Actually I have thought about it
. Problem is, without data on either I can only speculate... My speculation more specifically is based on the idea that one has two options - either use an OS WG, this demands a planar wavefront, or a conical horn, this demands a spherical wavefront The point of either is that the side walls be perpendicular to the wavefront I presume.
Compression drivers attempt to supply a planar wavefront, but achieve this imperfectly. Dome drivers are meant for flat baffle, i.e. a 180 degree coverage conical horn. The curved surface even if vibrating axially must to some exent produce a spherical distortion of the wavefront, reduced due to the phase lag of the center of the dome: if I understand Dickason correctly cones are cones in part to give the edge of the cone a phase lead w.r.t the center where the voice coil drives them, similarly domes are domes for the opposite reason, to give a phase lead to the center. So, no real spherical wavefront and no real planar wavefront either for a dome.
The point was rather, is a dome's output a better match for a conical WG than an OS WG - is its actual wavefront closer to the conical's ideal wavefront than the OS's ideal wavefront. And, is the deviation from the ideal in the dome-conical case worse than the deviation from the ideal in the compression driver - OS WG case.
Do you have data on wavefront shape of domes? That would make a good case, unfortunately I don't have any so I can only speculate.
Hello Patrick
"In a nutshell, Danley describes how the drooping frequency response that you see with a compression driver is *preferable* to the flat response you see with a dome tweeter. This behavior has to do with the motor superiority of a compression driver, and it also affect it's phase response."
I always thought it was the mass breakpoint of the diaphram that gave you the 6db per octave roll off.
Rob
"In a nutshell, Danley describes how the drooping frequency response that you see with a compression driver is *preferable* to the flat response you see with a dome tweeter. This behavior has to do with the motor superiority of a compression driver, and it also affect it's phase response."
I always thought it was the mass breakpoint of the diaphram that gave you the 6db per octave roll off.
Rob
gedlee said:
Interesting. Why a conical instead of OS?
I know you don't care for domes, but I appreciate your commenting on them. For now CD's are expensive, and very few perform all that well. And the use of one radically alters and constricts what can be done with a design. By comparison for us diyers and small designers good performing domes are plentiful and we'd just like to harness some of the desirable performance traits of waveguides with these domes we have available.
group buy
Hi Bas, good idea actually!
I can't organize that either - I guess the head quarter needs to be in US anyways , I could function as european h.q. though. But as I kind of brought this up, I been counting too - here's my list.
Members that only expressed interest with no specific numbers got a '?'
edit:
Maybe it's a good idea if that 'central post' would also have the basic info about the g.b. like cost, what you get... Here Earl has posted the actual measurements of the ESP15.
@ Earl: Will the group buy. only be for the Summa / ESP15 waveguide or could we get the the smaller ESP12 and ESP10 wgs too?
best, LC
Hi Bas, good idea actually!
I can't organize that either - I guess the head quarter needs to be in US anyways , I could function as european h.q. though. But as I kind of brought this up, I been counting too - here's my list.
PHP:
USERNAME WGs Location
Jazr ? us
Patrick Bateman ? us
pooge 2 us
Brett 2 au
Pallas 3 eu
Total 7+Members that only expressed interest with no specific numbers got a '?'
edit:
Maybe it's a good idea if that 'central post' would also have the basic info about the g.b. like cost, what you get... Here Earl has posted the actual measurements of the ESP15.
@ Earl: Will the group buy. only be for the Summa / ESP15 waveguide or could we get the the smaller ESP12 and ESP10 wgs too?
best, LC
MBK said:
Actually I have thought about it
My speculation more specifically is based on the idea that one has two options - either use an OS WG, this demands a planar wavefront, or a conical horn, this demands a spherical wavefront The point of either is that the side walls be perpendicular to the wavefront I presume.
Compression drivers attempt to supply a planar wavefront, but achieve this imperfectly. Dome drivers are meant for flat baffle, i.e. a 180 degree coverage conical horn. The curved surface even if vibrating axially must to some exent produce a spherical distortion of the wavefront, reduced due to the phase lag of the center of the dome: if I understand Dickason correctly cones are cones in part to give the edge of the cone a phase lead w.r.t the center where the voice coil drives them, similarly domes are domes for the opposite reason, to give a phase lead to the center. So, no real spherical wavefront and no real planar wavefront either for a dome.
