3D printing 1/2 of a waveguide

All I've ever used is PLA.

I print at 205C and my bed is at 60C.

I do my first layer at 70C.

For me, here's the things that helped me improve my prints. It was a LOT of trial and error:

1) I bought a whole new printer. Made a huge difference.
2) My new printer has a heated bed and the old one didn't.
3) Cooling makes a HUGE difference. Adding a cooling fan on my old printer made a huge difference, and my new printer has two.
4) On my old printer I upgraded the "hot end." That helped. As I understand it, the older hot end designs radiated too much heat, and this leads to print curling and jamming in the print head.
5) replacing your nozzles periodically helps.

I haven't noticed any significant difference in curling when using various print head sizes. The use of a larger print head will shorten print times dramatically, and make the parts stronger. But it also uses up a lot more filament. For instance, if you go from a print head that's .3mm to a print head that's .6mm, you're basically using four times as much filament (because the diameter is double) but you can also print in half the time (because you can double the height of each layer.)

It's pretty astounding how strong the parts are when you print with a .7mm head, I've been seriously considering a 3D printed bicycle! I have 3D parts that I can easily stand on without deformation, and I weigh 250lbs!

I think it would be particularly interesting to take this to the 'next level' by figuring out how to do 'honeycomb' type structures that would be impractical to make with conventional manufacturing technologies.

There's also carbon fiber infused filaments out there, but I've read that they destroy your printer in a hurry.
 
I'm using a 1mm nozzle (and a 40W hot end called a "Volcano"!) because I don't have the patience needed to print something this large with a more normal nozzle. Detail suffers (and talk about a lot of trial-and-error getting it to print even successfully, forget pretty), but that's not much of an issue for something like these big waveguides.

Yeah, strength is not something I'm worried about on this, the printed parts with a 1mm nozzle are really tough. I used two outer perimeter walls on the prints shown, I may try one with a single wall, it should still be strong enough and print significantly faster. The slicer I'm using ("Slic3r", a freeware slicer program that works within "Repetier", another freeware thing) has a honeycomb infill option, but when I tried it early on it made a mess. I had a lot of other problems then, though, so I may try it again., though I don't really need more strength on the waveguides.
 
Yowza, 1mm is way too big. I went as big as .7mm and had to dial it back. I'm currently using .4mm, which is right about where I started.

The big nozzles make REALLY strong parts but they have some drawbacks:

1) you blow through a ton of PLA
2) Any type of horizontal 'bridges' tend to sag.

You can offset issue #2 with some really careful orientation of the waveguide, but I find the smaller nozzles are way less hassle.

For instance, as we speak I am 3D printing some enclosures for my current project, and I'm able to print shapes that are as low as 10 degrees. The 'general' recommendation is to keep that to 45 degrees or greater, but I can get away with ten degrees by using a fast print speed and a small nozzle. I'll upload a pic later to illustrate what I mean. They're enclosures for those Gento woofers you're using on your original Synergy project.

I'm also using Slic3r. I tried Cura but Slic3r is good for a control freak like me. Cura seems to print a lot faster but I could probably make Slic3r print faster too. I'm kinda paranoid about tweaking my settings because they seem to be working.
 
Actually, it isn't. The mouth end has about a 2mm gap between it and the bed due to warping. It also does an 'elephant foot' kind of thing there where the lip around the horn squishes out slightly. I printed the other half of this last night, and here is what they look like held together (both halves are the same file).
gap.png

Any suggestions? Turn down the bed heat (I ran it at 50C all layers)? I already have a strong fan on the nozzle cooling down the layers when placed. I could just use a sander and get the two parts to join better (being careful to not sand at the throat area. I could probably divide the model at a slight angle on the 'cutting plane' (this is in Fusion3 360) to compensate and make them fit a little better, but the on-bed edge would still need work to make it flat.

