Moving this from the thread on DIY spherical speakers...
For those who haven't seen the previous posts, I am looking at using a 3D printer to produce a "spherical" ESL with 12 drivers about 200 mm in diameter each arranged on the surface of a truncated dodecahedron. The printed parts will serve as structural elements as well as the stator supports for wire screen stators unless I can find a cheap source of perforated sheet metal discs. Diaphragms will be 6 um polyester with Licron coating.
The diaphragm side of the printed parts will print on the printer's bed so they will be smooth and flat. My printer has a 300x300 mm cast aluminum tooling plate bed with a layer of PEI. I am planning to use ABS to print the actual speaker parts and will solvent weld them together using acetone. I have printed some 100 mm stator pieces and a gluing jig to test the accuracy of the printer.
Today I redesigned and printed the gluing jig for the smaller test build before I commit to the full size speaker build, and welded one hemisphere together. I used ethyl acetate (sold as MEK substitute) to solvent weld the 3D printed PETG parts together. As I was gluing it occurred to me that the jig may not be necessary- strategically placed binder clips might do the job just fine. It takes just a tiny drop of the solvent and the joint is secure in a less than a minute.
When I set the glued up assembly down on a flat table all the surfaces on the bottom of the assembly touch the table so I think the printer is producing the prints with accurate angles. The surfaces that are welded looked like they were parallel to each other. I will print the other 6 individual stator pieces and glue them together and see how the two assemblies line up. Then it will be on to more of the design work for the actual drivers.
Note- the print quality is less than ideal- I have a 1mm nozzle on the machine so retraction doesn't work worth a damn resulting in blobby defects. I'll probably switch to a smaller nozzle where I can use retraction effectively for the final prints.
This is a truncated dodecahedron:
Three of the outer stators pieces (red) in the gluing jig (blue). The orange piece is one of the inner stator pieces:
Here is a hemisphere made from the small, test-sized outer stator pieces:
It will be interesting to see if the diaphragm tension causes the stator pieces to flex out of shape. I'll probably add some supports for the wire screen to the inner and outer stator pieces, as well as some means to add dust covers to each driver using the same 6um film as the diaphragms. I still have to work out details of the electrical connections.
For those who haven't seen the previous posts, I am looking at using a 3D printer to produce a "spherical" ESL with 12 drivers about 200 mm in diameter each arranged on the surface of a truncated dodecahedron. The printed parts will serve as structural elements as well as the stator supports for wire screen stators unless I can find a cheap source of perforated sheet metal discs. Diaphragms will be 6 um polyester with Licron coating.
The diaphragm side of the printed parts will print on the printer's bed so they will be smooth and flat. My printer has a 300x300 mm cast aluminum tooling plate bed with a layer of PEI. I am planning to use ABS to print the actual speaker parts and will solvent weld them together using acetone. I have printed some 100 mm stator pieces and a gluing jig to test the accuracy of the printer.
Today I redesigned and printed the gluing jig for the smaller test build before I commit to the full size speaker build, and welded one hemisphere together. I used ethyl acetate (sold as MEK substitute) to solvent weld the 3D printed PETG parts together. As I was gluing it occurred to me that the jig may not be necessary- strategically placed binder clips might do the job just fine. It takes just a tiny drop of the solvent and the joint is secure in a less than a minute.
When I set the glued up assembly down on a flat table all the surfaces on the bottom of the assembly touch the table so I think the printer is producing the prints with accurate angles. The surfaces that are welded looked like they were parallel to each other. I will print the other 6 individual stator pieces and glue them together and see how the two assemblies line up. Then it will be on to more of the design work for the actual drivers.
Note- the print quality is less than ideal- I have a 1mm nozzle on the machine so retraction doesn't work worth a damn resulting in blobby defects. I'll probably switch to a smaller nozzle where I can use retraction effectively for the final prints.
