• FAQ  • Search  • Memberlist  • Usergroups   • Register   • Profile  • Log in to check your private messages  • Log in 

3D printing mouthpiece experiments


Goto page 1, 2  Next
 
Post new topic   Reply to topic    trumpetherald.com Forum Index -> Mouthpieces
View previous topic :: View next topic  
Author Message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Mon Nov 02, 2015 7:12 am    Post subject: 3D printing mouthpiece experiments Reply with quote

I have access to a 3D printer (Ultimaker) and that plus my desire to experiment with strange mouthpiece designs has gotten me to printing a few. I first did a 1 1/2C trumpet mouthpiece I found on-line. It played extremely well for what it was; the major downside was the rim being too rough for the lips, and that can be fixed with some epoxy.

This weekend I hand-measured an antique cornet mouthpiece I had and did a print of it. Here are the results.





What I discovered is there is a systematic error of .3-.4mm, you can perhaps make out how the drill bore is somewhat smaller in the copy. Even with this discrepancy it plays remarkably well - better than most of the mouthpieces in my pile. I am going to shift the input numbers to compensate for this difference and do some more prints.

If I can get a good copy then the fun really begins! I can change around whatever parameter I want. Once I get a keeper I can send it to Kanstul for a "metal" print.

Anyone else doing 3D printer experiments?


Last edited by scottfsmith on Wed Sep 11, 2019 11:38 am; edited 2 times in total
Back to top
View user's profile Send private message
patdublc
Heavyweight Member


Joined: 02 Jun 2006
Posts: 870
Location: Salisbury, MD

PostPosted: Mon Nov 02, 2015 8:12 am    Post subject: Reply with quote

I've talked to several people (some of which produce mouthpieces professionally) about 3d printing them. The general feedback is that the printing technology is still not quite there for the required precision. But, those people also tend to have high dollar/high throughput machines that can crank out brass pieces.
In time, 3d printing will change a lot of business models allowing people to fabricate at home. You may still license a design from a company but actually build it at home. There's good and bad to this. It will be very difficult to share proprietary designs with someone. Selling PDF files is in a similar place. It is so easy to share unlicensed copies.
_________________
Pat Shaner
Play Wedge Mouthpieces by Dr. Dave exclusively.
Experiment with LOTS of horn makes and models.
Back to top
View user's profile Send private message
razeontherock
Heavyweight Member


Joined: 05 Jun 2004
Posts: 10548
Location: The land of GR and Getzen

PostPosted: Mon Nov 02, 2015 10:05 am    Post subject: Re: 3D printing mouthpiece experiments Reply with quote

scottfsmith wrote:


What I discovered is there is a systematic error of .3-.4mm


Very interesting! Is this error like either + .3 mm OR - .4 mm? Or is it consistently off in the same direction?

That translates to about .012" inches. (12 thou) The most accurate metal workers I'm aware of are tool and die machinists, who need to hold a tolerance of +/- .0002" (two tenths) The idea of that uber precision tolerance is that the final product will actually be accurate at the nearest thousandth of an inch. If you're saying that every dimension is consistently off in the same direction. and that the deviation stays between .3 and .4 mm, that translates to within 4 thou by the way I'm figuring it here.

Just to give you a basis for comparison. I would call that very promising! It'd be really interesting to do that with a material that is workable after printing, whether in a lathe and some sort of cutting / shaping tool, sandpaper, or however it might work. I would think the finishing details like smoothing out any hard lines in critical areas like the throat as cup funnels into venturi, the rest of the backbore from there .... perhaps improving the bite a touch ... you may be able to wind up with something that plays really well.
Back to top
View user's profile Send private message Send e-mail
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Mon Nov 02, 2015 12:27 pm    Post subject: Reply with quote

The error was all around -.3mm - its like it printed correctly and as the plastic cooled it shrunk ever so slightly.