edit: forgot -- I did the dividing in Meshmixer

Actually increasing the bed temp might help. I usually run at 55 to 60 degree but with PLA you could even go up to 70 degree. To prevent the print curling up, increasing the bed temp might do it. To get the best sticking print to the bed you'd have to run a few tests with varying the z start height (nozzle above bed). Moving it down just a little bit just might make it stick way better. Try a few small test prints, even though it might smear out that first layer, the adhesion usually improves a lot! It's a lot easier to remove the excess with a sharp blade. I've had prints stick that were almost impossible to get loose again. Some people like printing on a glass bed, others use a printing sticker, on our Velleman we're printing on what looks like a circuit board. I've worked with them all, but the best recipe has always been finding the right bed temperature + z height for the start of the print. Closer to the bed usually improves the adhesion. Print the first layer at a lower speed.

Another trick that helps is to add a brim around the printed part, making the part sticking to the bed larger in size and at the right temperature it will stick better. You could also incorporate a few flanges in your design to increase the surface area that sticks to the bed on corners or on that end. It could also help to glue them together.

Check the print temperature too, you're printing pretty big massive shapes and that builds up a lot of heat. Use the lowest temperature that still has good flow and print results. There are pretty big differences between PLA filaments. Even between colours of the same brand PLA. Some will allow lower printing temperatures, others come out crispy if the temperature gets too low.

Hope this helps...
 
Last edited:
Oh, that's a really good point. I totally forgot to mention that.

Literally 75% of my prints failed until I started adding a SERIOUS brim to every print.

Slic3r has a really wimpy 'brim' option, but I really take it to another level. Here's what I do:

In the model itself, I draw a 'ring' on the very first layer. I make the ring about 1.25mm high and 2.5mm thick. I have the ring go THROUGH my part.

Here's how this works:

If you have a big hunk of plastic, it will tend to 'curl' at the edges, like you're seeing. By having a 'ring' that goes through the part, the ring will keep the part from curling up. The ring is on the first layer, so the ring is basically 'glued' to the print bed.

The key here, is that the ring is relatively skinny. Since it's so skinny, it stays warm on the bed. (Preventing the ring itself from delaminating.)

The ring is so tightly laminated to the print bed, when I pull up the ring, it takes the tape right with it. It's basically glued to the print bed.

I've tried literally gluing the part to the print bed, but a big ol' ring works better and is less messy. On my old printer I used to zip tie the part right to the print bed, but my new printer and it's heated bed doesn't require these extreme measures.

I'll post pics later.
 
I'm also using Slic3r. I tried Cura but Slic3r is good for a control freak like me. Cura seems to print a lot faster but I could probably make Slic3r print faster too. I'm kinda paranoid about tweaking my settings because they seem to be working.

I love Cura but I'll admit Slic3r offers more control :)

The infill in Cura made for stronger parts, however the newest Repetier has Slic3r Prusa edition. That one has triangle infill from top to bottom.
 
I'm using a 1mm nozzle (and a 40W hot end called a "Volcano"!) because I don't have the patience needed to print something this large with a more normal nozzle. Detail suffers (and talk about a lot of trial-and-error getting it to print even successfully, forget pretty), but that's not much of an issue for something like these big waveguides.

Yeah, strength is not something I'm worried about on this, the printed parts with a 1mm nozzle are really tough. I used two outer perimeter walls on the prints shown, I may try one with a single wall, it should still be strong enough and print significantly faster. The slicer I'm using ("Slic3r", a freeware slicer program that works within "Repetier", another freeware thing) has a honeycomb infill option, but when I tried it early on it made a mess. I had a lot of other problems then, though, so I may try it again., though I don't really need more strength on the waveguides.

