.
So, the wider the baffle the worse the imaging (ignoring the potential improvement from round-overs) until the baffle edge is more than 9 inches away from the driver's edge. Then imaging abruptly improves.
A round-over can mitigate edge diffraction, but it must have a large enough radius to be effective, at least 1/2 wavelength and preferably 1 wavelength. The ear is most sensitive to diffraction at about 4 kHz, and at 4 kHz a wavelength is about 3.4 inches. That's a pretty big round-over! Point being, small round-overs probably don't do much.
Driver directivity can reduce the amount of energy that sees the cabinet edge in the first place, but care must be taken to not do more harm than good in the pursuit of increased directivity. "
So if we are to believe in the Precedence Effect then we need to minimise edges within 9" (22.5cm) of a drivers edge- hint = spherical) So that rules out most speakers. Tweeters which are most important in this respect are usually crammed in right at the top of a baffle with three 90 degrees edges very close by and two of them (maybe even three) at the same distance from said tweeter.
I am not fooling myself here, I have heard a quite dramatic improvement in SQ when removing my tweeter face plates and taping over any screw heads and holes. The face plates were the source of two 2.5mm edges and eight screw holes and screw heads.
I also accept that I will have changed the directivity of the tweeters by doing this as the first edge which was only 5mm from the periphery of the dome will have been acting like a very small waveguide and a very crap one at that. Altering the directivity has produced no bad effects. The whole effect of doing this is very addictive. As Peter said, it's all good - extra detail, smoother, less fatiguing and the sounstage entices and welcomes you in ( whoops a bit of reviewer speak there).
I am currently trying to devise a way of mounting the tweeters with no face plate at all. The plastic disc that was underneath that now forms the face plate will need leveling with filler, sanding and then painting to make it presentable. That will reduce the size a little bit more and make the angle where tweeter meets sphere even smoother. The only other way to do that is to make the tweeter sphere bigger but I don't want to do that as it pushes the four tweeter domes further apart and I don't want that. The proportions of 35cm & 15cm spheres together looks about right to my eyes.
Gosh what a long thread. It's going to take a while it catch up, but in the meantime...
Objects much smaller than the wavelength are ignored by a propagating wave, immovable objects much larger than the wavelength cause a simple reflection. It's where the object size is similar to the wavelength that funky stuff happens. Now an outside corner is small compared to most audio wavelengths and typically causes re-radiation (NOT diffraction). Diffraction happens when the corner size is similar to the wavelength and causes the wave propagation to bend around the corner smoothly. A roundover aids diffraction and minimises re-radiation. Re-radiation causes wave interference with the original source and thus amplitude ripples. The haas effect prevents us from hearing this re-radiation as a separate event. Wilson audio use surface mount felt to absorb waves propagating along the front of a loudspeaker, as does a few BBC speaker, Spica and other spring to mind too. Of course nothing is 100% effective and you are best spreading out the edge distances, so the tweeter should be 1 arbitrary unit the closest edge, 1.61 units from the next edge, 1.61^2 from the next .... This spreads out the interference 'eigenvalues', minimising ripple.
Is this any help?
Cheers, Rex
These users thanked the author r3xj0hn570n for the post:
CycleCoach wrote: ↑Fri Jun 25, 2021 11:41 am
I'm not sure the internal profile would be all that important: it's the reflections on the outside you're worried about right?
Just think how gnarly the insides of a Doc-Modded speaker are after random plates of steel have been slapped all over the place!
You may be right but, if not there is nothing that can be done once the panels have been mad into an external sphere. Acrulic spheres are the best Stu can get and I think that is probably right. A solid sphere with tweeters set into it might work though.
Not sure about the precision