air muscles
Tater Schuld
has anyone used these air muscles that i've seen used in the shadow
robot walker? anyone have any comments on them?
Tater Schuld (The Mad Scientist)
The Mad Factory
mad...@midway.tds.net
Edwin Wise
Schuld) wrote:
>hi
>has anyone used these air muscles that i've seen used in the shadow
>robot walker? anyone have any comments on them?
I've built a couple and experimented with them briefly -- they are
quite nifty! However, a single muscle acting alone is very springy
-- they say it helps if you use them in antagonistic pairs, and it
stiffens up the action nicely.
Edwin!
--
Edwin Wise (ew...@simreal.com)
Simulated Reality Systems, LLC
http://www.simreal.com
--> Check out our project-oriented beginner's robotics book
"Applied Robotics" at http://www.simreal.com/Robotics
Tater Schuld
wrote:
>On Sun, 05 Dec 1999 06:20:07 GMT, mad...@midway.tds.net (Tater
>Schuld) wrote:
>
>>hi
>>has anyone used these air muscles that i've seen used in the shadow
>>robot walker? anyone have any comments on them?
>
>I've built a couple and experimented with them briefly -- they are
>quite nifty! However, a single muscle acting alone is very springy
>-- they say it helps if you use them in antagonistic pairs, and it
>stiffens up the action nicely.
>
i was planning on doing so. so, why havent you used them more often?
Zaczek
octopod with them. We're still working on it because we have
not been able to make it fully walk...it has some trouble
supporting all its weight right now. We are using about 32 muscles
, 4 per leg in opposing pairs, but we have purchased an additional
8 large ones to help us make the robot support itself and finally
walk. These muscles are quite expensive though...smallest ones
are $20 so we even tried making our own. We were fairly successful
but the inner bladder we used was very imperfect so nearly half
of our self made muscles failed after about 1hr worth of cycling.
That's pretty good time (according to SHADOW) for a first try though.
Since then we just decided to buy them from Shadow instead.
You can check out the homepage for the project at:
http://atlantis.aae.uiuc.edu/~astro/octopod_project.html
We have a paper there that gives some details on how to build your own.
And we have the entire description of the control system we used.
Mario
On Sun, 5 Dec 1999, Tater Schuld wrote:
> hi
> has anyone used these air muscles that i've seen used in the shadow
> robot walker? anyone have any comments on them?
> Tater Schuld (The Mad Scientist)
> The Mad Factory
> mad...@midway.tds.net
>
>
>
<pre>
Projects & Design Competition: http://atlantis.aae.uiuc.edu/~astro
( NREMT-B ) Emergency Medical Technician - B
_____
/ | Graduate Student in Aero/Astro Eng. at the
.__./''''''| University of Illinois in Urbana/Champaign
._____________/ |/^^^^^^^\ " Center for the Simulation of
| Zaczek, `==="\_______/ Advanced Rockets "
`. Mariusz .___/^^^^^^^^\
`------------'~~~\________/ Office: 322D Talbot Lab
`........\ Urbana, IL 61801
`-------' (217) 333-4434
http://www.students.uiuc.edu/~zaczek
</pre>
Edwin Wise
Schuld) wrote:
>On Sun, 05 Dec 1999 15:23:37 GMT, ew...@simreal.com (Edwin Wise)
>wrote:
>>
>>I've built a couple and experimented with them briefly -- they are
>>quite nifty! However, a single muscle acting alone is very springy
>>-- they say it helps if you use them in antagonistic pairs, and it
>>stiffens up the action nicely.
>>
>i was planning on doing so. so, why havent you used them more often?
To get the strength I needed would take a large muscle -- which was
harder to build.
Muscles seemed harder to attach to the framework than standard
cylinders -- with more dead space to do so. I was getting a better
contraction-to-length ratio with cylinders; many of those calculations
*don't* include the dead space needed to attach the actuators.
The design for using antagonistic muscles was more complicated than
the design to use a two-direction cylinder.
I knew the cylinders would survive endless cycles of use, but there is
no wear data on the muscles.
And the cylinders seemed like they would be easier to control, again
because of the springiness of the muscles.