Hey I just realized Google probably knows the answer. It pulled up the following:

https://ultimaker.com/en/community/view/7984-pla-shrink-factor

-- it seems the PLA (the plastic used) is known to shrink by a certain amount. The fact that the center is hollow may lead to relatively more shrinkage.

Along with the general shrinkage the surface is not completely smooth, but that can be sanded down as you mention. I had to sand down the rim to get it playable, since the piece is sitting on the rim when it prints it has extra plastic there. There are also occasional printer errors - if you look at the outside of the shank you will see one little glitch on the print I did for example.

Overall I am hopeful of a +-.1mm tolerance once I get everything tuned. This should be good enough for serious experiments I hope.
Back to top
View user's profile Send private message
snichols
Heavyweight Member


Joined: 12 Apr 2010
Posts: 576
Location: Virginia

PostPosted: Mon Nov 02, 2015 12:44 pm    Post subject: Reply with quote

Not to be tacky or anything, but how much does the material cost to make one piece like that? I'm not super familiar with 3D printing, so I don't know if that's $5 in material or $50.... I can only imagine that over the years as they become more precise, commonplace, and material prices come down that they might be a real cost-saver...
Back to top
View user's profile Send private message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Mon Nov 02, 2015 2:41 pm    Post subject: Reply with quote

The material is PLA (plastic) filament and it costs around $30 for a 500' roll. A mouthpiece is probably a foot or so worth of material so thats 6 cents each.

One reason why its so little material is by default the printer only makes a .4mm thick wall, the inner part is mostly hollow. I am going to try to print some filled-in ones eventually, but they have more problems with shrinkage so am doing the hollow ones for now. Acoustically there is some difference with the hollow walls, but its less than I thought it would be.
Back to top
View user's profile Send private message
TrumpetBoy907
Regular Member


Joined: 15 Oct 2015
Posts: 18
Location: Anchorage, Alaska

PostPosted: Tue Nov 03, 2015 3:29 am    Post subject: Reply with quote

I have a friend that 3D printed a mouthpiece with a moustache on it... played pretty terribly, but looked great! Lime Green!
Back to top
View user's profile Send private message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Mon Nov 09, 2015 9:59 am    Post subject: Reply with quote

One more followup for those potentially interesting in doing this.. I have been having fun every evening the last week, doing a print each night more or less. Here are a few things I have learned.

Overall I am getting more and more impressed - I really can't measure any difference with my calipers between the original and the copy (once I have done a few corrections outlined below). They do not sound identical but even the hollow copies are just as musical and playable.

I am currently adding .15mm to all radius dimensions to account for shrinkage. Thats .3mm to the diameter. The drill is still coming out a bit narrow but I can sand it to exactly the width I need with a sandpaper-wrapped drill bit.

The rims are one of the biggest challenges, it is sitting on the rim to print and there are artifacts from the base. Basically I think its best to make little contour in the rim in the input file, and just use sandpaper to shape it after the print. This takes a lot of practice as mouthpiece rims have subtle curve structures.

In order to get a smooth rim I have been using progressively finer sandpaper up to 600, and then polishing by rubbing hard on fabric.

Its hard to measure mouthpieces by hand. I use calipers as well as a big set of drill bits with many fine-grained sizes. But I found I needed several passes of re-measurement to get things right. A 3D scanner would be very helpful.

I am getting an artifact in the shank about 1cm from the end, its when the print stops having a hollow interior because the walls are meeting each other. I am having to hand file down that spot to smooth it out.

Here are this last weeks experiments.






The two leftmost ones were initial trials. The 3rd one is a very faithful copy. The last three are various wacky experiments (notice the ribbing in the bowl bottom of the 4th one - that is intentional). The ribbing is a failure so far, but some of the experiments are proving interesting. The one on the far right has the choke point 5mm into the backbore and looks worth pursuing.

