Let's Design and Build a (mostly) Digital Theremin!

Posted: 6/9/2020 7:53:16 PM
tinkeringdude

From: Germany

Joined: 8/30/2014

Now all you need is a bit of that "mechanical distortion effect" (however that works) of clarinet and you can implement a digital Klezmer band, ha!
Probably a bit much effort for a gimmick  squeeky - though I wonder, is it just the maybe not entirely intuitive result of modulating the clarinet sound with the vocal sound, or is there more going on...

But seriously, esp. some of the not-bass clarinet notes sound pretty realistic to me (and I'm not familiar with the bass ones, but fast modulation might also make it sound less like one)

Posted: 6/9/2020 8:35:36 PM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

"Now all you need is a bit of that "mechanical distortion effect" (however that works) of clarinet and you can implement a digital Klezmer band, ha!
Probably a bit much effort for a gimmick  squeeky - though I wonder, is it just the maybe not entirely intuitive result of modulating the clarinet sound with the vocal sound, or is there more going on..."  - tinkeringdude

I think that's a lot of it.  Hugh Le Caine was able to get almost the exact sound of growling brass from his Electric Sacbut, and that was fairly primitive technology (tubes).  His joke was he reached a new peak in low-down. :-)  (See: Sugar Blues.)

"But seriously, esp. some of the not-bass clarinet notes sound pretty realistic to me (and I'm not familiar with the bass ones, but fast modulation might also make it sound less like one)"

Yeah, higher registers are almost always easier to synthesize for any given instrument, as there are fewer harmonic "telltales" for the ear to latch onto.  And clarinet is more of a tremolo than vibrato type of instrument, I should have really held back on my vibrato there.  I wonder if any Thereminists out there are able to do realistic tremolo w/ their volume hand?  I don't think I've ever seen that in any video.

Posted: 6/12/2020 4:24:41 AM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

In The Doghouse

Getting to know OpenSCAD.  Can't say I'll be a guru any time soon, but am learning the basic ins and outs.  Here's today's revision of a 1/2 scale LCD & encoders "doghouse":

I've since improved the LCD bezel opening tapering (for better or for worse, SCAD really makes you think long and hard about geometry).

Some thoughts:

1. Since it probably won't end up being a 100% solid part (i.e. it will have infill), specifying where each hole goes is fairly important, as the 3D printing process can reinforce these areas.
2. I'm thinking more and more that these parts (the LED tuner would be housed in a separate doghouse) could function much like the pickguard on an electric guitar: hide the huge nasty cutaway behind it, while providing access to it.  That would mean, I believe, through holes on the front, with bolt heads showing.  The Theremin main body could then be an otherwise sealed box.
3. I wonder if the LCD mounting holes could be incorporated into holding the thing down?
4. Roger says paint goes a long way towards hiding the 3D printing telltales (many otherwise pristine molded parts also end up painted).
5. The part looks rather aesthetically uninteresting?  No matter what they tell you, it's always a beauty contest.
6. I'm thinking of removing the holes on the right and left, and using the lower right encoder pushbutton to do acal.

Posted: 6/12/2020 10:13:08 AM
pitts8rh

From: Minnesota USA

Joined: 11/27/2015

Looking good, Eric!  The pain of learning 3D modeling seems sooooo much easier when you have virtually instant feedback from a printer.  Even with its flaws, you can't ignore the benefits of being able to print up conceptual pieces, templates, and in many cases finished parts.  I could kick myself for shunning 3D printers for so long.

1. Since it probably won't end up being a 100% solid part (i.e. it will have infill), specifying where each hole goes is fairly important, as the 3D printing process can reinforce these areas.

Sometimes it's just easier to print solid so you can drill and fit things on a draft print, then measure your features and update the model.   Usually when I get to the point of detailing or making things that mate I start getting itchy to get something printed without fully thinking through the fine points (such as hole placements).

2. I'm thinking more and more that these parts (the LED tuner would be housed in a separate doghouse) could function much like the pickguard on an electric guitar: hide the huge nasty cutaway behind it, while providing access to it.  That would mean, I believe, through holes on the front, with bolt heads showing.  The Theremin main body could then be an otherwise sealed box.

