Motion control and data acquistion system

RG wrote:

I'm surprised you haven't gotten an answer yet.

There is an almost infinite number of directions that you can go on this
with actuator, feedback & controller selections.  Much of what you want
to do will depend on how quickly you want it to go together and how well
integrated you want it to be when you're done.

I'm mostly an expert in building full-custom systems, but here's my
guess at how to do this with off-the-shelf hardware:

You've got three or four parts of the system to select: the actuator
(motor), the sensor(s), the controller (which may be more than one
piece) and the operator interface (which may be part of one of the
controller pieces).

I would use a servo motor with an encoder as the actuator, with an
amplifier that supported encoder feedback.  For a reasonably sized motor
you'd probably want to gear this down to your screw.  Direct drive would
require a huge motor, and for most applications you'd go insanely fast
(figure that a motor really wants to go 3000 to 6000RPM; if you use a
10TPI screw that's ten inches a second, once it gets to speed).

You'll have some backlash in your gear train and your lead screw, so
unless you can always go at your target position from the same direction
you'll want some feedback straight from the thing you're moving.  Even
if you _are_ always going from the same direction the feedback will be
nice.  You could use an LVDT here, but that has a limited stroke.  I'd
consider an optical encoder strip of the sort that is used for direct
read outs for machine tools -- dunno where you'd get one, but I'll bet
that it's out there.  If you have to go at the target position from two
directions the backlash will _still_ give you fits, but it won't be as bad.

I would consider using a net-enabled controller, and putting the
supervisory control, the data acquisition, and the user interface on a
PC.  This has some down sides with closing loops in real time, but it
has the strong advantage of giving you a flexible programming
environment that is (relatively) easy to work with and that folks have a
lot of familiarity with.  I would strongly consider using National
Instrument's software to make this work, particularly as a one-off.
Even if you were going to make multiple copies to sell, I'd still
advocate using N-I for the prototype, then moving the application to
stand-alone thing for production, to avoid Labview's limitations and
royalties.

If you call National Instruments and tell them you think they can do it
they'll assure you that they can (it doesn't matter whether they can or
not: they'll sell hardware either way) and they'll recommend a local
consultant (who may or may not be able to make it work), who will
certainly use National Instruments software and hardware to do it.

I'm being pretty cynical about N-I, but if you keep your BS filter
turned on you'll do OK.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google?  See http://cfaj.freeshell.org/google/

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

to the world:


It sounds like you are making a test gig for QA of some product. I
have lots of experience building these kind of rigs for factories.

As Tim said, there are lots of ways to go about this. Which way you go
depends on consideration of price, flexibility, application,
production scale, etc. When I build something like this for a
customer, I usually spend many hours going over their needs. Many
times the customer does not even consider things that affect design
until I bring this up, so several meetings happen before I do any firm
designing.

Are you really wanting to use LVDTs or are you using that term as
generic for any linear position sensor?

I normally use an AC or DC gear motor on applications like this with
an inexpensive variable speed drive. As far as controls go, this
depends on your needs. One customer I had wanted the test stand to be
tied to a PC where the entire test sequence was controlled, monitored
and recorded. They wanted that test data dumped to Excel Spreadsheets
they supplied, with a single click of a mouse. They also wanted the
ability to run the tests with the computer shut down and totally by
hand. All these needs had to be considered.

If you are interested in hiring me to help in some manner beyond what
I can do here, send me a note at paulm@wiredog.com

RG wrote:

Whereabouts are you based?

RG wrote:

The most difficult part of this will be going at a speed of .001/min.
The resolution of the feedback would need to be very fine to achieve
this slow speed without bumping from one count to the next.  I don't
think a steady speed of 0.001 in/min is achievable over a period of
time.  Moving 0.001 every minute is.  The rest is easy and the parts can
be bought off the shelf.  A good motion controller can not only do the
motion control but can also control or monitor the force and buffer data
  so NI labview can retrieve it over Ethernet.