Tater Schuld
<zac...@cassini.aae.uiuc.edu> wrote:
>Yes, I've used them quite alot. A friend of mine and I built an
>octopod with them. We're still working on it because we have
>not been able to make it fully walk...it has some trouble
nice bot. why did you use a total of four muscles per leg instead of
just three?
Gordon McComb
Shadow Project page, and the various write-ups in books and Web pages, it
seemed to me that they would be best suited if they were in miniature, and
where you definitely wanted compliancy. "Soft touch" end effectors (with
or without human-looking fingers) would be one application. All of the
Air Muscles that I've seen a physically large, and therefore positioned
against traditional pneumatic cylinders. If they were very small -- the
thickness of a pencil and about half the length -- I'd bet they'd become
as ubiquitous as the BASIC Stamp.
I'm interested in testing out the shape-memory alloy linear actuator that
Mondotronics and a few others sell...I'm wondering if it has the low-cycle
rate problem that affects Nitinol/Biometal. Anyone know the specs? How
fast will it cycle?
-- Gordon
Roger Doulis
Gordon McComb wrote:
>
> I have not yet used Air Muscles, but in reading the info on it at the
> Shadow Project page, and the various write-ups in books and Web pages, it
> seemed to me that they would be best suited if they were in miniature, and
> where you definitely wanted compliancy. "Soft touch" end effectors (with
> or without human-looking fingers) would be one application. All of the
> Air Muscles that I've seen a physically large, and therefore positioned
> against traditional pneumatic cylinders. If they were very small -- the
> thickness of a pencil and about half the length -- I'd bet they'd become
> as ubiquitous as the BASIC Stamp.
>
> I'm interested in testing out the shape-memory alloy linear actuator that
> Mondotronics and a few others sell...I'm wondering if it has the low-cycle
> rate problem that affects Nitinol/Biometal. Anyone know the specs? How
> fast will it cycle?
>
> -- Gordon
>
< Big Snip >
The shaped memory alloys cycle very slowly and CONSUME power.
Try looking at http://www.dynalloy.com/ for specs.
--
Roger Doulis
(A polar bear is a rectangular bear after a co-ordinate transform).
Dept Civil Engineering (Clayton Campus) Ph 9905-4964
Monash University Fax 9905-1483
Wellington Rd.
Clayton Victoria 3168
Australia
Edwin Wise
wrote:
>Air Muscles that I've seen a physically large, and therefore positioned
>against traditional pneumatic cylinders. If they were very small -- the
>thickness of a pencil and about half the length -- I'd bet they'd become
>as ubiquitous as the BASIC Stamp.
An interesting thought -- tiny little muscles could be very
interesting to work with. They should be entirely soft then --
molded at some expense, I suppose, so there are no metal plugs or
fittings at the end, just a big rubber thing with the mesh covering,
and a tube to splice onto.
I can see it now -- Lego Technic, Bionic Expansion Set
Gordon McComb
> An interesting thought -- tiny little muscles could be very
> interesting to work with. They should be entirely soft then --
> molded at some expense, I suppose, so there are no metal plugs or
> fittings at the end, just a big rubber thing with the mesh covering,
> and a tube to splice onto.
>
> I can see it now -- Lego Technic, Bionic Expansion Set
Are the Air Muscles covered by a patent of some type? If not, I'd think
it would be fairly easy to construct something like this. What came to
mind were surgical balloons, but I think these would be too expensive
(sterilized and all that). There has to be a source of small, yet
durable, balloons for use as the bladder. A constrictive mesh, perhaps
finer than the kind they use on the Air Muscles, might work as the
container.
-- Gordon
Gordon,
Chain mail might last longer than mesh.
Jim
Roger Doulis
I don't think it is (#include disclamer.h) The work was originaly done
be McKibben and the device is knows as a McKibben pneumatic muscle. I
have some copies of the original papers at home somewhere.
Do a search for Pneumatic Muscle or McKibben Muscle (and other combos)
will turn up some good info including how to make them.
--
Roger Doulis
(A polar bear is a rectangular bear after a coordinate transform).
