[simpits-tech] Re: motion platform, and outside view.

Thu, 5 Dec 2002 17:17:23 -0500

Hi Gordon,

It sounds like you've done some homework on this.  I did some searching and
reading of technical literature on motion platforms.  Some of them read like
a physics Ph.D. thesis paper (great for insomnia). :)  Indeed, the
algorithms for actuating the platform to simulate inertial forces are not
trivial.  Here are some web links:

http://www.nlr.nl/public/facilities/f111-04/index.html
http://mclab.me.pusan.ac.kr/english1/reserach/project/simulator.htm
http://wwwrobot.gmc.ulaval.ca/~ricard/Articles/Delft_ceas_95/delft_html.html
http://www.virtualsimulation.com/reports/a975001.pdf
http://www.virtualsimulation.com/reports/a99xxxx.pdf
http://www.dfrc.nasa.gov/DTRS/1988/PDF/H-1371.pdf
http://www-sdb.larc.nasa.gov/aiaa/pdf/aiaa-99-4328.pdf

Is anybody actually coding stuff like "washout"?

Craig R.

-----Original Message-----
From:	simpits-tech-admin@simpits.org [mailto:simpits-tech-admin@simpits.org]
On Behalf Of Gordan Sikic
Sent:	Thursday, December 05, 2002 2:32 AM
To:	simpits-tech@simpits.org
Subject:	Re: [simpits-tech] Re: motion platform, and outside view.

Hi,

On Thursday 05 December 2002 05:22, you wrote:
> I just read a book about material strength - "Building Large
> Dobsonian Telescopes" (I think...). Unfortunately I took it back to
> the library! I'll see if it's there, and copy the formula. The
> author provided a very straightforward way of figuring it all out.
> It went something like this:
>
> Double the size, cube the weight, but the stiffness depended upon
> the wall of the tube. Double the wall (e.g. 2mm instead of 1mm), and
> you quadruple (I think) the stiffness. They also said
> aluminium/aluminum was as good as steel in most applications. (If
> memory serves, I saw the same thing in a book on chassis design). It
> was also considerably lighter. I understand that it is difficult to
> weld, however (I don't know how to weld).
>
> The triangulation is similar to building trusses and some skyscraper
> beams. I was once involved in a project that dealt with the
> construction of a (very) large skyscraper, and one of the engineers
> told me that you need less material than you would think, but it has
> to be laid out properly. McGraw-Hill probably publishes something on
> this.
>
> Good luck!
> /Ian
>
> PS From experience of perching monitors in precarious positions, I
> can tell you that the point of balance shifts the moment you look
> away from that particular monitor... (I can still see the monitor
> pitching slowly forward!)
>

I adore jumping into middle of discussion :)

I mostly agree with you, but there is ine fundamental difference in
buildong telescopes and otion platform, and it is connected to the
type of motion.

With telescope, one wants as fine positioning as possible, without
disturbances (and to be able to follow the motion of the stars), but
the _time_ needed to achieve position os not a constraint.

On the other hand, with motion platform you have some kind of (much
faster) servo system, where you want position, speed and accelerations
to follow some calculatd criteria, and this is quite a different
problem.

I do agree that the platform should be as stiff as possible, but
additional requirement here is that the platform must be as "light" as
possible; when I state light, I mean that the mass is as low as
possible, but moments of inertia also. Of course, one might argue
about mass distribution including the pit, pilots, monitors and all
other equipement attached on the platform :)

This is importent because the platform should imitate dynamic system
with different characteristics, compared to the characteristics of the
platform itslef; you want one dynamic system to behave as different
one, and it requires power (and I mean POWER)

BTW,
I would to use the oporunity to put couple of problems, other than
mechanical ones.

(mostly addressedt to Jonas)
pls don't be angry with me for writing it down. I do appoogise if you
already know this.

What is the program that will calculate required motion of the
platform (position/speed/accelerations)? It is not enough to have
motion of the AC in order to drive motion platform; you have to use it
just as a input for the program which calculates required motion of
the motion platform (which is quite a task, beleive me:)

Another problem is concerned with safety. You are sitting on top of
this, and you don't want to allow (for example) max acceleration (of
wrong sign) when the platform is at max displacement. The platform
would hit the "suspension" couple of times, and then break.

Another safety problem is connected to the driving program. you must
not allow _any_ kind of MS related problems while anybody is on the
platform. One single GPF (general protection fault within windoze), or
computer lockup might drive platform crasy. One solution for this is
to use specialized OS for work in real time, like qnx (www.qnx.com)
profesional OS which is free to use for non commercial use. The other
one is to write the program what uses RT extensions for Linuux
(www.rtlinux.org); Needles to say, this involves writing program to
calculate required motion of the platform

You would also need to think of "panic button" which will cut off
power suply, and allow for quick platform abandoning, in case of fire,
for example.

How many degrees of freedom are you planing to allow? 3,4, 6? Having
more DOF's requires more complicated driving program, but having less
might (suprisingly) pose mechanical problems; in this case you are
constraining couple of DOF's using linkage, and if it is not well done
(like loose bearings), it might introduce unwanted motion, which will
be very noticable.

There are couple of "technical issues" with mechanical construction of
the platform, but (as Ian said), this should not be the problem for
mechanical/civil engineer.

If you reached this part of the message, I do admire you!
Once more, I do appologize if you already know about these problems.

ciao,

Gordan Sikic
mailto:gordan.sikic@inet.hr
_______________________________________________
Simpits-tech mailing list
Simpits-tech@simpits.org
http://www.simpits.org/mailman/listinfo/simpits-tech
To unsubscribe, please see the instructions at the bottom of the above page.
Thanks!