Loading ...
Sorry, an error occurred while loading the content.

Re: little walking insects

Expand Messages
  • Nick Fox
    I actually bought the stiquito robotic kit a while ago that uses the nitinol wire. I think it may be time to go into Edison mode and start prototyping. I was
    Message 1 of 11 , Sep 6, 2006
    • 0 Attachment
      I actually bought the stiquito robotic kit a while ago that uses the
      nitinol wire. I think it may be time to go into Edison mode and start
      prototyping. I was wondering if it was possible to get enough current
      out of these solar cells to run the thing:

      http://pagermotors.com/Flex_Cells.htm

      It looks like from the previous discussion that these might not be
      enough. Is it possible to use these in addition to the watch batteries
      to run the robot and continuously recharge the batteries?

      Nick


      --- In SeattleRobotics@yahoogroups.com, PeterBalch <PeterBalch@...> wrote:
      >
      > > From: "Branden Gunn" <engunneer@...>
      > > Nitinol contracts when it's hot,
      >
      > Actually, I don't think that's the right way to look at it.
      >
      > Nitinol doesn't "contract", it changes its elasticity. Its rest-length
      > always remains the same.
      >
      > When it is cool, it has a low modulus of elasticity. It is easily
      > stretched.
      >
      > When it is hot, it has a high modulus of elasticity. It acts like a
      strong
      > spring.
      >
      > But its rest-length is always the same.
      >
      > If you try to design an actuator based on "Nitinol contracts when it's
      > hot", your maths will go wrong.
      >
      > > Also note that most nitiol wires only shrink about 4% in length
      > > when active.
      >
      > Hmmm. Once again, perhaps that's not the ideal way to look at it.
      From my
      > understanding of the specs:
      >
      > If you stretch a Nitinol wire by much more than 8%, you'll permanently
      > change its rest-length.
      >
      > If you stretch a Nitinol wire by 8% by tensioning it, you can heat
      it and
      > cool it and it will maintain its rest-length but you can only do
      that a few
      > times until it becomes permanently stretched.
      >
      > If you stretch a Nitinol wire by 5% or less by tensioning it, you
      can heat
      > and cool it lots of times (more than 100k???).
      >
      > But what if you stretch it by 5% then fix the ends and heat it? The
      tension
      > in the wire will now be much greater (because its modulus of
      elasticity has
      > increased). Will that permanently change the rest length? I think it
      might.
      >
      >
      > So should you be worrying about the not exceeding a maximum length
      (105% of
      > rest-length)? Or should you be worrying about the not exceeding a
      maximum
      > tension (that tension needed to expand cool wire to 105% of
      rest-length)? I
      > think it's the latter but I haven't seen any docs discussing it. Does it
      > matter? Well, it would if you've made a gripper and the gripper is
      holding
      > a very hard object. What if your walker gets a leg stuck? Should
      > Nitinol-powered limbs always be springy so they can't exceed the wire's
      > maximum tension?
      >
      > > which is really easy to do by running current
      > > through it,
      >
      > The large current is, of course, the huge disadvantage of Nitinol for
      > mobile robots. You might need 100mA per "muscle" and, with a 6 leged
      insect
      > with 2 DOF per leg, that's maybe 600mA on average. Plus you have to keep
      > the current running to maintain a stationary pose. What miniature
      cells can
      > provide 600mA? How long will they last?
      >
      > Secondly, Nitinol heats-up very much faster than it cools-down which
      makes
      > walking hard to control. My experience is that the wire's response to
      > temperature is non-linear: above a threhold, more current makes it
      hotter
      > (and so it takes longer to cool) but doesn't change the elasticity much.
      > for the fastest contraction, you need a big current but for the fast
      > cooling, you want just enough current.
      >
      > The ideal way to control a wire is to give it a big current to make it
      > heat-up quickly then, after a 200mS, reduce the current so it just
      > maintains its temperature above the threshold. That way, you can
      conserve
      > power and relax the muscle easily. But it's hard to know what is the
      right
      > maintainance current - a breeze will cool the wire. I guess you should
      > continuously measure its resistance (which will indicate temperature -
      > maybe).
      >
      > > Nitinol wire only expands because the metal leg acts as a spring to
      > > strech it out again, which is where most of the inefficienciers
      come in,
      >
      > you can use two wires in opposition so the wire doesn't have to fight
      > against the spring but it's hard to control them thanks to the very
      > different rates of heating and cooling.
      >
      >
      >
      > Has anyone experience of building an "insect" that carries it's own
      > batteries? How long was it between recharges? how many steps per second
      > could it achieve?
      >
      > At fist sight, Nitinol looks like a great solution for all sorts of
      robot
      > problems, especially for walking or hands, but I think the disadvantages
      > far outweigh the advantages.
      >
      > Does anyone have any mobile robot success stories?
      >
      > Peter
      >
    • Branden Gunn
      I agree that my description above is not fully correct for all cases, but for commercial muscle wire , I think it is appropriate. Peter makes some very good
      Message 2 of 11 , Sep 6, 2006
      • 0 Attachment
        I agree that my description above is not fully correct for all cases,
        but for commercial "muscle wire", I think it is appropriate. Peter
        makes some very good points (thank you) in his reply, especially about
        the current consumption, but I have some more to add.