Regnirps
>> Are the Air Muscles covered by a patent of some type? If not, I'd think
>> it would be fairly easy to construct something like this. What came to
>> mind were surgical balloons, but I think these would be too expensive
>> (sterilized and all that). There has to be a source of small, yet
>> durable, balloons for use as the bladder. A constrictive mesh, perhaps
>> finer than the kind they use on the Air Muscles, might work as the
>> container.
>>
>> -- Gordon
>Gordon,
>Chain mail might last longer than mesh.
I think it isn't mesh like a net, it is braid like wire protector or Chinese
finger puzzle. It has to shorten when the diameter increases. But I have seen
fine wire rope versions of the braid for protecting bundles and for gripping
the ends of heavy cables.
Charlie Springer
Edwin Wise
<gmc...@gmccomb.com> wrote:
>Edwin Wise wrote:
>> An interesting thought -- tiny little muscles could be very
>> interesting to work with. They should be entirely soft then --
>
>Are the Air Muscles covered by a patent of some type?
The original work by McKibben is ancient...
>mind were surgical balloons, but I think these would be too expensive
>(sterilized and all that). There has to be a source of small, yet
>durable, balloons for use as the bladder.
I've used surgical tubing which is dirt cheap, since it is used for
many different applications. For larger tests, I've tried bicycle
innertubes, as well.
The best would be to do some mass-production and make cast-latex (or
something) "guts" that include the connectors, etc.
> A constrictive mesh, perhaps
>finer than the kind they use on the Air Muscles, might work as the
>container.
I've found that the cable-cover mesh that is sold here and there works
very well, and itsn't too expensive, and comes in many sizes. It's
been a few years since I did the research, but I found a couple of
source I think.
I actually have a medium-sized muscle in my toolbox still; most of the
bulk and nuisance of it lies in the plugs, connectors, clamps, and
other hardware necessary to put it together and attach it to the
robot.
Darrell Gerber
about
10-15,000 reps. How does this compare with everyones experiences? For them
to
be a commercially viable alternative I would think that number would have to
be
much higher.
Another question, how do they usually fail? Does the mesh covering fail or
the tube?
Does it fail in the middle or near the clamps? Changes to the design can be
made to
get additional life depending on the mode of failure. One thing that comes
to mind
is using a different material besides steel for the mesh so that it won't
rub on the tubing
and wear it out.
Just random thoughts on the subject but, like I said, not backed up by
experience.
Darrell
Edwin Wise <ew...@simreal.com> wrote in message
news:384d15e9...@news.texas.net...
> On Mon, 06 Dec 1999 15:49:21 GMT, Gordon McComb <gmc...@gmccomb.com>
> wrote:
>
> >Air Muscles that I've seen a physically large, and therefore positioned
> >against traditional pneumatic cylinders. If they were very small -- the
> >thickness of a pencil and about half the length -- I'd bet they'd become
> >as ubiquitous as the BASIC Stamp.
>
> An interesting thought -- tiny little muscles could be very
> interesting to work with. They should be entirely soft then --
> molded at some expense, I suppose, so there are no metal plugs or
> fittings at the end, just a big rubber thing with the mesh covering,
> and a tube to splice onto.
>
> I can see it now -- Lego Technic, Bionic Expansion Set
>
mac...@2.1.t.com
> There has to be a source of small, yet durable, balloons for use
> as the bladder.
Surgical rubber tubing (I got mine off a slingshot) sorta works, but
the walls are so thick it takes a certain amount of pressure just to
get it to *start* expanding. Once is starts, it expands & contracts
OK, and stretches quite a bit, but something better must be available!
> A constrictive mesh, perhaps finer than the kind they use on the
> Air Muscles, might work as the container.
I'm holding it in my hand right now. Samples I got at a trade show,
after dazzling (or confusing) the salesman with explanations of how
I was going to make air muscles with it. The factory thinks it's
for slipping over cable bundles. It works good for air muscles,
and comes in lots of sizes. Lots of colors, too, if you want your robot
to be eye-dazzling. <g>
The stuff expands to about 3 to 4 times it's smallest "at rest" diameter.
A 1/2" dia. piece easily holds a golf ball, like a Boa Constrictor
eating a Chihuahua.
The sample piece says TECHFLEX_I_ (800) 323-5140.
I don't know if they have a website.
-------------------------------------------------------------
* anti-spam address: Change "2.1.T" to "21T" to send email.