That's a good approach. Keep the back enclosure (wood box or whatever) simple and put the details into the configurable and easily reproduced printed front panel.  Another thing to think about is making the panel itself three dimensional.  Although individually you can only print flat-faced surfaces, nothing says that you can't build up a front panel from multiple pieces.  Think of something like a Moog synthesizer, where instead of silk screened borders delineating functional divisions on the front panel you break these areas up into separate elevated panels or islands on a larger back panel.  Aesthetic design features and trims can be added where the elements can be made from different materials using print patterns that work best for that individual part alone. 

4. Roger says paint goes a long way towards hiding the 3D printing telltales (many otherwise pristine molded parts also end up painted).

Yes, but there is a problem that is not unique to painting rough 3D printed parts.  Any surface that has imperfections such as pinholes, or lines in the case of 3D printed parts, does not readily accept paint to form a nice smooth surface, particularly if sprayed.  The surface tension of paint tends to make it crawl up on the islands and avoid filling holes or grooves, so instead of flowing out and filling imperfections it tends to aggravate those flaws.  If you've ever tried to spray a finish onto some open-grained wood like mahogany or oak you will know what this is like.

There are a few ways around this.  Unless you are dealing with a flat or highly organic printed shape, sanding the surface to smooth it very much is not practical.  Plus PLA and PETG don't sand (or machine) well because the heat makes the material gummy.

It is easier to fill the voids in the material so that the fluid paint is given a chance to flow out and level rather than crawl up into islands due to surface tension. You can do this by brushing on a first coat of sandable sealer, since this is much more effective at filling voids than spraying.

Another spray-only method involves the use of high-solids paint applied in very dry building coats, so dry that flow and surface tension never come into play. A few of these dry coats will actually begin to bridge the surface gaps, and with that a normal flowing series of final coats can be applied.

Thirdly, a high-solids paint can be applied in multiple heavy wet coats and the voids will eventually fill, although this will take more paint and can cause some loss of definition around details and at edges.  Paints like hammertone and some of the new textured finishes can hide a lot.  I've recently experimented with textured truck bed liner paint, and it works really well for some less detailed parts.

Fortunately, there is no real need to seek out special paints for plastics.  Tooth adhesion to the 3D printed surface precludes any bonding issues from what I have seen.

5. The part looks rather aesthetically uninteresting?  No matter what they tell you, it's always a beauty contest.

Yes, but adding curves and interesting design elements in the cad model is easy, and for fabrication the printer doesn't care - it is a most compliant employee!

Posted: 6/13/2020 5:28:22 PM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

Thanks for that Roger!  What printed wall thickness should I be aiming for?  Also, I'm thinking countersunk Philips wood screws to hold the thing down to the box below, is that what you might do?  Or would you do a snap-on type thing?

How might you connect together a multiple doghouse stack, glue?  Have you tried gluing PLA or PETG?

A reel of black Overture PETG arrived yesterday, perhaps my next doghouse sample will be from that.

===============

Yesterday I made it so ACAL happens when pressing the lower right UI screen selector encoder.  This is a natural: there are only UI 20 screens, so it's easier to spin the encoder to go to the main page than to press the encoder (thus zeroing it).  And the knob function is the same on all pages, so one can press it anywhere and get an ACAL to happen.  Don't know why I didn't think of that literally a year ago.

Also, I just realized that I had a moral imperative! to remove the "air space" taken up in RAM by defunct parameters.  Four at 10 bytes each recovers 40 bytes of precious RAM for potential future use.  So one last jumbling of the parameters took place, undone by the "uf" (update file) command in the editor, and I can sleep better at night knowing that the parameter space is now contiguous and more orderly.

===============

I don't know why, but I'm drawn to the bass version of almost every conventional orchestral instrument.  And, in my mind, the regular instrument can go down lower than it physically can in real life, and I prefer that to the middle & higher end.  Weird.

Looking at the bass flute today, it's a real head scratcher in terms of behavior.

Posted: 6/13/2020 5:46:51 PM
tinkeringdude

From: Germany

Joined: 8/30/2014

 Also, I'm thinking countersunk Philips wood screws to hold the thing down to the box below, is that what you might do?  Or would you do a snap-on type thing?.