        Nitinol as a material has alot of interesting properties, most notably
        it's "shape memory". Basically, nitinol has three critical
        temperatures, though I don't know offhand what he technical names are,
        but I'll call them T1 through T3. T1 is the coldest and T3 is the
        hottest. All three are "tunable" by changing the ratio of Nickel to
        Titanium in the Nitinol (when it's made).
        When you heat the material past T3, it will "save" it's shape in the
        crystal micro-structure of the material. In the case of commercial
        "muscle wire", the "saved" shape is a short, straight wire. Once the
        material is cool (below T1), you can so anything you want to it
        (anything that is considered a "reversible deformation"). In the case
        of our muscle wire, it means we can stretch it (4% to 5% is safe), or
        bend it. Any time you heat the material above T2, without going hotter
        than T3, the material will return to it's "saved" state. Once the wire
        has cooled below T1, it's shape can once again be changed at will.

        In the case of muscle wire, if you heat it up using electric current
        (above T2), it will return to it's saved state that was saved when it
        was made, which is a short straight wire. If you leave the electric
        current on it for too long, it will heat up past T3, it will "save"
        it's current shape, which may be slightly longer than it's original
        state, or might be bent. If you heat it up past T4 (melting point),
        the wire breaks.

        I agree that the major problem with nitinol is getting it to cool down
        after use, and the fact that you have to use current all the time to
        keep it at the right temperature. Peter makes a great point that the
        amount of current needed changes as a function of the current
        temperature, and isn't really linear at all.

        For those interested, I wrote a paper on shape memory alloys for a
        materials science class about 5 years ago. The paper isn't the
        greatest (or longest), but anyone who wants to can read it:
        http://gunn.ath.cx/projects/SMA%20Material%20Paper.pdf

        --Branden Gunn
        engunneer@...