-------------------------------------------------------------
Gordon McComb
> the walls are so thick it takes a certain amount of pressure just to
> get it to *start* expanding. Once is starts, it expands & contracts
> OK, and stretches quite a bit, but something better must be available!
Actually what I was thinking of were balloons for angioplastic surgery.
These are very small, and have fairly thin walls, yet are strong. They are
made to be inserted into an artery and "pumped up" to free any clog in
there. In one "dry" demonstration I saw the balloon expanded more in one
direction than the other -- sort of what we want for miniature air muscles.
I well well imagine balloons angioplasty costing a $100 each, though. I don't
know, but that kind of surgical stuff can get really steep. Then again,
maybe they only cost $3 and the hospitals charge $100! <g>
> I'm holding it in my hand right now. Samples I got at a trade show,
> after dazzling (or confusing) the salesman with explanations of how
> I was going to make air muscles with it. The factory thinks it's
> for slipping over cable bundles. It works good for air muscles,
> and comes in lots of sizes. Lots of colors, too, if you want your robot
> to be eye-dazzling. <g>
Is this the stuff slit on one side? If not I'm not sure I've ever seen this
stuff. Anyone know of a souce on the net where one could see it firsthand...?
-- Gordon
Bob
multiple wires or cables into one large cable. It's usually plastic, but
I've seen it in metal as well. I've seen it at Radio Shack and at American
Science & Surplus in various lengths and diameters. Some of this is fairly
rough, so I'm not sure how well it would do regarding abrasion and
wear-and-tear on the bladder. If you could get this in Teflon or some other
low-friction material, it would probably last longer. I'm also not sure how
easy in would be to anchor it at the ends.
Oh Well. Just some ideas. Good Luck.
Bob
mac...@2.1.t.com wrote in message ...
>Gordon McComb <gmc...@gmccomb.com> wrote:
...
>> A constrictive mesh, perhaps finer than the kind they use on the
>> Air Muscles, might work as the container.
>
>I'm holding it in my hand right now. Samples I got at a trade show,
>after dazzling (or confusing) the salesman with explanations of how
>I was going to make air muscles with it. The factory thinks it's
>for slipping over cable bundles. It works good for air muscles,
>and comes in lots of sizes. Lots of colors, too, if you want your robot
>to be eye-dazzling. <g>
>
Dennis Clark
[snip]
:> I'm holding it in my hand right now. Samples I got at a trade show,
:> after dazzling (or confusing) the salesman with explanations of how
:> I was going to make air muscles with it. The factory thinks it's
:> for slipping over cable bundles. It works good for air muscles,
:> and comes in lots of sizes. Lots of colors, too, if you want your robot
:> to be eye-dazzling. <g>
: Is this the stuff slit on one side? If not I'm not sure I've ever seen this
: stuff. Anyone know of a souce on the net where one could see it firsthand...?
Nope. Its a "tube" of nylon (in the stuff I have seen) that is a woven
mesh. When its expanded in the middle the two ends contract towards each
other, strongly. I find it in various sizes at electronic surplus shops,
you know, the ones that are lit like caves and have people that don't
speak my language behind the counter to haggle over the cost of everything.
Too darn many projects!
DLC
--
============================================================================
* Dennis Clark Aristocrat at heart d...@verinet.com www.verinet.com/~dlc *
* Be well, do good work, and stay in touch -- Garrison Keillor *
============================================================================
Gordon McComb
> mesh. When its expanded in the middle the two ends contract towards each
> other, strongly. I find it in various sizes at electronic surplus shops,
> you know, the ones that are lit like caves and have people that don't
> speak my language behind the counter to haggle over the cost of everything.
Oh *that*! I was imagining some solid plasticky thing that had ribs or
something on it. I think Fry's has some of that as well (also lit like a
cave; try reading part numbers in their components section under .9
lux!!). I never thought of trying to use that for an ersatz Air Muscle
but I'll have to try it. I think I'll work on using LEGOs for
connections. Might make it easier to experiment with, but that still
doesn't overcome Edwin!'s objections of a completely pliable structure.
-- Gordon
Steve Lajoie
Edwin Wise wrote:
> I knew the cylinders would survive endless cycles of use, but there is
> no wear data on the muscles.
