I'm going to try everything that I have at hand, the DFM-2535R00-08 being one of those, so I suppose yes.
Smooth-On’s FOAM-iT!™ Series
Has anyone tried to print petals with a sparse gyroid infill (which is porous all the way through) and then fill with a SmoothOn's FOAM-iT ("Castable Rigid Urethane Foam")?
I'm considering e.g. the 10-SLOW, it seems that these don't get hot during curing, and it's still pretty fast. The petals can be easily made thick and hollow (basically), ideal for filling with something like this, IMO. Otherwise, especially the big ones, tend to resonate. Could this help or is still too light for an effective damping? The type 26 - the most dense/least expanding - is 26 lb/ft³.
I've tried various different things as a filler and I really need someting fast-curing.
Has anyone tried to print petals with a sparse gyroid infill (which is porous all the way through) and then fill with a SmoothOn's FOAM-iT ("Castable Rigid Urethane Foam")?
I'm considering e.g. the 10-SLOW, it seems that these don't get hot during curing, and it's still pretty fast. The petals can be easily made thick and hollow (basically), ideal for filling with something like this, IMO. Otherwise, especially the big ones, tend to resonate. Could this help or is still too light for an effective damping? The type 26 - the most dense/least expanding - is 26 lb/ft³.
I've tried various different things as a filler and I really need someting fast-curing.
Last edited:
I wouldn't use an expanding foam, the pressure it builds up when expanding is significant and it could burst your petals. Also, I believe that filling it with resin is better than foam for structural rigidity.
I'm done printing this small enclosure using 5 wall layers and 5% gyroid infill and will pour Neukadur Multicast 27 mixed with Neukadur Multicast Hardener ISO 5 which is the low-shrinkage and low-exotherm version (but it takes longer to cure than other variants). Will report back - give me a couple of days... My plan is to do the same with one of your designs eventually.
If you want something fast-curing choose a resin with a short de-molding time, e.g., Neukadur MultiCast 7 (120min de-molding time; 3mins pot life, so you have to mx and pour really fast). Probably all manufacturers of PU resin have a product with a short curing time.
I'm done printing this small enclosure using 5 wall layers and 5% gyroid infill and will pour Neukadur Multicast 27 mixed with Neukadur Multicast Hardener ISO 5 which is the low-shrinkage and low-exotherm version (but it takes longer to cure than other variants). Will report back - give me a couple of days... My plan is to do the same with one of your designs eventually.
If you want something fast-curing choose a resin with a short de-molding time, e.g., Neukadur MultiCast 7 (120min de-molding time; 3mins pot life, so you have to mx and pour really fast). Probably all manufacturers of PU resin have a product with a short curing time.
Last edited:
Try it and see if you like it. I think it's fun and satisfying - there is some magic to creating solid objects from soft glass tissue and liquid resin. To start, watch the the Rutan Composite Construction video on youtube (start at minute 25) to get an idea how it works (- there are other tutorials; search for Mike Arnold's work) - the basics of airplane building are the same as for composite speaker building. Also look at what boat builders are doing.
If you want to go "full composite construction" you could print (or CNC) the inner cone and lip shape of the horn (instead of the actual horn), sand/fill/polish/wax it (see the "Epoxy wipe method" here how to make a shiny surface with little work), then build the horn on top of the cone "inside out" by applying gelcoat of your chosen colour and a couple of layers glass/epoxy. Only worth it if you want to pull several horns from the cone mould. Printing and resin filling is probably much faster (and cleaner) for prototyping or building a small number of horns.
Wear gloves, glasses and a respirator when working with resin and when sanding the stuff! I think that you need a dedicated space to do composite work; a garden shed or garage is good enough, but you shouldn't do this in a room in your apartment.
If you want to go "full composite construction" you could print (or CNC) the inner cone and lip shape of the horn (instead of the actual horn), sand/fill/polish/wax it (see the "Epoxy wipe method" here how to make a shiny surface with little work), then build the horn on top of the cone "inside out" by applying gelcoat of your chosen colour and a couple of layers glass/epoxy. Only worth it if you want to pull several horns from the cone mould. Printing and resin filling is probably much faster (and cleaner) for prototyping or building a small number of horns.
Wear gloves, glasses and a respirator when working with resin and when sanding the stuff! I think that you need a dedicated space to do composite work; a garden shed or garage is good enough, but you shouldn't do this in a room in your apartment.
I managed to run the simulation, many thanks to @mabat . However I am getting some surprising results. This is what the conical waveguide looks like in ABEC:
That looks exactly as I intended. What surprised me were the results:
Is this to be expected from this type of horn? Especially the dip and peak below and above 1 kHz on axis. The configuration is attached.
