3D printing of mouthpieces .....

[jacobg]

Dr. Sloan, what software do you recommend for creating a design?

[sloan]

Dr. Sloan, what software do you recommend for creating a design?

I use OpenSCAD, netfabb, and custom written Java code. Note - that's what *I* use, not necessarily what I recommend. Note that the "custom written Java code" is not an option for most designers. [all you youngsters out there: LEARN TO CODE!]

Local Engineering students are trained in CREO.

Many students have already picked up Blender.

No one was ever fired for specifying AutoCAD.

An up and comer is Rhinoceros, esp. when paired with Grasshopper.

Essentially, there are two design methods - you can design the surface, or you can design the solid. I think working with solids (basic objects such as spheres and cylinders plus "constructive solid geometry" operations (set theoretic ops on 3D point sets) is easiest, for me. Others prefer to directly model the surface.

Commercial grade printers now seem to uniformly want a surface definition - and that surface must be closed, watertight, and manifold. It must enclose a well-defined solid. Software which comes with the printer will slice this surface model and plan tool paths for the printer. My printers all work by Fused Deposition (stuff comes out a nozzle and stays where you put it). Other printers (esp. those that do metal) work by laying down layers of powder and selectively bonding the powder. It amounts to the same thing in the end. The difference is that some material is easier to melt and other material is easier to make into a fine powder.

When selecting software for 3D Print design, ask if it can easily produce "STL" files. This is the common "assembler language" used today. Hobbyists still work with "G-Code" which allows you to directly drive the printer assembly. Fun to play with, but not where you want to go if you are just starting out.

[eutubabone]

I personaly use GODZILLA. The "footprint" is larger than the Rhinocerus or grasshopper, and even larger than ELEPHANT. By the way, I have no idea what I'm talking about. I just got done teaching beligerant, obstinate 5th grade woodwinds and this is a good stress reliever.

[WitchyPsycho2]

At my college one of the Industrial tech professors has been making some for our Tuba/Euph studio. But mostly just as cut away mouthpieces and rims. And our cost for these are reasonable.

[Mark]

This is way out of my area of expertise; but since most mouthpieces are symmetrical around an axis running through the center of the shank, could you just use a 2D design and then spin it around the that axis to generate a 3D design?

[J.Harris]

This is way out of my area of expertise; but since most mouthpieces are symmetrical around an axis running through the center of the shank, could you just use a 2D design and then spin it around the that axis to generate a 3D design?

Mark, yes that is one method. I printed a tuba mouthpiece of my own design using that exact method. Unfortunately the end product was not refined enough to determine how effective the design was. The printer was set to layer in .010" increments which produces a fairly rough surface.

[sloan]

This is way out of my area of expertise; but since most mouthpieces are symmetrical around an axis running through the center of the shank, could you just use a 2D design and then spin it around the that axis to generate a 3D design?

Mark, yes that is one method. I printed a tuba mouthpiece of my own design using that exact method. Unfortunately the end product was not refined enough to determine how effective the design was. The printer was set to layer in .010" increments which produces a fairly rough surface.

I alluded to this above. To make a radially symmetric mouthpiece, all you need is a single polygon (think about slicing your mouthpiece down the middle - the cross-section will consist of two polygons which are mirror images of each other). It is then simple to spin that polygon around the central axis to form the solid model.

A *good* printer can produce quite reasonable models with 0.01" layers. And, ABS can easily be post-processed to smooth out the jaggies. One popular technique is to dip in a weak Acetone solution. Acetone dissolves the plastic, and then evaporates. When the Acetone disappears, the plastic re-forms. It's also possible to sand (polish) the surface.

Cheap printers may advertise 0.01" layers, but not really deliver. For a little bit extra, my best printer will do 0.007" layers. But, if you think you need a mirror finish, then some post-processing is required. I have produced mouthpieces that play just fine right off the printer - but if I were to use one for an extended period of time I'd probably want to polish the rim. TANSTAAFL.

[oldbandnerd]

Sloan .....
right now these printers use plastic and some sort of binder or glue to produce whatever it is you want made. Do you think this technology will someday progress and do the same thing with metal? I suppose that instead of cutting big pieces of metal down to smaller formed pieces of metal like in a lathe it would use some form of metal shavings or powder and melt it together. Do you see this happening ?

[WoWwYnAtoR]

...Commercial grade printers now seem to uniformly want a surface definition - and that surface must be closed, watertight, and manifold. It must enclose a well-defined solid. Software which comes with the printer will slice this surface model and plan tool paths for the printer. My printers all work by Fused Deposition (stuff comes out a nozzle and stays where you put it). Other printers (esp. those that do metal) work by laying down layers of powder and selectively bonding the powder. It amounts to the same thing in the end. The difference is that some material is easier to melt and other material is easier to make into a fine powder....