This is a truncated dodecahedron:
Three of the outer stators pieces (red) in the gluing jig (blue). The orange piece is one of the inner stator pieces:
Here is a hemisphere made from the small, test-sized outer stator pieces:
It will be interesting to see if the diaphragm tension causes the stator pieces to flex out of shape. I'll probably add some supports for the wire screen to the inner and outer stator pieces, as well as some means to add dust covers to each driver using the same 6um film as the diaphragms. I still have to work out details of the electrical connections.
Printed and welded the other hemisphere:
And do they fit together? Like chocolate and peanut butter!
With a sawbuck for scale (sorry, I have no bananas):
I'd say it's a go. I'll start work on the "final" driver design including screen support, electrical connections, and dust covers. I'll also design and print the triangular pieces that will fit between the drivers to close up the sphere. I'm hoping to make this a sealed enclosure but it might be hard given all the seams and potential leak points. I might have to add some silicone to the joints (or use some thick construction adhesive when joining the pieces) to seal it completely.
These prints used a total of 160g of filament. I printed 6 of the stator frames at a time in two batches. The frames are 100mm across the edges. After gluing these hemispheres I'd say the gluing jig is a good idea. Using binder clips only would require very careful placement and I'm not sure you'd be able to get the same results.
I don't have a bunch of test equipment to measure frequency response, etc., so the ultimate evaluation is all going to be done by my not so golden, 65 year old ears. I will post links to stl files for printed parts in case anyone wants to try to duplicate or improve on my results.
And do they fit together? Like chocolate and peanut butter!
With a sawbuck for scale (sorry, I have no bananas):
I'd say it's a go. I'll start work on the "final" driver design including screen support, electrical connections, and dust covers. I'll also design and print the triangular pieces that will fit between the drivers to close up the sphere. I'm hoping to make this a sealed enclosure but it might be hard given all the seams and potential leak points. I might have to add some silicone to the joints (or use some thick construction adhesive when joining the pieces) to seal it completely.
These prints used a total of 160g of filament. I printed 6 of the stator frames at a time in two batches. The frames are 100mm across the edges. After gluing these hemispheres I'd say the gluing jig is a good idea. Using binder clips only would require very careful placement and I'm not sure you'd be able to get the same results.
I don't have a bunch of test equipment to measure frequency response, etc., so the ultimate evaluation is all going to be done by my not so golden, 65 year old ears. I will post links to stl files for printed parts in case anyone wants to try to duplicate or improve on my results.
The idea is great, but keep in mind esl panels are not very suitable for use in a sealed enclosure. They are used in OB for a reason.
Have you ever tried it? Have you ever heard it? Neither have I...
This is the idea. The tube will carry the wires up to the speaker from the transformers and bias supply located in the box. The angles that the drivers make on the surface of the "sphere" will require some specific height based on the listening distance because they are angled slightly downward. At 10' the driver facing the listener will have to be about 1.9' above the ear which will probably put it close to midway between the floor and ceiling in most rooms.
Even if the sealed sphere concept proves less than ideal, using a 3D printer to make the drivers will yield some useful results for making open baffle speakers.
This is the idea. The tube will carry the wires up to the speaker from the transformers and bias supply located in the box. The angles that the drivers make on the surface of the "sphere" will require some specific height based on the listening distance because they are angled slightly downward. At 10' the driver facing the listener will have to be about 1.9' above the ear which will probably put it close to midway between the floor and ceiling in most rooms.
Even if the sealed sphere concept proves less than ideal, using a 3D printer to make the drivers will yield some useful results for making open baffle speakers.
Last edited:
An ESL diaphragm is almost acoustically transparent and lacking any stiffness, so I suspect this would cause problems with enclosure pressure and reflections in a sealed enclosure. I'd love it to work though!
(And no, I've not tried it...).
(And no, I've not tried it...).
I tried once yes. And abandoned the idea straight away. Not only do you loose the advantage of distortion free uncoloured sound reproduction that an esl can offer, but it will also raise the frequency of the fundamental resonance of the panel so much that it will only be usefull as a tweeter at best. Unless you use huge enclosures, I mean really huge. Vas for an esl panel is simply massive.
https://www.diyaudio.com/community/threads/cabinet-for-electrostatic-mini-panel.296880/
https://www.diyaudio.com/community/threads/cabinet-for-electrostatic-mini-panel.296880/
why not take a polystyrol ball and some dml driver making the ball emanate sound in all directions.