I checked another simulation I made before this one, that one had the roundover smaller and specified with wrong parameters. The above used parameters found via Desmos for the desired roundover shape. Anyway, the result is this:
With a dip and peak even worse. Does this mean that I am more or less on track and by modifying the roundover (making it larger, probably
I can improve the situation further?
This is a FDW gated on axis measurement with reflections in room - of the horn without any roundover - do I see coorectly the pattern there as well?
If yes, I am really amazed and the great ATH will allow me to optimize this construction method to build large round MEHs
That looks exactly as I intended. What surprised me were the results:
Is this to be expected from this type of horn? Especially the dip and peak below and above 1 kHz on axis. The configuration is attached.
I checked another simulation I made before this one, that one had the roundover smaller and specified with wrong parameters. The above used parameters found via Desmos for the desired roundover shape. Anyway, the result is this:
With a dip and peak even worse. Does this mean that I am more or less on track and by modifying the roundover (making it larger, probably
I can improve the situation further?This is a FDW gated on axis measurement with reflections in room - of the horn without any roundover - do I see coorectly the pattern there as well?
If yes, I am really amazed and the great ATH will allow me to optimize this construction method to build large round MEHs
Yeah looks about right, edge diffraction. Form periodicity of the interference you can directly calculate where it happens, how big of a delay there is after direct sound when another sound comes after.
If you want to get rid of the interference altogether, the waveguide curve must be smooth from where it starts to expand all the way around basically avoiding any abrupt changes in curvature. Any "discontinuity", or change in the curve, could send diffraction related backwave, which is basically secondary sound source that emits backwards where the sound came from (simplified) with opposing polarity and limited bandwidth. With such device you've got about maximal diffraction backwave happening at the mouth edge.
The thing is, the simulation uses ideal sound source, and thus the frequency response could be absolutely smooth with no wiggle whatsoever. Any wiggle you see is result of some interference, basically precense of at least one extra sound source, perhaps a reflection or diffraction related backwave that emits when sound bends (diffracts) around a corner.
It's impossible to get rid of wiggle altogether, so all you need to do is relieve it enough so that you don't find it audibly disturbing.
If you want to get rid of the interference altogether, the waveguide curve must be smooth from where it starts to expand all the way around basically avoiding any abrupt changes in curvature. Any "discontinuity", or change in the curve, could send diffraction related backwave, which is basically secondary sound source that emits backwards where the sound came from (simplified) with opposing polarity and limited bandwidth. With such device you've got about maximal diffraction backwave happening at the mouth edge.
The thing is, the simulation uses ideal sound source, and thus the frequency response could be absolutely smooth with no wiggle whatsoever. Any wiggle you see is result of some interference, basically precense of at least one extra sound source, perhaps a reflection or diffraction related backwave that emits when sound bends (diffracts) around a corner.
It's impossible to get rid of wiggle altogether, so all you need to do is relieve it enough so that you don't find it audibly disturbing.
For this construction, I can get smooth transitions both on the throat side and the mouth side as well - albeit with a long flat conical part - it actually only needs to be long enough to fit the drivers.
Maybe it would be possible to use a three part construction (throat adapter - round to octagon, 8 sided conical middle part and octagon to round mouth made from eight parts) to mimic what mabat does with a 3D printer. It could be easier than CNC a fully round horn from slices and then fitting drivers on the outside.
Maybe it would be possible to use a three part construction (throat adapter - round to octagon, 8 sided conical middle part and octagon to round mouth made from eight parts) to mimic what mabat does with a 3D printer. It could be easier than CNC a fully round horn from slices and then fitting drivers on the outside.
Not really - I would have thought opposite - a finer mesh would be better? The dip is almost exactly where predicted, below the driver rolls off, so the peak is not that visible. I think I see some in the mids - and when going off axis, I see similar irregularities as in the simulation. I need to do better measurements some day.
I have a new addition - an analytic approach to segmented horns: http://www.at-horns.eu/seg-horn.html
This is the way all new EXAR waveguides will be designed. The current ones were created with more numerical (brute-force) methods and that was somewhat ugly. This is way more elegant and perhaps even smoother (and much, much easier to work with).
This is the way all new EXAR waveguides will be designed. The current ones were created with more numerical (brute-force) methods and that was somewhat ugly. This is way more elegant and perhaps even smoother (and much, much easier to work with).
Last edited:
The middle part of the R-OSSE horn is pretty flattish - if there is no other way, the middle part petals could be machined to be slightly curved on the inside while keeping the outside flat to be able to mount the MF and LF drivers easily. I need ca 45 cm of flat surface on the outside, the rest can be whatever.