[joh_tuba]

It's relatively affordable to have ponoko print your stainless mouthpiece but the finished product would need some refinement.
http://www.ponoko.com/make-and-sell/sho ... less-steel

OR, you can use this company's software(design exactly what you want) and have it machined to your specs, polished, and plated, and shipped to your door: http://www.emachineshop.com/

[sloan]

It's relatively affordable to have ponoko print your stainless mouthpiece but the finished product would need some refinement.
http://www.ponoko.com/make-and-sell/sho ... less-steel

Note the costs. Stainless steel (rough - requiring post-processing) starts at $15 per cubic CENTIMETER. There are 2.54*2.54*2.54 = 16.4 cubic centimeters in a cubic inch. So, about $250 per cubic inch.

I charge my customers about $20 per cubic INCH for ABS plastic.

Some time ago, when my lab was new, an engineer brought me the design for a small box with lots of holes in it and a few stiffeners. He told me (others have confirmed) that he could have it made out of aluminum - but it would take 3 weeks and cost $1000. I made one for him in 12 hours and charged him $150. Three days later, he was back with a modified design. Two months later, he ordered six (6) for training purposes. When it flies to the ISS, it will still be made in aluminum (NASA understands aluminum, but ABS plastic is not on their "approved materials list") - but the plastic versions allowed him to evolve and test the design much faster and much more inexpensively. I think it's the same with mouthpieces - 3D printing is appropriate for fast prototyping (and perhaps customization), but mass production is not (yet) threatened.

For those who want to try this at home: my "industrial strength" ABS printers range in price (just the hardware) from $15k-50k; we have a proposal in to acquire a metal printer - again, just the hardware - $750,000. Don't throw away your lathes (for parts that *can* be made on a lathe!). On the third hand, I believe I could churn out "playable" ABS mouthpieces with a $4k printer (plus hand polishing) for something like $25 in consumables (plus your time for hand polishing). And, the 3D printed versions can include asymmetrical features that can't be turned on a lathe.

Or, you could ramp up to injection molding. Let's see - the first one will cost $10,001. Two will cost $10,002. Ten Thousand will cost $20,000. How many do you need?

[joh_tuba]

Excellent points, of course.

3D printing is obviously the future.

Another alternative, albeit not quite as cost effect as Dr. Sloan is this: http://www.emachineshop.com/
You could potentially have a one off brass mouthpiece turned and shipped for under $50. They'll also polish and plate before shipping for an extra fee.
The cost to do the same design in stainless is a lot more... but cheaper than the price Dr. Sloan referenced for Ponoko.

If nothing else, I encourage everyone to download the eMachineShop software and explore it. It's really too bad that you can't export much because the software is excellent.

[pjv]

We're only talking mouthpieces.
Often brass players wonder how their axe would respond with an alteration in just one part of the horn. This can be very expensive because the brass itself is expensive and shaping conical parts is time consuming and requires skill.
I wonder what the practical applications are of plastic parts for tuba (a mouthpipe for example) with the technical perfection these printers can achieve. Again, for testing purposes.

[sloan]

We're only talking mouthpieces.
Often brass players wonder how their axe would respond with an alteration in just one part of the horn. This can be very expensive because the brass itself is expensive and shaping conical parts is time consuming and requires skill.
I wonder what the practical applications are of plastic parts for tuba (a mouthpipe for example) with the technical perfection these printers can achieve. Again, for testing purposes.

Larger parts are mostly limited by the build envelope of the printers. Large size and fin precision are difficult to get at the same time. My largest printer can do 8"x8"x12".

I could, perhaps, do a piston valve set...that doesn't come apart. Probably not practical, but "interesting".

[joh_tuba]

Alright, let's go down the rabbit hole.

For those that want to design their own stuff, what are some of the design conventions one should generally adhere to?

What I've gathered:
Total Length: 3 7/8"
Outside largest diamter: 1 7/8"
Throat: generally around 1/3"
Shank tip diameter for American shank: According to Bach .519"
Euro shank tip: ?????
Shank length: 1 3/4"
Morse taper for shank: 1.4321 degrees OR 0.05/in
Cup diameter: generally 29-33mm
Safe wall thickness at shank tip for machining: ???

Is that all correct? What is a common cup depth to the smallest part of the throat? Conventions for shaping entry into the backbore? Conventions for shape of backbore?

[joh_tuba]

https://drive.google.com/file/d/0B2QfbI ... sp=sharing

Link is to an eMachineShop file describing a mouthpiece I just whipped up based on limited knowledge and my personal aesthetic.
Specs:
.323 Throat
Bach American shank
cup depth 2 1/8"
Cup width 1 3/8"
Flat rim 1/4" wide

Pretty sure I'm describing a pretty wide, deep funnel cup with a sharp rim and Laskey sized backbore on an american shank.

I encourage folks to visit http://www.emachineshop.com/ and download the software to see it for youself. You can view a 3D rendering of the finished product and get instant cost estimates to have it made in various materials with different finishes etc. An unpolished stainless steel 416 version will set you back $122 after shipping with the price dropping dramatically the more you produce. Raw unfinished brass is about $80 and you could polish it up at home and have it plated locally after you decide you like it.

Also, looks like DXF and STL files can be exported but I've not experimented with that. Might be helpful if you want Dr. Sloan to give it a try in plastic first.

Feedback?

