Don Foreman wrote:
On Mon, 01 Sep 2008 07:54:47 -0700, Tim Wescott
wrote:
You don't mention material.
Thank you -- this is why I post these questions to the group, because
you know all the things I should be saying.
It's gotta not rust in a living room environment, and stainless (at
least from McMaster) is cheaper than brass. 6061 Aluminum may work at a
larger diameter than 3/16, but then I'd have to triple-step the shafts
or have huge bearings (which I don't want).
OK, so can't rust implies steel is OK if plated, otherwise brass, ally
or SS. If making bazillions of them they'd be plated steel because
the plating cost is negligable in large batches.
I don't understand the bit of triplestep but ally isn't that much
cheaper than brass so if you have reservations with ally I'd just
exclude it. 3/16" brass is about a buck a foot or about $0.45 per
gizmo. Some might find a few cents more for material to be
advantageous in terms of processing time. BTW, brass runs real nice in
plastic bearings, particularly delrin.
Brass won't rust but it will tarnish after a while. If that'd be
objectionable you need to say so. Then 303 would probably be the
best choice. 303 is very easy to machine in a rigid lathe but not as
easy as brass in a less-rigid lathe.
So what's drafting-ease for "make the shoulder any damn thing you want
as long as it's cheap"? Ditto for "cut the off-end of the shaft any
damn way you want as long as it's cheap"?
Just dimension the length with tolerance suitable for your
application. You don't have to say "cheap"; nobody ever wants "costs
more than would be necessary". With your specs I'm not sure if I'd
part off blanks in the lathe or just shear them.
BTW, if you have a concentricity requirement for the shaft body and
machined part, you need to specify it. A shiny spiffy shop would have
machines that hold a couple of tenths but guys that might want such a
job could well have chucks that are out .005".
A big part of my problem is that while I've gotten a bit of drafting in
high school, and I've worked around mechanical engineers for years, it
was all for very high-falutin stuff. So everything gets engineered down
to the gnat's eyelash, and "expensive" is a $10 part that keeps a
$500,000 system from working, not a $50 part that works perfectly all
the time.
So any instincts that I have are all skewed the wrong way.
This sounds more like a "consumer" deal. None of them will need to
work flawlessly for thousands of hours, right?
You're making me think this through, Don, which is one of the things I
was hoping for with my original post. I'm not sure what motivated me to
reject brass over stainless, but I see that 360 is only a hair more
expensive than 303 stainless, and it'll be way way easier to machine.
Maybe I'll get one stick of each from McMaster for prototyping, and see
what I think of it for myself.
One of my great frustrations on this project is that the driver for this
shaft design is a great little potentiometer with poorly defined
specifications. It costs about 1/3 as much as anything else that would
be remotely feasible, it has low friction, it slaps onto a surface-mount
board with all the same processes as the other parts, and has other good
features.
But the data sheet (at least the English-language one) doesn't give any
specs for how close to center the thing has to be run -- and I have to
account for shaft runout, bearing slop, _and_ board misalignment in my
error budget. At this point I'm guessing that .005 would be on the edge
of what's acceptable; I'll have to go over my measurements of the part
to make sure.
An alternative that I had rejected, but am now thinking may make sense,
would be to use a 1/4" shaft that's stepped down to 3/16 then again to
4mm. If this is done all in one step it would make it easy to get
really good concentricity (I'm not sure how good, but .0005 ought to be
quite achievable, yes?).
I wonder how much extra that would cost... I could still go 1/4" on the
off-end to allow it to just be cut in a shear. The cost would be two
sizes of bearings, but I don't see that as a big deal if it makes the
shaft enough less of a deal.
(here's the bearing I'm planning on using:
http://www.igus.com/igubal/igall2.asp?p=Clip).
If I could categorize the finished assembly it'd be "pro consumer". I'm
going to be teaching some control systems seminars, to take embedded
software engineers and spin them up on applying control systems theory
in software. In order to do this I'm making a gizmo with a simple & fun
mechanism, with a board that controls it. The intent is to get the
price to a point where I can send one home to each attendee (it's not
going to be a cheap seminar).
So it needs to work well for occasional use by one person for a year or
so, at which point it'll gather so much dust it won't be recognizable.
--
Tim Wescott
Wescott Design Services
http://www.wescottdesign.com
Do you need to implement control loops in software?
"Applied Control Theory for Embedded Systems" gives you just what it says.
See details at
http://www.wescottdesign.com/actfes/actfes.html