Repetier also has Cura build in as a slicer. Just try it. It's a bit more stripped down but very good results can be had with it. It features a top to bottom rectangular infill and usually is a lot faster to print than Slic3r. I 've also used Slic3r's honeycomb but the rectangle pattern in Cura convinced me it was the way to go. I only noticed the difference due to one of our printers (a Chinese clone) which only had a stripped down Cura version. Even very thin walled objects turned out to be very rigid. Like I mentioned before, the new Slic3r Prusa edition might be fun too! Just download the new Repetier 2.0.0

My biggest nozzle is 0.5 mm. Yes the small nozzles take forever :). Though speed isn't a factor if you want the smoothest most detailed results. Not really needed here though.
 
awesome work bwalso. I've been following your work since the cosine. i've even got a stash of the aura ns6 to build my own cosines. cant wait to see how these turn out for you. :)

Those NS6 are gold. I would've designed those into this waveguide if they were still available (even if at a lot more than the $9 each they were going fo when PE was blowing them out)
 
The fun part with 3D prints is you can use complex shapes and it won't be any harder to print. I use fillets and chamfers everywhere to make it more printable. Gravity is the only thing holding us back in design. Unless you add a second extruder that only prints supports and uses a dissolvable filament. One day I'll try that, as all overhang design limitations would be a thing of the past. (at the cost of even more printing time of course)

I'm amazed how many printable objects online look square and without imagination or thought put into it. That's one thing where 3D printing allows us more freedom.

In your waveguide case I'd increase surface area touching the bed to see if that helps. Just a couple of flanges and a brim all around and increase bed temp to 60 degrees. Play with the starting height of the nozzle, move it closer by 0.05 mm at a time in some test prints.
 
I somehow expected your prints to include a volume filler under the cones but I guess you would be able to incorporate that only after a few iterations used for testing :).

Should be fun, working out a waveguide that incorporates all small details. Curious to see/hear how they perform for you. Maybe over time you'd be able to find a recipe for sand filled prints to dampen the horn :D.
 
Founder of XSA-Labs
Joined 2012
Paid Member
There is a big trial and error phase to get the settings just right so that it doesn't warp or peel off the bed. I printed a big adapter to hold my woofer to the waveguide and that warped at first. Then I did some experimenting with temperature profiles in time or as function of number of layers. Start with hotter bed when laying down initial layers to get good adhesion. Then slowly cool it as you get higher. The differential temperature is what causes warping. Minimize temperature differentials. So I start with 55C bed after 10 layers I cool to 50C. Brims and skirts help but I don't use a fan to cool. I don't use a fan as that causes more differentials. Shut off the workpiece fan but leave the one cooling the nozzle assembly. The other trick is to mix 50:50 Elmers white glue and water. Use a brush and apply thin layer to 50C glass bed and let dry. That thin layer keeps workpiece firmly bonded to glass bed. Play with nozzle temps too. I found a few degrees hotter made all the difference and it was smooth and pimple free. Maybe 118C to 120C for PLA.

The first time I made this it was warped. Here is how nice and flat it was and warp free after applying the changes.

527487d1453821187-bookshelf-multi-way-point-source-horn-horn-woofer-adapter-04.jpg


527484d1453821187-bookshelf-multi-way-point-source-horn-horn-woofer-adapter-01.jpg
 
Last edited:
I somehow expected your prints to include a volume filler under the cones but I guess you would be able to incorporate that only after a few iterations used for testing :).

The thinking was that I could test that using modeling clay like I have in the past. But the cone on that midrange driver isn't very deep at all, I might need to actually add some front volume (dig away some plastic) if playing with the apertures doesn't do it. I would like to work the aperture to not be so much at the edge of the cone, but there isn't a lot to work with, will have to see how it looks at its high end. With some luck I should see tomorrow.
 
Yeah, the technology is pretty amazing. I had gone to a local makers' club meeting here about a year ago where some guys were giving a talk and showing stuff they printed -- they were little toy type trinkets, didn't look very good and the presenters talked a lot about how tricky it was to set up to get prints right. I concluded that the method wasn't of much value and forgot about it. But Patrick Bateman showed some waveguide prints he had made a few months ago and got me to take a 2nd look. And now, I spend way too much time obsessing about 3D printing instead of just about speakers!