        On 9/6/06, PeterBalch <PeterBalch@...> wrote:
        > > From: "Branden Gunn" <engunneer@...>
        > > Nitinol contracts when it's hot,
        >
        > Actually, I don't think that's the right way to look at it.
        >
        > Nitinol doesn't "contract", it changes its elasticity. Its rest-length
        > always remains the same.
        >
        > When it is cool, it has a low modulus of elasticity. It is easily
        > stretched.
        >
        > When it is hot, it has a high modulus of elasticity. It acts like a strong
        > spring.
        >
        > But its rest-length is always the same.
        >
        > If you try to design an actuator based on "Nitinol contracts when it's
        > hot", your maths will go wrong.
        >
        > > Also note that most nitiol wires only shrink about 4% in length
        > > when active.
        >
        > Hmmm. Once again, perhaps that's not the ideal way to look at it. From my
        > understanding of the specs:
        >
        > If you stretch a Nitinol wire by much more than 8%, you'll permanently
        > change its rest-length.
        >
        > If you stretch a Nitinol wire by 8% by tensioning it, you can heat it and
        > cool it and it will maintain its rest-length but you can only do that a few
        > times until it becomes permanently stretched.
        >
        > If you stretch a Nitinol wire by 5% or less by tensioning it, you can heat
        > and cool it lots of times (more than 100k???).
        >
        > But what if you stretch it by 5% then fix the ends and heat it? The tension
        > in the wire will now be much greater (because its modulus of elasticity has
        > increased). Will that permanently change the rest length? I think it might.
        >
        >
        > So should you be worrying about the not exceeding a maximum length (105% of
        > rest-length)? Or should you be worrying about the not exceeding a maximum
        > tension (that tension needed to expand cool wire to 105% of rest-length)? I
        > think it's the latter but I haven't seen any docs discussing it. Does it
        > matter? Well, it would if you've made a gripper and the gripper is holding
        > a very hard object. What if your walker gets a leg stuck? Should
        > Nitinol-powered limbs always be springy so they can't exceed the wire's
        > maximum tension?
        >
        > > which is really easy to do by running current
        > > through it,
        >
        > The large current is, of course, the huge disadvantage of Nitinol for
        > mobile robots. You might need 100mA per "muscle" and, with a 6 leged insect
        > with 2 DOF per leg, that's maybe 600mA on average. Plus you have to keep
        > the current running to maintain a stationary pose. What miniature cells can
        > provide 600mA? How long will they last?
        >
        > Secondly, Nitinol heats-up very much faster than it cools-down which makes
        > walking hard to control. My experience is that the wire's response to
        > temperature is non-linear: above a threhold, more current makes it hotter
        > (and so it takes longer to cool) but doesn't change the elasticity much.
        > for the fastest contraction, you need a big current but for the fast
        > cooling, you want just enough current.
        >
        > The ideal way to control a wire is to give it a big current to make it
        > heat-up quickly then, after a 200mS, reduce the current so it just
        > maintains its temperature above the threshold. That way, you can conserve
        > power and relax the muscle easily. But it's hard to know what is the right
        > maintainance current - a breeze will cool the wire. I guess you should
        > continuously measure its resistance (which will indicate temperature -
        > maybe).
        >
        > > Nitinol wire only expands because the metal leg acts as a spring to
        > > strech it out again, which is where most of the inefficienciers come in,
        >
        > you can use two wires in opposition so the wire doesn't have to fight
        > against the spring but it's hard to control them thanks to the very
        > different rates of heating and cooling.
        >
        >
        >
        > Has anyone experience of building an "insect" that carries it's own
        > batteries? How long was it between recharges? how many steps per second
        > could it achieve?
        >
        > At fist sight, Nitinol looks like a great solution for all sorts of robot
        > problems, especially for walking or hands, but I think the disadvantages
        > far outweigh the advantages.
        >
        > Does anyone have any mobile robot success stories?
        >
        > Peter
        >
        >
        > Visit the SRS Website at http://www.seattlerobotics.org
        > Yahoo! Groups Links
        >
        >
        >
        >
        >
        >
        >
      • PeterBalch
        Branden ... It looks like you are much more of an expert than me. I ve read that description before in just about every discussion of memory metals ( your car
        Message 3 of 11 , Sep 7, 2006
        • 0 Attachment
          Branden

          > When you heat the material past T3, it will "save" it's shape in the
          > crystal micro-structure of the material. In the case of commercial
          > "muscle wire", the "saved" shape is a short, straight wire. Once the
          > material is cool (below T1), you can so anything you want to it
          > (anything that is considered a "reversible deformation"). In the case
          > of our muscle wire, it means we can stretch it (4% to 5% is safe), or
          > bend it. Any time you heat the material above T2, without going hotter
          > than T3, the material will return to it's "saved" state. Once the wire
          > has cooled below T1, it's shape can once again be changed at will.

          It looks like you are much more of an expert than me. I've read that
          description before in just about every discussion of memory metals ("your
          car will be able to straighten its bodywork after a dent") but it didn't
          seem to make sense when I played with Nitinol wire. The description makes
          it sound like Nitinol elastically deforms over a greater range of strains
          at high temperature than it does at low temperature; so a low-temperature
          plastic deformation becomes an elastic deformation when the metal is
          heated.

