| |
OK, I'm a little bit proud of this... and, mostly, amazed that it functioned at
all. I was selected to participate in the HALE (High Altitude LEGO Extravaganza)
after submitted two proposals in the hopes one would make it... and instead,
both were accepted. Well, today was the day - both payloads flew:
http://thenxtstep.blogspot.com/2008/07/hale-launch-day-coverage.html
One was a camera platform, (based on an earlier robot I posted on NXTlog called
"Nadar"), named "Gypsy". Besides using an NXT to control the platform pitch and
an off-the-shelf camera for video & still photography, it had three sensors
(pressure/temperature, sound, and 3-axis accelometer) as well as a HiTechnic
protoboard (I love that thing) that networked two more temperature sensors, a
light sensor (CDS cell), and could be used to turn a separate heater system on
and off. The biggest unknown on this one was actually it's support system -
using LEGO turntables and studless beams (no glue, no metal, no reinforcing with
non-LEGO components) to support the whole payload. That might be tough, when you
consider the exterior of the payload where this LEGO was exposed ended up around
-60° C, and was almost certainly shock-loaded to at least 2-3 G's during
descent.
The Hardware description:
http://thenxtstep.blogspot.com/2008/07/gypsy-hardware-t-6-days.html
(I particularly had fun dressing the minifigs for the mission, of course)
And the Software description:
http://thenxtstep.blogspot.com/2008/07/gypsy-software-t-24-hours.html
(probably the most complex NXT-G program I've put together)
The really cool thing? _The_LEGO_Didn't_Break_. The payload seems to have worked
OK (I'll know more later when I look at the data), but the fact that the ABS
actually handled the load under those very extreme conditions (bitter cold and
near vacuum) really amazed me.
The other payload was even riskier: Lil' Joe was cut free from the main payload
at 80,000', and had to deploy its own parachute under control of the NXT...
after waiting for some time in a high-altitude free-fall in imitation of Col.
Joe Kittinger. Anything went wrong, and it was DOOMED:
http://thenxtstep.blogspot.com/2008/07/lil-joe-t-7-days.html
Well... something did go wrong. BADLY wrong. While my program seems to have
worked flawlessly, and the deployment of the parachute functioned, there was no
way I could test the dynamics of a parachute being deployed from behind a foam
block at a speed of several hundred miles per hour in an incredibly thin
airstream (evidently driving around like a fool holding the payload outside my
car window at 35 mph isn't the same thing). On the way down, at some point, the
parachute and the tailfin became hopelessly tangled and the parachute must have
collapsed. I've no idea yet how fast it hit the desert floor, but it must have
been HARD. The guys who found it said the NXT and SPOT transmitter that were
inside the inch-thick-plus styrofoam shell punched straight through it upon
impact (I'm still waiting for pictures).
The amazing thing? _The_LEGO_Didn't_Break_. OK, there may have been pieces
deformed or bent (I've not seen the payload yet), but when the ground team
picked it up, the NXT was still running the program, logging data, even after
that horrendous impact. Absolutely amazing. I was worried that the NXT might not
survive a "normal" impact, which would have taken place at a speed in excess of
1000'/min, let alone one where the parachute collapses leading to something 2-3
(or more) times as fast (30 mph or more likely far faster). Simply unbelievable.
I can't wait to get the data back, but I've got to say this is BY FAR the
riskiest thing I've ever done with my LEGO... AND IT HANDLED IT!!
Brickshelf gallery of the payloads before they left on the mission:
http://www.brickshelf.com/cgi-bin/gallery.cgi?f=324334
I'll put up post-mission details when I get them.
--
Brian Davis
|
|
| |
Give you joy for your very successful mission. And a glass of wine with you.
Incredible. Will be expecting with much anticipation your final report to the
Admirality and the subsequent publishing of results in the LEGO Gazette.
Tommy Armstrong
|
|
| |
Give you joy for your very successful mission. And a glass of wine with you.
Incredible. Will be expecting with much anticipation your final report to the
Admiralty and the subsequent publishing of results in the LEGO Gazette.
I hate it when I misspell words.
Tommy Armstrong
|
|
| |
It certainly was the most exciting robot project which we participated in.
Many thanks:
1. to the organizers : Dr. Eric Wang and his team at Nevada University. Great
job guys !!!
2. to the sponsors : Nevada Space Grant, LEGO Company, Energizer, National
Instruments
3. to the contributors :
* all the teams, although, honestly, I love most the FLL's contribution -under
the guidance of David Levy- and ex aequo the 4th grades' marshmallows under the
supervision of Barbara Bratzel and Chris Rogers. These kids leave us adults
ashamed.
* Brian Davis's freefall robot and Gypsy prove reliability and hopefully will
deliver lots of good pictures
* Eugene Tsai's Brix-Catcher : wow !
* Jurgen Leitner and David Leal Martinez' REEL-E : this is real research ! But
please explain the name.
* The LEGO Mindstorms mysterious payload leaves us under suspens.
