Re: [CAD_CAM_EDM_DRO] Re: A "simple" Hexapod project - Joints

Stephan,

I agree that Stewart platform machines are larger than xyz machines for a
particular working volume, especially if the y and z ranges are smaller
than the x. If we compare a SP machine to a Bridgeport knee mill design,
it seems that as the y axis travel increases, the mill must get heavier
(and a bit larger) to handle the greater moments coming from the same
forces expressed on the deeper throat.

I also found, like you, that range of movement in the joints was important,
and is a limiting factor on the working volume. I studied universal joints
available in catalogs, as well as ball joints, but didn't find pre-made
joints that offered enough range of movement. That's why I went to
gimbals. It is also why the platform gimbals are built of 1/2" stock, and
the portion of the leg that attaches to the gimbal is similarly built of
1/2" bar. Going to a wider material for either forces the gimbal to scale
to be longer for the same angular movement.

Another design parameter that impacts machine size is the ratio of the leg
at its longest to the leg at its shortest. Larger ratios give more
positioning flexibility, and smaller machines. For my leg design, the
ratio is less than 2:1. Some designs have a ratio greater than 2:1.

Stewart platform machines, or the tri-glide machine you are building,
require complex design tradeoffs. I find that to be part of the pleasure.

-- Carl

At 09:09 AM 6/27/2003 +0000, you wrote:

>I have done the inverse kinematics for a triaglide system (general
>closed form) as it is fairly simple and just a 4th order polynomial
>to solve.
>A couple of comments:
>These machines are fairly large i.t.o space occupied for a similiar
>xyz machine. So you are trading floor/table space for building cost
>
>2 things play a dominant role, working volume and joint range.
>As the working volume is irrelevant in the x direction (you simply
>add more rail length), the problem is deciding y and z ranges and
>these are interrelated. If you want a wide y with a shallow z, to
>minimise the length of rail, you need a tall machine. For a deep
>working volume you need to increase the x direction.
>More important, if you choose to use ball joints, the limit of the
>sideways range will decide your minimum leg length. Plus you would
>like some clearance above the spindle holder.
>Another problem creeps into the system. If you want to minimise rail
>movement, you have to operate with the legs as near to vertical as
>possible, but this creates the problem that your sideways stiffness
>is at a minimum. Conversely if you operate the machine with maximum
>sideways stiffness, your vertical stiffness is at a minimum.If you
>operate the machine with both vertical and horizontal stiffness,
>your machine size is at a maximum.
>
>another thing to look out for. Since this is a non linear movement
>machine, there are points where a small movement in the rail can
>cause a large movement in the tool position, So the resolution is
>not equal over the entire working volume. Steppers would not be the
>best of choices.
>
>Despite these difficulties I am working on mine and will keep you
>informed
>
>Stephan