The point was rather, is a dome's output a better match for a conical WG than an OS WG - is its actual wavefront closer to the conical's ideal wavefront than the OS's ideal wavefront. And, is the deviation from the ideal in the dome-conical case worse than the deviation from the ideal in the compression driver - OS WG case.
Do you have data on wavefront shape of domes? That would make a good case, unfortunately I don't have any so I can only speculate.
Good questions.
The dome (normal and inverted) radiation was handled for a flat baffle by Suzuki in the early 80's in JASA. On a sphere this problem is done by Skudryk in his Acoustics book, and maybe in my book too (I don't remember).
Now: "is a dome's output a better match for a conical WG than an OS WG " I would say yes, but there are other variables that complicate things (see below).
And :"is the deviation from the ideal in the dome-conical case worse than the deviation from the ideal in the compression driver - OS WG case." This is absolutely the case since the ideal compression driver is a perfect match for an OS waveguide while even in the ideal case the dome and conical don't match.
BUT: More importantly, in the "typical" case the compression driver is a very good match to a well matched waveguide - nearly ideal, but the dome, in practice, is a poorer match than the ideal because of the need to accomodate the surround. One cannot make the waveguide go all the way down to the domes surface and this gap at the edge is a major problem.
You guys keep arguing for domes and I'm telling you they just don't win in any category but cheap. I have never seen the point of doing DIY "on the cheap" - I suppose my time is worth too much to me. You can save a lot of money sure, but why cut every corner? Cut the ones that don't affect performance so much and spend some money on those aspects that dominate the performance. Thats my design philosphy and it works well. Summas can be made for about $500 each if you cut the right corners yourself! Try and buy a speaker of that quality for that price.
I do agree completely with Geddes on the dome versus compression drivers in waveguides. It seems that a lot of people are afraid of using compression drivers because all the major high end manufactureres are using domes so there must be a good reason for that they might think, but are they all using waveguides ? No, don't think so.
See it as if you really like transmissionline or backloaded horn speakers and you have to choose a proper woofer for it. Would you go for the same driver in both applications ? I guess not.
Another misunderstanding is that using dome tweeters is cheaper. I think it's most often the other way around. When I compare my Focal Audiom TLR tweeters with my BMS4540ND waveguides the latter has much better performance at even one third of the price. The only concern is that you have to take more care about the crossover. Plugging the BMS in as a straight replacment of the Focals will let your ears bleed and yourself running for 100 miles away from them as fast as you can. With the filtering done right everything smoothens out.
See it as if you really like transmissionline or backloaded horn speakers and you have to choose a proper woofer for it. Would you go for the same driver in both applications ? I guess not.
Another misunderstanding is that using dome tweeters is cheaper. I think it's most often the other way around. When I compare my Focal Audiom TLR tweeters with my BMS4540ND waveguides the latter has much better performance at even one third of the price. The only concern is that you have to take more care about the crossover. Plugging the BMS in as a straight replacment of the Focals will let your ears bleed and yourself running for 100 miles away from them as fast as you can. With the filtering done right everything smoothens out.
Sjef said:
This is indeed correct and a major issue. Waveguides and ANYTHING cannot be used without EQ. But if you are going to do the EQ then do it for a compression driver NOT a dome. If you can't do a good crossover design (and that ain't easy!) then stay away from waveguides, you'll only be disappointed.
Good point. Fair enough also re: ultimate quality with a compression driver. I am not really arguing for domes, I somewhat academically just wanted to know. Plus, due to the inherent compression, compression drivers do also have disadvantages as disussed above. In any case I still want to use a compression driver in a WG in the next upgrade of my system...
Something else occurred to me re: HOMs and OS / conical WGs. How is sound reflected off an angled surface? Does the optics law of entrance angle to reflecting surface = exit angle, apply to a sound wave in the typical WG case where wavelength >> WG internal dimension?
The point where I am getting at is, if so, then geometrically any waveguide with an angle to the propagation axis of >45 degrees (total coverage > 90 degrees) can not produce more than one internal reflection: The worst case internal WG wave propagation angle is 90 degrees relative to propagation axis. Hitting a 45 or more degrees WG wall will then deflect the wave at worst to an angle of zero, i.e., parallel to the axis of propagation. In other words this wave then has no chance of hitting any internal WG surfaces another time.
