"Tim Wescott" wrote in message
...
On Mon, 12 Nov 2012 16:03:47 -0800, mike wrote:

On 11/12/2012 2:27 PM, Tim Wescott wrote:
So, I'm working on a gizmo to go with a series of control systems
training seminars that I'm putting on.

The gizmo is basically a fan on a pivoted stick, with a control system
that works to maintain the angle of the stick relative to its mount at
a commanded value. It does a good job of giving people a visceral
understanding of how a feedback control system works, and I don't think
it's going to cost me much to produce.

But I went and took it on an airline flight for the first time this
week, and it didn't survive well. The position feedback from the pivot
is provided by a nice inexpensive potentiometer with a D-shaped hole,
into which one inserts a shaft of the correct dimensions. Here's a
close-up of the pot mounted on the board, with a shaft (and, if you've
sharp eyes, a little paper shim that keeps things snug).

https://docs.google.com/open?id=0B5l...TBCc3VzWDYxZ0E

On the flight (two flights each way, Portland to Ottawa and back), the
pot broke off the board. It was obviously punched out of the board by
the force of the shaft acting on the back of the pot. Fortunately the
training is for engineers, and it was at a corporate site, so my
customer was able to repair the thing and I was able to use it for
demonstrations.

Unfortunately -- even though I thought I had identified the problem and
fixed it -- it broke on the way back, too.

Now, one solution to this may just be that I need to find a different
way of putting the whole thing together so that it's easy to
disassemble for shipping, and then don't ship it assembled.

But I also want to put it out to the group for suggestions: the shaft
needs to be shimmed to a snug fit in the hole of the pot, or the slight
play between pot and shaft messes up the control (the arm will hunt
within the slop of the connection). But shimming things seems to set
up a problem with the shaft transmitting too much force to the pot, and
-- ping!!!

This thing has experienced a moderate amount of knocking around in my
shop, and use both on the property and around the local area without
breaking. But as soon as I go and ship the damn thing it breaks. So
not only am I very concerned about shipping, I'm a bit concerned about
this being a point of fragility in an otherwise reasonably stout
mechanism.

Comments appreciated.

Expecting that pot to survive shipment with a big lever attached is
optimistic.
Switch to a real pot that mounts by the shaft bearing, not the element
contacts and a setscrew to remove the lever for shipment.

Or you could leave it the way it is and demonstrate that the electrical
solution is worthless if the mechanical implementation fails. That's
probably a more valuable lesson.

That pot is less than $2.00, and very low friction.

Your "real" pots are big, clunky, expensive, high friction, and require
brackets which must be purchased which further ads to the cost.

So finding a way to isolate that pot from damage while still having it
work would be a big plus.


Perhaps the other useful lesson is that it is best to nail down and publish
all the desired specs before having a design review or calling the
consultants. I am still not clear about what sort of weight and torque is
on that small pot. Just how little friction do you need? How cheap does the
solution have to be?

It sounds like you want a stronger bearing to take up the mechanical loads,
and then have a connecting arm from the pivoting portion to the pot. The
stationary part of the bearing would be mounted with a bracket to board
above the pot with its axis in-line with the pot. The connecting arm could
extend through the center of a ball bearing, or it could be U-shaped and
connected to the outside if you need less than 360 deg. movement and the
bearing is not hollow. A cheap, easy-to-make, super-low-friction pivot can
be made by tapering both ends of a hard shaft to a point, and then clamp
onto the ends with two cup-point setscrews. You adjust the set screws to
have minimal friction with no slop. You also need threadlocker or a locking
nut to keep the setcrews in place.

Another low friction pivot design for small angles is a cross-flexure. For
some reason these are expensive to buy but relatively easy to make if you
have some spring steel shim stock. You can find one in the read/write arm
pivot in some hardrives. Hell, any bearing they use there would probably be
good for your purposes.