3D printing is actually not new -- it has been around for quite a while
now.  What is new is the consumer level low cost plastic printing
machine.  Printing on a cheap machine will produce coarse porous parts
but this can be minimized on better machines with better resolution
precision and careful control over things like extruder temperature and
feed rates.

In most cases where you need a good finish or true water-tightness,
you would apply what are referred to as secondary operations.  With
ABS, you can soak the part very briefly in acetone and all but really
sloppy layering will be made air tight.

A well-made printed plastic part can be pretty strong, but machined or
molded parts are almost always stronger.  A pattern can be made using
printing and this used (after improving the finish if necessary) to
cast in metal or other material.  There are several very strong low
temperature melting alloys that can be cast in plaster molds or even
high temperature silicone molds.  The metal can be melted in a pot on
the stove.

If the part has undercuts you will need either a flexible mold such
as silicone (can only directly handle molten metal of the lowest
temperature range) or you can make a flexible mold and then a wax
investment casting in plaster by the conventional methods.  I know of
people using this technique to make beautiful metal parts.

What has been discussed previously is having parts made by professional
additive manufacturing companies.  They can do excellent work in all
sorts of materials and the parts can have all the properties needed.
They're generally not cheap but, considering all aspects of the job,
can be very cost competitive, especially for small volume production.
You can also proof your parts in plastic on a cheap machine to know
that the money spent with the pros will not be a waste.

Generally 3D printing is not cost competitive for high volume production
although there are exceptional cases.  Each job needs to be considered
on an individual basis.  I seem to always get the tough jobs to produce
cost effectively -- never large enough to do it right nor small enough
to work as an artisan making parts by hand.  So I always have to
consider different methods and torture myself for weeks figuring out
the best approach.

As for making a sprocket by printing, this is an example of a part
that could be made very easily by conventional methods.  Turn the blank
on a lathe and cut the cogs with an end mill on a vertical mill using
a rotary table or spindex [spin index fixture] for indexing or other
indexing method if you have one.  Then smooth the teeth with a file.
The whole job would probably take less than an hour.

I could also cut the sprocket on my CNC mill but it would take longer
to design the part than to make it unless you do this regularly.  I have
a program for making gears that I could use but it would take me a day
to learn it.

If you got this far, you are probably interested in methods for
producing parts for mechanical music devices.  Bruce Newman and I are
organizing the technical workshops at the 2019 Seattle AMICA National
Meeting in July 2019.  We are hoping to have a workshop on 3D printing
and maybe CNC parts production.

Spencer Chase