Here is a print in action

Back to top
View user's profile Send private message
ProAm
Heavyweight Member


Joined: 05 Feb 2008
Posts: 810

PostPosted: Tue Nov 10, 2015 1:38 pm    Post subject: Reply with quote

Looks like some fun experiments! I am enjoying reading about them. Please keep posting updates.
Back to top
View user's profile Send private message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Wed Nov 11, 2015 7:42 pm    Post subject: Reply with quote

Sure! I'll throw out an update now and then.

The most recent thing I learned is there are standard techniques for polishing the plastic. I already had some buffing wheels and compound but didn't realize one of the compounds (red) was meant to buff plastic. It seems to work very well, the rims are shiny now.

My recent experiments are all with moving the choke point further down the bore. I had an old mouthpiece (~1860's Sudre piece) with a choke point almost 15mm further in, and the "frontbore" (the part between the bowl and choke point) is like the backbore but in reverse - it starts out like a really fat bore by the cup and then narrows down to the choke point. This mouthpiece has this amazing "throaty" (sic) sound that I would like to replicate. Its not a good overall mouthpiece, it has no high range, but this one aspect is really impressive. So I took the original mouthpiece I measured and made a variation where I moved the choke point 5mm further in, and another where I moved it 10mm in. I did this by moving the choke point and then eyeballing the move on the others to make a nice curve. Here is the profile of the 10mm move:



- the point at 30mm was at 20mm originally. These guys clearly have the throaty sound I am looking for, and I am enjoying playing on them. They have more midrange and a less "reedy" sound.
Back to top
View user's profile Send private message
razeontherock
Heavyweight Member


Joined: 05 Jun 2004
Posts: 10548
Location: The land of GR and Getzen

PostPosted: Thu Nov 12, 2015 12:22 am    Post subject: Reply with quote

As you adjust the "frontbore," (what some mouthpiece makers refer to as the "throat," then using the term drill size or venturi) you will need to adjust the internal dimensions to preserve the cylindrical section of the venturi. And also make some attempt to preserve the backbore. The cylindrical section is really important though. One thing you want to test for is what happens to intonation as you exercise dynamics. You want ampc to hold pitch as you change from pp to FF, while doing nothing at all to compensate.

Some very expensive mpcs do not have any dynamic range at all where they are steady in pitch. Makes it impossible to tune!
Back to top
View user's profile Send private message Send e-mail
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Thu Nov 12, 2015 5:56 am    Post subject: Reply with quote

The original mouthpiece has no cylindrical section. My vague impression is the early Bach mouthpieces (or someone else in that period) had two things new: they had a pronounced cylinder section in the drill, and they also had a parabolic backbore shape in many of them. None of my old mouthpieces has a cylindrical section at the choke point, and they all have either conical or slightly flaring backbores.

Here is the original one's measurement, the Courtois mouthpiece pictured above:



You can see there is no cylinder on it. In looking at the stretched throat profile above you might say it in fact has a cylinder of sorts, there is a long stretch where the backbore does not change much -- a "flared cylinder"

Re: pitch, the ribbed one was too unstable on some pitches. The extended throat mouthpieces are at least decent but have not been carefully evaluated for pitch. With the longer backbore they are actually more pitch-stable than the original.
Back to top
View user's profile Send private message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Sun Nov 22, 2015 11:14 am    Post subject: Reply with quote

One more update for those interested.. I have continued to learn more as I have done more prints. Here are some of the ones I have made so far:



I switched from copying this old Courtois mouthpiece to a similar period Courtois Arban mouthpiece, this is the kind that Arban himself designed and played on. I think they may be the earliest mouthpieces with a more modern cup profile - a more shallow C-ish shape and no "sharp edge" going into the throat. My original goal for doing prints was to make a version of this guy with a good rim on it - the original is an amazing player but has a cookie cutter type rim.

I also switched to doing solid prints, by making the wall thickness 4mm - its "all wall" then, no hollow fill. This did not end up making that much difference in anything, but I am going to keep doing the all solid prints since they take about the same time.