If the above is correct then for coverages above 90 degrees, axisymmetric conical WGs can not produce HOMs.
Am I missing something?
"If the above is correct then for coverages above 90 degrees, axisymmetric conical WGs can not produce HOMs. "
Earl will correct me if I'm mistaken, but I had envisioned HOM's spraying in all directions (or at least in many "wrong" ones) from the throat area.
These will strike the WG walls at oblique angles ionstead of traveling along them, reflecting back and forth until they exit.
Their longer path lengths is why the foam absorbs them more than the "good" waves.
Earl will correct me if I'm mistaken, but I had envisioned HOM's spraying in all directions (or at least in many "wrong" ones) from the throat area.
These will strike the WG walls at oblique angles ionstead of traveling along them, reflecting back and forth until they exit.
Their longer path lengths is why the foam absorbs them more than the "good" waves.
XT1086 shape
I made a mold of the interior of the XT1086 with expanding urethane foam so that I could slice it up into 1/4” (6.35 mm) thick pieces for patterns to make foam plugs.
It turns out Vaseline is not an adequate mold release, so I’ll have to do it again, but now we can see the shape better.
In the vertical (narrower) axis the shape looks very nice, but in the horizontal it has a definite kink.
Also the Vaseline makes the edge more apparent in the picture looking into the horn.
Earl, or anyone else who feels qualified to comment, what’s the worst that could happen if that edge is smoothed over? BTW the wall thickness is about 4 mm and stiffening ribs run right over that area, so I think it’s safe to do.
Will it have an effect on the coverage angle? I’d actually like narrower horizontal coverage because the L and R speakers have very different sidewall distances.
I’d actually considered rotating the horns 90 deg to have 60° x 80° instead of 80° x 60° H x V coverage, but I’m limited in the vertical dimension of my boxes.
Vertical axis:
I made a mold of the interior of the XT1086 with expanding urethane foam so that I could slice it up into 1/4” (6.35 mm) thick pieces for patterns to make foam plugs.
It turns out Vaseline is not an adequate mold release, so I’ll have to do it again, but now we can see the shape better.
In the vertical (narrower) axis the shape looks very nice, but in the horizontal it has a definite kink.
Also the Vaseline makes the edge more apparent in the picture looking into the horn.
Earl, or anyone else who feels qualified to comment, what’s the worst that could happen if that edge is smoothed over? BTW the wall thickness is about 4 mm and stiffening ribs run right over that area, so I think it’s safe to do.
Will it have an effect on the coverage angle? I’d actually like narrower horizontal coverage because the L and R speakers have very different sidewall distances.
I’d actually considered rotating the horns 90 deg to have 60° x 80° instead of 80° x 60° H x V coverage, but I’m limited in the vertical dimension of my boxes.
Vertical axis:
Attachments
"what’s the worst that could happen if that edge is smoothed over?"
Nothing bad. You'd have "less" diffraction because the corner wouldn't be as sharp. Don't know if the difference would be audible.
They cost very much?
Foam. I got some pretty thick pieces (@30 ppi, reticulated) from a foam shop. He sells it to pond owners for filtration.
Nothing bad. You'd have "less" diffraction because the corner wouldn't be as sharp. Don't know if the difference would be audible.
They cost very much?
Foam. I got some pretty thick pieces (@30 ppi, reticulated) from a foam shop. He sells it to pond owners for filtration.
But how do you shape a big block?
I was thinking you could make a pantograph setup to follow a mold on a turntable, using a router with a convex abrasive cutter, but that would get to be a pretty good project on its own.
I think the layers would work pretty good.
The critical area is near the throat, where the wall angle is small and missing material from the "stairstepping" would be minimal.
Also, the HOM waves are not following the wall, so that material won't be missed.
Earl, does that sound right?
Thanks
I was thinking you could make a pantograph setup to follow a mold on a turntable, using a router with a convex abrasive cutter, but that would get to be a pretty good project on its own.
I think the layers would work pretty good.
The critical area is near the throat, where the wall angle is small and missing material from the "stairstepping" would be minimal.
Also, the HOM waves are not following the wall, so that material won't be missed.
Earl, does that sound right?