* Finally our own team: thanks to Francis Massen, Jean Mootz and Jean-Claude
Krack, the LCD students and the Convict robotics group: now let's hope we get
valuable data. (http://www.convict.lu/htm/rob/hale2.htm and
http://www.legoengineering.com/content/view/98/65/ (thanks Morgan!)
Claude Baumann
|
|
| |
In lugnet.robotics, Brian Davis wrote:
> OK, I'm a little bit proud of this... and, mostly, amazed that it functioned at
> all.
You should be proud! Excellent success. I hope ALL your data will be published.
The nerd in me really wants to know what the terminal velocity of a box of
Styrofoam filled with Lego is...
And pictures, I'm assuming the ground recovery crew went crazy with pictures of
the cras.. ummm landing sights.
Again, congratulations on a successful mission.
Tom
|
|
| |
In lugnet.robotics, Claude Baumann wrote:
> * Jurgen Leitner and David Leal Martinez' REEL-E : this is real research !
> But please explain the name.
Surely you've heard of WALL-E by now...
|
|
| |
In lugnet.robotics, Thomas Atkinson wrote:
> I hope ALL your data will be published.
A couple of folks have asked me about that. Trust me, anything I can will be
free for anybody who wants it, but I'll have to figure out a way to put it up.
Brickshelf is really for pictures, and I push it by posting images of data
graphs... but here, I think some folks want the actual program and actual data.
> The nerd in me really wants to know what the terminal
> velocity of a box of Styrofoam filled with Lego is...
The nerd in you is not, at all, alone :).
I know it was dropped from roughly 80,000', and made the trip to the ground in
under 23 minutes (perhaps as fast as 13 minutes). That's an average speed of
between 40 and 70 mph, but that's average - it should have been much higher in
the high, thin atmosphere, before parachute deployment. Lil' Joe was programmed
to take care of its own 'chute deployment: it timed how long it was going up,
assumed an ascent speed of at least 14' per second (data already recovered shows
it was closer to 20 ft/sec), and uses that to calculate an assumed altitude. It
then assumes a "hard deck" at 15,000', calculates a maximum drop distance, and
bases it's "free-fall" time on an average descent rate of 900 ft/sec (613 mph, a
very safe overestimate). Based on knowing all that, the free-fall time should
have been somewhere between about 40 to 70 seconds (depending on when it "turned
on"), leading to pulling 'chute at an elevation between 54,500' (likely) or as
low as 8,200' (ulp! But a drastic overestimate). during a true free-fall of 40
seconds, Lil' Joe would reach close to 870 mph, again almost certainly an
overestimate, so treat that as an upper limit.
For comparison, the main payload (coming down under a large parachute) seems to
have hit descent speeds of about 160 mph shortly after the balloon popped,
decreasing smoothly to a ground impact velocity that looks to have been around
14 mph if not less.
Where this gets interesting is in figuring out what Lil' Joe's potential impact
velocity was. The entire trip down took no more than 13 to 23 minutes, but
without any parachute the first 30,000' or more whips by in a few minutes at
*best*, even with air friction. That means the bulk of that time was spend
descending from below 54,000'. A little more math implies a ground impact
somewhere in the range of 30 to 50 mph.
Think of driving your car at slightly below highway speeds into a concrete
bridge abutment, with about an inch of stiff Styrofoam and a little bit of soft
foam as your safety system. Ouch.
And yeah, I figured all this out before the ground team had found the final
payload. This sort of thing really matters to me if I can predict things *in
advance*, rather than just casual observation after the fact. A product of my
education (physicist).
> And pictures, I'm assuming the ground recovery crew went crazy
> with pictures of the cras.. ummm landing sights.
See for yourself. I'll grab some of these for Brickshelf later on, but Eric has
started a gallery:
http://gallery.me.com/lego.professor#100014
In particular there's a picture of Lil' Joe "in situ", with the bottom blown out
(along with the tether, anchor pad, and black foam insert that should have been
inside the shell), but the NXT & SPOT not visible through the breach. From some
hasty analysis of the pictures, I think Lil' Joe actually *bounced* hard enough
that the payload shell ricocheted up while the internals were still coming down
to do that - its the only explanation I've come up with that seems to explain
the crash/landing site.
> Again, congratulations on a successful mission.
Thanks. I can't wait to see what else worked (& what didn't).
--
Brian Davis
|
|
| |
In lugnet.robotics, Brian Davis wrote:
> (evidently driving around like a fool holding the payload outside my
> car window at 35 mph isn't the same thing)
Brian do you have one of these near you?
http://www.airkix.com/pix_flix/flix_advert1.asp
This just happens to be over the road from my local LEGO shop. (local being 85
Miles, needless to say I dont go there that often!)
Mike.
|
|
| |
In lugnet.robotics, David Laswell wrote:
> In lugnet.robotics, Claude Baumann wrote:
> > * Jurgen Leitner and David Leal Martinez' REEL-E : this is real research !
> > But please explain the name.
>
> Surely you've heard of WALL-E by now...
Not necessarily. WALL-E has not been released entirely worldwide. Much of Europe
is still waiting for its release.
Joe
|
|
|