The printer has four quality settings, I was printing on "normal" (100 micron = .1mm layer thickness) which is the second from the most coarse. I recently switched to one more fine, 60 microns. This has both made the shapes come out more smooth and has made the prints more accurate. The print length is about double. You can see the mouthpieces on the left in the picture are a bit less rough on the outside. I am now trying a 40 micron print to see if that helps any more; its three hours print time. The printer goes all the way to 20 micron but supposedly the improvement has diminished to little by then. The speed can also be adjusted, I might try a slower print head speed to see if that improves accuracy.

Anyway for more serious prints 60micron layer thickness is needed - I found that I had errors of up to .1mm in the 100 micron which went away at the 60 micron. Also in terms of sonic fidelity the 60 micron copy is getting extremely close to the original; the 100 micron are still nice players but have a slightly different character to them. In the metal mouthpiece world, a silver re-plate will add 20 microns, and that is under what most players will notice. Overall, I don't think the current 3D printers can make a perfect copy but will make something you are just as happy with playing even though you will notice subtle differences. Even with 1 micron accuracy I would not be getting a perfect copy since my hand measurements are probably accurate to 50 microns or worse.

Speaking of hand measurements, I have continued to refine the measurements by directly comparing the original and the copy - with calipers you can keep going back and forth and get more accuracy on a comparison measurement than on the absolute number. Based on this I see where the copy differs and re-measure the original several more times in that spot, do another print, refine again, etc. I am still using .15mm as a uniform shrink factor and its working very well, I am finding errors are more attributable to inaccuracies in measurement or too coarse a setting on the printer.
Back to top
View user's profile Send private message
Leadherent
Veteran Member


Joined: 27 Dec 2002
Posts: 134

PostPosted: Tue Dec 01, 2015 12:23 am    Post subject: Reply with quote

This is interesting stuff. Keep us posted on the progress, please!

Leadherent
_________________
Confidence is what you have before you understand the problem - Woody Allen

1929 Conn 22B New York Symphony
1967 Conn 60B
Marcinkiewicz E16

Courtois 154R Flugelhorn
Bob Reeves 40HF
Back to top
View user's profile Send private message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Tue Dec 01, 2015 5:30 pm    Post subject: Reply with quote

Sure!

There were still some glitches in the bowl and I realized it is because as the print is running it is building the "roof" of the mouthpiece dome layer by layer and each layer added overhangs a little from the previous and then droops a touch before it cools. The solution: Print upside-down! The problem on the inner bowl turns into an overhang on the outer side of the bowl which is not dimension-critical. I had to make a base for it to stand on but once I got that the bowl came out beautifully. As an added bonus I can now print the rim exactly how I want. If you look at the Kanstul mouthpiece comparator site you can see rims have subtle curves on them which can make a big difference in how they feel. Those rim curves I am now getting great reproductions of.

I also slowed the speed down by half and lowered the temperature from 210C to 190C (how hot the plastic is being melted) and those two also increased the accuracy. I am also always using .04mm layers. Now I feel like I am getting within 50 microns or less of my measurements.

Here are some pictures:



Back to top
View user's profile Send private message
Richard III
Heavyweight Member


Joined: 22 May 2007
Posts: 1860
Location: Amador County, CA

PostPosted: Tue Dec 01, 2015 5:48 pm    Post subject: Reply with quote

So cool. When are you going to start taking orders?
_________________
Richard

For Sale: Cornets: Conn 77A, 80A, King Long Cornet Silvertone, Roth-Reynolds Professional, King Cleveland Superior.
Back to top
View user's profile Send private message Send e-mail
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Wed Dec 02, 2015 6:13 am    Post subject: Reply with quote

No plans there.. my main instrument for years was oboe, and making my own mouthpieces is what I have done since I was 15 years old. Old habits die hard