Peter Nachtwey

 

Heidenhain Linear scales for NC control such as the LC series offer
resolutions of 0.005 micron.
http://www.heidenhain.com
--
Anthony

You can't 'idiot proof' anything....every time you try, they just make
better idiots.

Remove sp to reply via email

RG wrote:


Sorry for delay but just returned from a conference.


That is quite a large effort (1.3 tonne of force or about about 140N.
 

The effort you are putting in requires a torque intensive gear-motor for
which a multi-phase AC drive system would probably be the best suited. This
would utilise a full vector AC drive package that can apply torque at low
speed. You will also need quite a high gear ratio.

[%X]

How fast do you need the max speed to be. This should be known as we could
not optimise the gear ratio required without this figure.


The last points are a matter of system architecture and selection of
suitable components. All quite easily accomplished.


I would keep the position sensing but would need more details about such as
maximum stroke length and whether or not you could consider zero-backlash
gearing with such loads. If the drive back-lash were minimised position of
the carriage could be evaluated from the rotation of the drive shaft with
multiple geared synchro/resolvers. If the LVDT used for monitoring the
position could give you a resolution over the whole length of less than
0.001 inches you may stand a chance. Achieving the positional accuracy at
the slow speed you mention would not be too much of a problem if you could
ramp down the motor for the positional landing. However, if you have a
high-ish rate of movement required (like the unspecified max speed) then
this may get to be more difficult (not necessarily impossible).


Your motion control system will need access to the position and force data
to ensure maintaining proper control over the actuation. You will also want
to monitor motor temperature and motor current.
 

Company's like Infranor <http://www.infranorusa.com/>  do packages for AC and
Brushless DC. They can handle single and multi-axis solutions and I have
worked with them in the past (two high integrity projects). Also worth a
look would be Emersons/Control Techniques
<http://www.emersonct.com/index.asp>  who do similar packages. I have used
some of their stuff but not the drives or servos.

If you can provide the other information I or others in this group may be
able to get a bit more specific about a suitable, and most likely off the
shelf, solution.

--
********************************************************************
Forth based HIDECS Consultancy .....<http://www.amleth.demon.co.uk/>
Mob: +44 (0)7811-639972
Tel: +44 (0)1235-811095
Going Forth Safely ..... EBA. www.electric-boat-association.org.uk..
********************************************************************

Paul E. Bennett wrote:

to max speed

I once worked on a lead-screw device for ruling precision gratings. (My
work consisted of design and installing the position-measuring system.)
The first discovery when the instrumentation was in place was that the
initial motion when the motor started up was retrograde. Small motions
at high loads often harbor surprises.

Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯

proclaimed to the world:


A 20 ton calendar I reworked used hydro assist to ease the load on the
adjusting screws used to set the nip gage. I had .0001 in sensors on
the bearing blocks to the frame and, under preload calibrated the nip
indicators. Take the 2 ton preload off the bearings and the nip
changed .0070 in. It is very hard to work at these tolerances and
loads.

On Sat, 17 Feb 2007 23:46:30 +0000, "Paul E. Bennett"


Not really. At the reduction ratio he is talking about a pretty small
motor will do the job. Loss is high for many gearbox setups, but there
are ones on the market to do this at these speeds. I did a 15 tonne
press for a test lab with a 2 hp.

Paul M <PaulMatWiredogdotcom> wrote:


As the OP did not state what his maximum speed requirement is it is hard to
tell. We know what he wants the slow speed to be. When I said "torque
intensive" I was rather referring to the output torque of the gear-motor
combination that would be needed to drive the ball-screw or ACME threaded
shaft. If the speed requirement is not too high then a fractional hp motor
might suit.

--
********************************************************************
Forth based HIDECS Consultancy .....<http://www.amleth.demon.co.uk/>
Mob: +44 (0)7811-639972
Tel: +44 (0)1235-811095
Going Forth Safely ..... EBA. www.electric-boat-association.org.uk..
********************************************************************

On Sun, 18 Feb 2007 21:49:09 +0000, "Paul E. Bennett"


True, but I don't know of any system out there that has a turn down
ratio so high as to need this.