Thanks
gedlee said:
Hi Earl,
well, your very own listening tests OS/CD vs horn/dome provides a compelling argument for domes as long as very high SPL requirements are not required. The direct radiator horn loaded dome of the Gradient didn't seem to put it at any disadvantage to the OS/CD Summa, at least in these tests. I concede that a CD is capable of much greater output, but I don't think that is what is being discussed here.
Perhaps further tests such as these (along with extensive measurements) are called for.
cheers,
AJ
But how do you shape a big block?
Sorry, I was in hurry. The pieces I got were 1" thick. I had a square cross section and drawings of verticle and horizontal profiles so it was easy to cut sections with an electric meat carving knife ($8 Walmart). - Some sort of reciprocating or jig saw might be better - Then I just put them in.
These were old exponential wood horns we made years ago, and the rough surfaces just hold everything together and inside. I didn't have to compress anything. These are kind of large - mouth is 5 X 22" and I couldn't face using 1/4' thickness.
If you produced a clean enough molded profile you could cut it in 1" sections, (even make paper cross section patterns from them), and they probably wouldn'be too hard to copy.
I don't think it matters much if the final result isn't perfect. I think I remember Earl Geddes writing somewhere that when he was doing this he just cut a foam cone and put it in. But we might ask Earl about that - memory plays tricks!
If the incredibly good industrial rent everything company we have in town has foam cutters, I'll rent one next time.
The foam shop guy had this incredibly long tiny-toothed saw held horizontal in a graduated jig and he cut me the 1" thicknesses off a bigger block. For the next project I'll take the waveguide in, let him put his hands on it, and get get his advice on how to do it.
Sorry, I was in hurry. The pieces I got were 1" thick. I had a square cross section and drawings of verticle and horizontal profiles so it was easy to cut sections with an electric meat carving knife ($8 Walmart). - Some sort of reciprocating or jig saw might be better - Then I just put them in.
These were old exponential wood horns we made years ago, and the rough surfaces just hold everything together and inside. I didn't have to compress anything. These are kind of large - mouth is 5 X 22" and I couldn't face using 1/4' thickness.
If you produced a clean enough molded profile you could cut it in 1" sections, (even make paper cross section patterns from them), and they probably wouldn'be too hard to copy.
I don't think it matters much if the final result isn't perfect. I think I remember Earl Geddes writing somewhere that when he was doing this he just cut a foam cone and put it in. But we might ask Earl about that - memory plays tricks!
If the incredibly good industrial rent everything company we have in town has foam cutters, I'll rent one next time.
The foam shop guy had this incredibly long tiny-toothed saw held horizontal in a graduated jig and he cut me the 1" thicknesses off a bigger block. For the next project I'll take the waveguide in, let him put his hands on it, and get get his advice on how to do it.
Hello
You may like to know that Seas are about to bring out a range of tweeters using " waveguides " developed with " www.mike-thomas.com/products.asp ". If you click on DXT you will see their guides.
I had an email from them saying they hope to have a DXT website by the end of August. This will provide information for the Diy market on using the DXT tweeters with other speakers. But it appears that the guides will not be available as a bolt-on modification as on the Monacor DT300, shame ! Still you can always guarantee that Seas will will come up with excellant tweeters
Roy
You may like to know that Seas are about to bring out a range of tweeters using " waveguides " developed with " www.mike-thomas.com/products.asp ". If you click on DXT you will see their guides.
I had an email from them saying they hope to have a DXT website by the end of August. This will provide information for the Diy market on using the DXT tweeters with other speakers. But it appears that the guides will not be available as a bolt-on modification as on the Monacor DT300, shame ! Still you can always guarantee that Seas will will come up with excellant tweeters
Roy
First off, the "home depot" foam insulation cans are generally closed-cell foam to provide ... well, insulation (i.e. prevent the movement of air). I'm not sure they'd work so well as a transmissive phase plug.
You can probably search around to find open cell versions online...
Second, you'll need a urethane compatible mold release. there are several spray-on products available from anywhere that sells molding products. not expensive.
It would pay to make a casting of your WG then make a mold off of that so you dont make a mess of the WG itself.
Matt
You can probably search around to find open cell versions online...
Second, you'll need a urethane compatible mold release. there are several spray-on products available from anywhere that sells molding products. not expensive.
It would pay to make a casting of your WG then make a mold off of that so you dont make a mess of the WG itself.
Matt