          You'd expect that if you bent a loose piece of wire (while cold) then
          heated it, it would straighten. Or if you stretched it (while cold) then
          released it, then heated it, it would shorten. But I never saw any movement
          of loose wire.

          The only effect I ever saw was a change in the modulus of elasticity. It
          wasn't at all what I was expecting. It simply doesn't match the description
          I've read in books.

          Either I'm missing something important or Nitinol is a special kind of
          memory metal.

          > http://gunn.ath.cx/projects/SMA%20Material%20Paper.pdf

          An interesting paper. There are a couple of things I don't understand.

          As you say, there are three states for Nitinol:
          (S3) very hot = soft: to make it "learn" its shape
          (S1) cold when it can be deformed
          (S2) hot when it returns to its "learned" shape

          Out of interest, is the Martensite-Austensite transition between S1 and S2
          or between S2 and S3? Martensite-Austensite suggests that Nitinol has two
          states but we know it has S1, S2 and S3.

          Is the hysteresis you describe important when you use Nitinol in robotics?
          Is the hysteresis between S1 and S2 or between S2 and S3? If it's between
          S1 and S2 then it would make it very hard to use Nitinol in a servo loop.

          Over what temperature range is the S1-S2 transition? Your diagram makes it
          look quite sudden. That would mean that you couldn't assume that a
          particular current would move your robot actuator to a particular position:
          even a small breeze or change in room temperature would make a big
          difference in position.

          > > Has anyone experience of building an "insect" that carries it's own
          > > batteries?
          > > Does anyone have any [Nitinol] mobile robot success stories?

          Complete silence!

          Should we all just file Nitinol in the "interesting but useless for robots"
          category?

          There's an interesting article in this week's New Scientist about using
          polymers for robot muscles.

          Peter
        • eshezo ayomi
          hi all iam an electronic engineering 4th year working in my final project it is 6 legged robot but i have some questions please ... and please if some one has
          Message 4 of 11 , Sep 16, 2006
          • 0 Attachment
            hi all iam an electronic engineering 4th year working in my final project
            it is 6 legged robot but i have some questions please"
            --- use a servo motor or the step motor
            --- use optical encoder or potentiometre


            and please if some one has an experiance in robotics ,please tell me as i have more questions need answers





            thanx before


            cccccccccccccccccccccccccc
            c ccccccc ccccccc c
            c ccc ccc c
            c ccc c ccc c
            c ccc c
            c ccccc c
            c c c c
            c c c c
            c cccccccccccccccccc c
            cccccccccccccccccccccccccc

            c u
            in good health

            ---------------------------------
            Get your email and more, right on the new Yahoo.com

            [Non-text portions of this message have been removed]
          • Howard Toburen
            Your Question Is Not Clear To Me. Are You Talking About The Six Legges Crawler Robot Kit From Parallax? [robo_seeker@yahoo.com] wrote: hi all iam an
            Message 5 of 11 , Sep 17, 2006
            • 0 Attachment
              Your Question Is Not Clear To Me. Are You Talking About The Six Legges Crawler Robot Kit From Parallax?

              [robo_seeker@...] wrote:
              hi all iam an electronic engineering 4th year working in my final project
              it is 6 legged robot but i have some questions please"
              --- use a servo motor or the step motor
              --- use optical encoder or potentiometre


              and please if some one has an experiance in robotics ,please tell me as i have more questions need answers





              thanx before


              cccccccccccccccccccccccccc
              c ccccccc ccccccc c
              c ccc ccc c
              c ccc c ccc c
              c ccc c
              c ccccc c
              c c c c
              c c c c
              c cccccccccccccccccc c
              cccccccccccccccccccccccccc

              c u
              in good health

              ---------------------------------
              Get your email and more, right on the new Yahoo.com

              [Non-text portions of this message have been removed]
            Your message has been successfully submitted and would be delivered to recipients shortly.