I do think the technology has reached the level where it could be really useful. I am going to do some experiments changing the backbore to see if I can get a more open sound from it. In general someone can take their favorite mouthpiece data, make ten different tweaks on it, and see how they work out. You can do it by hand with metal but its a lot harder to get the exact tweak you want. So, if I were starting out as a mouthpiece designer I would work only with 3D prints. Just contract someone to manufacture the designs that work out.
Back to top
View user's profile Send private message
schrödinger's cat
Regular Member


Joined: 21 Dec 2015
Posts: 33
Location: Texas

PostPosted: Tue Dec 22, 2015 10:28 am    Post subject: Reply with quote

As a mechanical engineer in an R&D facility, I deal with additive manufacturing every day. That said, perhaps this advice will be helpful.

Firstly, the tolerances you're observing are not too far outside the capability of quality fused-deposition-model (FDM) machines, which are typically in the realm +/-0.125mm or .005". If you want better consistency, tighter tolerances and overall better surface resolution, I'd look into stereolithography (SLA). This is the process we use and tolerances for external features can easily be held to +/-0.05mm with good design practices. I often print vertical bores to even tighter tolerances of +/-0.02mm but the operator has to stay on top of the calibration to achieve this as we've recently discovered...

All this aside, you should be able--as you've already alluded to--compensate for shrink in your model and achieve more accurate sizing. Keep in mind that areas of greater cross-section will pull (sink) more than those which are more consistent. This is why most injection-molded components feature consistent thickness ribs/walls/etc.. Many of the same design practices used for molded components apply here.

How much control do you have of your .stl model tolerances?

It might not matter much within the realm of FDM processing but will play a role in surface quality of higher resolution options like SLA. Something to keep in mind if you decide to make that investment.

Best of luck to you in what appears to be a very fun project!

Regards
Back to top
View user's profile Send private message
scottfsmith
Veteran Member


Joined: 27 Jun 2015
Posts: 333
Location: Maryland

PostPosted: Mon Jan 18, 2016 8:54 am    Post subject: Reply with quote

Hi, just noticing the above.. I do think I am about at the tolerance of the machine. By doing print/measure/adjust-stl/print/measure/adjust etc I can cancel out a lot of shrinkage etc error but at some point there is no more juice to squeeze. The good news is I am getting a very consistent product now and its pretty darn close to the original. The model itself I have total control and arbitrary accuracy on.

One last update on the printings, I found some bronze filament, its 40% bronze dust and 60% plastic. Its called BronzeFill. One advantage of BronzeFill is the mouthpieces have more heft to them and it makes them somewhat less bright sounding. Here are some pictures.





The bronze shines up nicely, and it has a real heft to it. If I made the exterior bigger I could get more dampening on it and may try that at some point. There is also a BrassFill version but its in "beta test" now and supposedly does not yet work well on 3mm filament machines like mine.

I also included the original piece I have been copying, a Courtois Arban #5. This mouthpiece is my favorite out of 20+ pieces old and new that I have, but the rim is too narrow. The two prints above aim to copy exactly the #5 profile except for the rim, and also they are a very slight touch more narrow at the top - its like a "#4.5" in the Courtois scale as I find the #5 a touch wide.

I tried a few other variations on the original #5 profile but the original beat them all out. Also the many copies I made while refining the accuracy were not quite as good players as the final ones where I pretty much nailed the original's measurements. It gets me more and more impressed with how subtle mouthpiece design is.
Back to top
View user's profile Send private message
sfbaytrumpet
New Member


Joined: 24 Aug 2015
Posts: 5

PostPosted: Thu Feb 11, 2016 9:21 pm    Post subject: Reply with quote

I had a middle school student experimenting with this as a project. He did a good job on the shape of a mpc, but throats and backbores never were close.
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic    trumpetherald.com Forum Index -> Mouthpieces All times are GMT - 8 Hours
Goto page 1, 2  Next
Page 1 of 2

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB © 2001, 2005 phpBB Group