Steve Lajoie
Roger Doulis wrote:
>
> I don't think it is (#include disclamer.h) The work was originaly done
> be McKibben and the device is knows as a McKibben pneumatic muscle. I
> have some copies of the original papers at home somewhere.
>
> Do a search for Pneumatic Muscle or McKibben Muscle (and other combos)
> will turn up some good info including how to make them.
mac...@2.1.t.com
> > The factory thinks it's
> > for slipping over cable bundles. It works good for air muscles,
> > and comes in lots of sizes. Lots of colors, too, if you want your robot
> > to be eye-dazzling. <g>
>
> Is this the stuff slit on one side?
Nope. It's like a mesh "Chinese Handcuffs". It doesn't work well *as*
Chinese Handcuffs, because the mesh is Nylon or something similar, and
it lets my fingers slide out too easily.
Techflex is one mfr., there are probably others. I'm still looking
thru the tradeshow pile for more info, or a website.
Zaczek
inner tubing (although it had some imperfections in our
case)...you can get the info from my paper on the
walking octopod project I did with a friend...go to:
http://atlantis.aae.uiuc.edu/~astro/octopod_project.html
Mario
On Wed, 8 Dec 1999, Edwin Wise wrote:
> On Tue, 07 Dec 1999 09:51:00 -0800, Gordon McComb
> <gmc...@gmccomb.com> wrote:
>
> >Edwin Wise wrote:
> >> An interesting thought -- tiny little muscles could be very
> >> interesting to work with. They should be entirely soft then --
> >
> >Are the Air Muscles covered by a patent of some type?
>
> The original work by McKibben is ancient...
>
> >mind were surgical balloons, but I think these would be too expensive
> >(sterilized and all that). There has to be a source of small, yet
> >durable, balloons for use as the bladder.
>
> I've used surgical tubing which is dirt cheap, since it is used for
> many different applications. For larger tests, I've tried bicycle
> innertubes, as well.
>
> The best would be to do some mass-production and make cast-latex (or
> something) "guts" that include the connectors, etc.
>
> > A constrictive mesh, perhaps
> >finer than the kind they use on the Air Muscles, might work as the
> >container.
>
> I've found that the cable-cover mesh that is sold here and there works
> very well, and itsn't too expensive, and comes in many sizes. It's
> been a few years since I did the research, but I found a couple of
> source I think.
>
> I actually have a medium-sized muscle in my toolbox still; most of the
> bulk and nuisance of it lies in the plugs, connectors, clamps, and
> other hardware necessary to put it together and attach it to the
> robot.
>
> Edwin!
>
> --
> Edwin Wise (ew...@simreal.com)
> Simulated Reality Systems, LLC
> http://www.simreal.com
> --> Check out our project-oriented beginner's robotics book
> "Applied Robotics" at http://www.simreal.com/Robotics
>
>
<pre>
Zaczek
Failure in our projects case usually resulted in the inner tube
failing. We had no failures with the mesh and only limited failure
with the connectors we used. The tube on the other hand, we noticed,
had some small air bubbles trapped within it and caused failures
after some number of cycles. Actually, in our case , Shadow told
us that we did pretty good for a first try...half of our 32 lasted
for close to 45 minutes constantly being cycled at a rate of one
fill-exhaust per second. You can check out our site at:
http://atlantis.aae.uiuc.edu/~astro/octopod_project.html
Mario
> > An interesting thought -- tiny little muscles could be very
> > interesting to work with. They should be entirely soft then --
> > molded at some expense, I suppose, so there are no metal plugs or
> > fittings at the end, just a big rubber thing with the mesh covering,
> > and a tube to splice onto.
> >
> > I can see it now -- Lego Technic, Bionic Expansion Set
> >
Zaczek
we wanted full control in the four directions....It was
easier for us to control the motion with four since they
were opposing pairs. If you use 3 then you will have to
partially fill two to move the leg beack and up and that
kind of precision is very hard with all the various
variables such as Pressure, flow rate and valve speeds.
It was much easier to program for four on the M68332
microcontroller than it would be for 3.
Mario
P.s> Thanks for checking out our site.