How about using those metal "soldering iron" inserts I once linked to? If you intend to open that thing more than a couple times, that should have better life time of the threads.

Posted: 6/13/2020 7:32:26 PM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

"How about using those metal "soldering iron" inserts I once linked to? If you intend to open that thing more than a couple times, that should have better life time of the threads."  - tinkeringdude

Those are definitely on my mind!  And I really like the video you linked to.  But in this instance I'm thinking more of through holes in the doghouse with screw heads showing, with the screws going down into the wood and biting into that.  Much like screws in an electric guitar pickguard.

Posted: 6/13/2020 9:11:16 PM
pitts8rh

From: Minnesota USA

Joined: 11/27/2015

If I could stick my nose into this topic...

I advised Dewster to try a couple methods as an alternative to the brass inserts.  I've tried them ( the brass inserts) in a couple applications with both good and bad results.  While they look nice and work well for light duty applications such as enclosure covers or light-duty attachments, they have a few weaknesses.  They can't take much of a pulling force before they pull right out of the plastic, and if they are overloaded they can seize and spin.  Some are better than others with large knurls to melt into a hole, but they still pull out.

For applications with rear access you can just print in a hex pocket for a standard stainless nut (below).  These too can spin if overloaded, but the idea is to make the nuts a press (or pull-in with a screw) fit so that they resist rotation.


Another method uses the unique ability to print voids or pockets in the structure to contain a nut internally.  A square nut is pressed into the pocket shown below from a side access. These are both stronger and less prone to seizing and turning than brass.

What printed wall thickness should I be aiming for? -Dew

For small boxes something like 2.0 to 3.2mm might be adequate.  Large things can be 4 to 6mm thick but with only 3 or 4 shells and 10-20% infill.  Like corrugated cardboard.

Also, I'm thinking countersunk Philips wood screws to hold the thing down to the box below, is that what you might do?

Trilobulars are best for plastic, but for hardware store screws I use sheet metal screws, not wood screws.  Flat, oval, pan, or truss head sheet metal screws. The thread and the taper is different, and you don't want the unthreaded shank that wood screws have.

Or would you do a snap-on type thing?
Snaps are good in some places, and they are oh so satisfying.  I'm printing some snap together Raspberry Pi cases, and that snap lid is almost as satisfying as peeling protective film off a brand new electronic display.

How might you connect together a multiple doghouse stack, glue?  Have you tried gluing PLA or PETG?
Use rear screws into bosses where possible, or glue is fine if you are essentially building up a single molded unit that never needs to come apart. If you have adequate glue pegs (like a pin in a boss hole) something like JBWeld is my first choice, but many people report success with cyanoacrylate CA.  I would used a thickened CA for its gap-filling ability and so it doesn't run out of control and up your sleeve (ask me how I thought of that!).  None of the readily available adhesives are going to solvent-weld the joints, so you are relying on tooth adhesion, which is generally fine because 3D prints a full of "tooth".   Solvent welding can be done with a dichloromethane mix that is sold for use with hard acrylics and polycarbonates.  It's not something you'll find at Home Depot; Delvies plastics has it, though.  It's water thin and will go everywhere if you don't have a good applicator.

Posted: 6/13/2020 9:41:55 PM
dewster

From: Northern NJ, USA

Joined: 2/17/2012

"Trilobulars are best for plastic, but for hardware store screws I use sheet metal screws, not wood screws.  Flat, oval, pan, or truss head sheet metal screws. The thread and the taper is different, and you don't want the unthreaded shank that wood screws have."  - pitts8rh

Ah, yes, I meant sheet metal screws - I almost never use wood screws for anything.  Thank you for the glue info too!  That's a really heinous texture on your first pic!

Posted: 6/13/2020 10:16:10 PM
pitts8rh

From: Minnesota USA

Joined: 11/27/2015

That texture is from the flex plates that Prusa uses.  I actually like it much better than the texture on the plates that I have been able to get.   It probably looks so severe because that part is really small, like 25mm or so, and that is a 3mm nut.

You must be logged in to post a reply. Please log in or register for a new account.