Re: Re: SS 2009 good batch?
On 25.3.2011 23:21, ray_harlan wrote:
> Tapio, how do you test your rubber? Windup or stretch? Details.
First of all, I test full 30 gram F1B motors. By stretching, the method
is adapted from Bror Eimar, but to my understanding originates from Fred
Pierce. I pull the motor to estimated 90% of breaking load, then unload
and record the pull force as a function of stretch. Summing these
measures integrates the force over stretch, giving energy return.
The original Eimar method used a lookup table to find the load/stretch
combination for 90% of breaking load (the longer the motor is stretched,
the lower the force. I found this complicated to use, so it was one
major reason to build a computerized system. While unloading, Eimar took
readings at steps of 10 or 20 cm, and again I felt that this number of
datapoints is too coarse to get good results.
I built the system almost 15 years ago. Did not know how to use
microcontrollers then, so I used a computer mouse as the
interface/AD-converted. The force is measured by a large spiral spring,
that stretches about 3in under the full load of about 45 kiloponds
(100lb). The movement of the spring goes around a pulley, and this
rotational movement is taken via a rubber belt to one of the tiny axes
of the mouse. So force = mouse movement. The other axis of the mouse, or
rather the optoforks measuring the movement, are moved to a boat winch,
whose movement (rotations) are measured. My software compensates for the
spring stretch shortening the rubber motor, and the changing diameter of
the winch drum, as the belt is wound around it.
The Eimar tables used the motor unstretched length, stretch factor and
the pull force. In practise, the unstretched length is not needed, as
always the unstretched times stretch factor, i.e. the stretched length
is used. So the table sums up into a simple equation, for a given
stretch it calculates the force that equals the target load on the
motor, and my software compares this to the measured, and instructs me
to keep stretching the motor until the target is met (with more stretch
the target force is reduced and actual force on the motor increased).
Reaching the target, I start unloading, and the software measures the
force and stretch, stores force at steps of 1cm, and sums these up to
integrate the energy return.
I do not break motors to find the actual breaking load for each batch,
but assume this to be constant. I have been considering this, but ended
up to the way I do it for a few reasons. First of all my rig measures
only full F1B motors, so it would waste a lot of rubber to do
destructive testing, and further I do not know if the rig would stand a
repeated load from breaking motors. Second, I consider that when using
the motors I wind them to more or less the same load/torque anyway, so
this assumption is more in harmony with my actual use of the motors.
Sort of test the motors for their intended use, not for maximum
potential energy return. On the other hand, I think the difference is
slight.
My testing suggest very small variation of energy return within a batch,
or at least within the box I have tested, even when I cross a boundary
from one rubber sheet to another. Therefore I feel that testing
individual motors is futile. However, I get the stretched length for
each motor to use as reference for expected maximum turns. Also I get
some nice comparison data for different batches. I am convinced that my
measurements are valid, as even though the energy return within a batch
stays constant, for different batches tested on the same session I get
different readings, so my rig is NOT producing just similar figures for
the same session!
I also use the same rig to break in the motors. This time 80%, keep the
motor under stretch for 4 minutes, slightly increasing the stretch while
the motor breaks in and the pull force is reduced. Then let the motors
rest until the next day before the test.
-Tapio-
Received on Fri Mar 25 2011 - 22:47:36 CET
This archive was generated by Yannick on Sat Dec 14 2019 - 19:13:46 CET