<pre>
Steve Lajoie
Zaczek wrote:
>
> Failure in our projects case usually resulted in the inner tube
> failing. We had no failures with the mesh and only limited failure
> with the connectors we used. The tube on the other hand, we noticed,
> had some small air bubbles trapped within it and caused failures
> after some number of cycles. Actually, in our case , Shadow told
> us that we did pretty good for a first try...half of our 32 lasted
> for close to 45 minutes constantly being cycled at a rate of one
> fill-exhaust per second. You can check out our site at:
>
> http://atlantis.aae.uiuc.edu/~astro/octopod_project.html
The University of Washington reports that Latex tubing
last longer than the Manosil Silicone Rubber.
Interesting stuff on your web page.
Tater Schuld
wrote:
>Zaczek wrote:
>>
>> us that we did pretty good for a first try...half of our 32 lasted
>> for close to 45 minutes constantly being cycled at a rate of one
>> fill-exhaust per second. You can check out our site at:
>>
what is the mean time between failures on these things....just got my
order in from images comp........will be testing
Zaczek
the smallest ones and have cycled them alot....I can't really
give you an accurate answer but I can probably approximate.
We've owned them for over 2 years now and have probably run
a total of about 7 total hours with cycles of about one
fill/exhaust of every 1-2 seconds. I think I might be on the
low side though but that's about approximate. We only strained
them to 80 psi max and usually had it at about 50 psi due to
the air source we had. I find that they work ok....the stretch
length was a little disappointing as then only stretch/compress
about 10 mm (maybe more, can't remember).
Good luck with your muscles and let us know what you do with
them.
Mario
<pre>
Tater Schuld
wrote:
>On Tue, 07 Dec 1999 09:51:00 -0800, Gordon McComb
><gmc...@gmccomb.com> wrote:
>
>I've used surgical tubing which is dirt cheap, since it is used for
>many different applications. For larger tests, I've tried bicycle
>innertubes, as well.
do you know where i could get some surgical tubing?
>The best would be to do some mass-production and make cast-latex (or
>something) "guts" that include the connectors, etc.
this was my idea even before i knew these muscles even exsisted
>I've found that the cable-cover mesh that is sold here and there works
>very well, and itsn't too expensive, and comes in many sizes. It's
>been a few years since I did the research, but I found a couple of
>source I think.
could you supply me with a few sources?
>I actually have a medium-sized muscle in my toolbox still; most of the
>bulk and nuisance of it lies in the plugs, connectors, clamps, and
>other hardware necessary to put it together and attach it to the
>robot.
hmmmmm looking at the muscles i just got...i'd say they compare even
LESS than the dead space you get when using pneumatic cylinders
Edwin Wise
Schuld) wrote:
>do you know where i could get some surgical tubing?
Medical supply places from your phone book -- many of these places
will sell to the public, for people with "medical needs". Also, a
variety of exercise machines use surgical tubing for their resistance
mechanisms... also slingshots.
>>I've found that the cable-cover mesh that is sold here and there works
>>,,,
>could you supply me with a few sources?
Erm, it's been a long time so I don't recall the precise supplier I
used -- however my Newark and Allied catalogs list "Expandable
Sleeving" and "Expandable Braided Sleeving" in their Wire and Cable
Accessories section. The thing to do may be to go to your favorite
electronics or wire supplier and see what they have.
>hmmmmm looking at the muscles i just got...i'd say they compare even
>LESS than the dead space you get when using pneumatic cylinders
That is true, on the muscle itself, but when I took into account the
connection to the framework, my setup (at least) was not as efficient
(stretched length to contracted length, and wrt strength) as the
pneumatic cylinders. Your mileage may vary!
Edwin
Zaczek
I've made my own air muscles and I have a listing of the
supplier and part numbers in a paper I wrote on the project.
Go to:
http://atlantis.aae.uiuc.edu/~astro/octopod_project.html
I think I have a part number for the inner tubing but not sure
about the weave part number although I do have a listing of
a know distributor.
Mario
P.S. I'd appreciate any comments people have about the project
above and any of the projects that are on my page...we're
always looking for new ideas and ways to improve. Please, if
possible post to the comment board of the site. Thank you.
<pre>