Some damping foam inside wich does not touch the walls.
Sound exciter could be outside of the ball on top.
Some damping foam inside wich does not touch the walls.
Sound exciter could be outside of the ball on top.
I would try acoustical foam also. I agree that the absorption inside will be very important, but haven't tried this exact approach, so can't say what the best absorber configuration is. The lower the frequency, the tougher it gets to absorb effectively.
Even if the device only works down to a few hundred hertz, it could still be interesting.
One of those drivers attached to a foam ball will not make the ball expand and contract as a spherical source- the waves will radiate from the driver along the surface, and maybe through the ball. Try it and let us know if it works.
I'm specifically interested in 3D printing ESL drivers and the spherical idea is almost secondary to that.
I will try with and without stuffing of some sort.
Hello Mark,
I like your idea. Whatever the result, you will be furthering your knowledge and sharing it, so that's a winner.
Actually, I'm responding for another reason:
I had not seen you on this forum before, and seeing this post is my opportunity to say hello.
Your original ESL website was a great inspiration for me when building my first ESL in 2008 (I shamelessly copied your bike tube tensioning jig).
Afterward I continued checking your older website but you didn't update it for many years so I figured you had lost interest or moved on to other things. I later found your Tech Topics page and added a link to it on my website.
Glad to see you are still around and building ESLs.
Keep up the good work!
Charlie
My website: https://jazzman-esl-page.blogspot.com
I was reading a post by a guy who was printing a transmission line enclosure and it occurred to me, why not print the stuffing too.
just a thought.
just a thought.
Hi ... inspiring to read about your idea and approach .. Made me think of these:
https://www.muraudio.com/px2
They are omnidirectional although the shape of the electrostatic panel is different. Apparently the crossover frequency is 450 Hz.
Cheers, Jesper
https://www.muraudio.com/px2
They are omnidirectional although the shape of the electrostatic panel is different. Apparently the crossover frequency is 450 Hz.
Cheers, Jesper
4, 6, or 8 seems like the angles between the drivers would result in very "beamy" high frequencies. 20, while closer to ideal, is too many to make, and would require them to be small in a reasonably sized speaker. 12 is a compromise between all those problems. The resulting speaker is a reasonable size and the number of drivers is not overly burdensome to make (12 is going to be enough trouble). Each driver has parasitic capacitance in the form of the portion of the stator plates that are glued to the supports. The more drivers you use, the greater that non-working capacitance.
Hey CharlieM, I've been to your site many times. You do some great work. I recently retired and have more time now so I'm getting back into projects. I still have about 1000m of 6um polyester, so there's bound to be a few ESLs in coming months.
I've seen prints of lions that have furry manes made by exploiting the stringiness of some types of filament but it's so much cheaper, faster, and easier to buy a bag of polyfill that it's pointless to try to print it. Especially 60 liters of it. The tapered decahedron I am printing for the drivers would be very hard to make by any other means, so I think it's a better use of the printer.
Last edited:
Have you considered making it roughly umbrella shaped instead? It'd be an open design which could be made to mimic a point source, since Quad have already demonstrated this. Also, every panel would be identical which would be a bonus, production wise.
I have about 20 years of rapid prototyping experience and one concern you'll have to either reconcile or become comfortable with is the longevity and stability of the materials being used. Not all materials are UV stable. Some will become brittle over time and some will not appreciate repeated heat cycling.
Good luck with an interesting project.
kind regards
Marek
I have about 20 years of rapid prototyping experience and one concern you'll have to either reconcile or become comfortable with is the longevity and stability of the materials being used. Not all materials are UV stable. Some will become brittle over time and some will not appreciate repeated heat cycling.
Good luck with an interesting project.
kind regards
Marek