[sloan]

Alright, let's go down the rabbit hole.

For those that want to design their own stuff, what are some of the design conventions one should generally adhere to?

I recommend starting with the Bach mouthpiece brochure (Google is your friend) - which
has a lot of useful "standard" dimensions.

First, define a standard shank (for extra credit, parameterize it) and outside shell.

That takes you from the tip of the shank to the lower boundary of the rim. Add a parameter for the radius of the rim.

Next, work on the inside - starting at the tip of the shank. Inside diameter at the tip is pretty much nailed down by your standard shank. The next point to define is the end of the (straight line?) inner surface through the shank.

Now, you get to work on the detailed shape from the shank, through the bowl, and around the rim. The only constraint is that you join up with the outer shell (just under the rim) and the entry to the inner shaft of the shank. Use your favorite method of defining smooth curves to get from A to B. You now have a polygon that defines (one half) of the cross-section through the axis of the mouthpiece.

For a rough sketch, perhaps as few as 10 points (x,y - from where the shank enters the bowl to he outer edge of the rim) should be enough. On the cheap, you could build a mouthpiece from this. For a little bit more, interpolate the non-straight-line portions of the polygon. "fit a spline to it"

To turn the polygon into a solid model, you rotate it about the axis. Programs such as OpenSCAD provide this as a single operation. You can easily produce more triangular patches than ABS plastic can resolve, so all (unwanted) sharp edges can be smoothed out.

Now - if you want asymmetric features (an oval rim, or a rim that changes thickness as a function of angle), the problem becomes slightly more complicated. The easiest way is to repeat the process above for n different angles around the axis. Then, you can either use those radial cross-sections directly, or interpolate. If that's too simple, you'll have to move up to a real 3D CAD program.

All of this can be done with 2D drafting tools, plus some fairly easy processing to turn the 2D drawing(s) into a solid model.

If you can't hack the last step, most people working in the area can probably get you there on the cheap. For example, I'd be happy to accept n radial cross-sections and print from that. Writing the program to read in n radial cross-sections and spit out a model suitable for printing is probably a 15 minute task. An hour, if you want documentation.

So...just to make this concrete, here's a file format:

N - number of radial cross-sections
n0 - number of x,y points in the first cross-section
x0,y0 - n0 x,y pairs
x1,y1
...
n1 - number of x,y points in the second cross-section
x0, y0 - n1 x,y pairs

x,y should be Real numbers, in either mm or inches (please specify - mm preferred)

By convention, the polygon should proceed counter-clockwise.

for example, here's a pipe that has an inner radius of 10mm, a length of 100mm, and a wall thickness of 2mm:

=============
1
4
0,10
100,10
100,12
0,12
=============

I'll bet that a quite reasonable (radially symmetric) mouthpiece can be specified in this format in fewer than 100 lines. If you want asymmetry, multiply that number by 128.

You create the file, I'll print it (and send a bill...). Let the games begin!

[sloan]

Also, looks like DXF and STL files can be exported but I've not experimented with that. Might be helpful if you want Dr. Sloan to give it a try in plastic first.

Feedback?

STL is *exactly* what I prefer for vanilla print jobs. Please check with something like "netfabb" - it's free - to be sure the STL actually defines a closed, watertight, manifold surface.

but, see my previous post for people who want to try out a mouthpiece idea who own only a pencil and graph paper.

And....when it can be "machined", that will probably always be the technology of choice. The fun starts when the machinist starts shaking his head in disbelief. Want to see a functioning Rubik's cube, printed as one (no assembly required, no dis-assembly allowed) piece? Or, the iconic sphere-within-a-sphere-within-a-sphere-within-a-sphere-within-a-sphere?

For tubists, consider the possibility of a complete piston valve assembly - that does not come apart. You probably would not be happy with the tolerances (you'd need really heavy oil...) but it could be done.

[sloan]

If nothing else, I encourage everyone to download the eMachineShop software and explore it. It's really too bad that you can't export much because the software is excellent.

My loyal staff tells me that eMachineShop does NOT do a good job of exporting STL (our preferred format) - but it does export (more competently) a couple of formats that can be read by freeware (we like "netfabb") that will not only verify (and repair as needed) the design, but also then output excellent STL. I *think* we are almost at the stage where I could consider taking an eMachineShop file and printing from it. [nominal extra charge for doing the file format conversion for you...]

[waiting patiently for reports on the tuba mouthpieces I have printed for TubeNetters recently...some of the designs look, um, "interesting"]

[joh_tuba]

I am guilty of an 'interesting' design and plan to write a very thorough review of my observations very soon.

Short version observations:
1) Apparently if there is a big end and a little end it'll probably play like a mouthpiece.. even if it's 'interesting'.. in fact it might work better than most in it's own 'special' aka not general purpose way.
2) A hollow shell of a mouthpiece made of ABS plastic feels, plays, and sounds just like a metal one.
3) 3D printing is a cost effective quick way to test ideas but probably not a great solution for a refined production product
4) I sent an STL file exported from eMachineshop that at the time Dr. Sloan said looked fine.. maybe not?