Range of Motion

In the server repository there is an algorithm to measure the range of motion openflexure-microscope-server!335 (merged) . Running this on microscopes has shown clearly that the limits to the motion are different for screwing the actuator column or down, and that the centre of the range of motion is not the same place as the top of the 'sphere cap' motion of the stage caused by the parallelogram flexure mechanism. At first sight this would be related to #403 (closed), that the column motion hits a hard stop when the actuator column hits the inside of the top of the actuator column, but there is no hard stop when the actuator column goes down into the foot. However it was also noticed that the sample clips can limit the motion by colliding with the illumination dovetail (not colliding with the illumination arm).

I set up a microscope with a ball-point pen in a location jig in place of the illumination condenser lens:

Pen-microscope

A piece of paper on top of a slide is held firmly enough that it does not move on the slide when the stage is moved under the pen.

I used a microscope body without assembled actuators, so that I could push the stage around, either by pushing the actuator levers up from the base of the microscope, or by pushing the stage top itself. Holding feet onto the microscope indicated that bridges inside the feet do not provide an end stop to the travel.

With the sample clips in the design position the motion is clearly not symmetrical about the neutral position, and is not square.

ROM_with_clips_annotated

At the bottom of this image the stage is moving away from the actuators, and there are straight lines at right angles to each other, where the actuator columns hit the upper limit for the x and y axes. The distance is about 8mm from the central relaxed position.

At the top of this image the stage moves towards the actuators and the sample clips hit the illumination dovetail. The distance is about 6.3mm from the central relaxed position.

At the sides towards the top of the image the stage legs close to the x and y actuator columns hit the side of the casing of the z-axis. The distance is about 6mm from the central relaxed position.

If I position the sample clips out of the way on the 'wrong' side of the stage top (as in the photo showing the pen setup), the motion is:

ROM_moved_clips_annotated

This is the same story, but with more motion towards the illumination.

After initial discussion with @williamwadsworth @rwb27 @julianstirling @Chish36:

  • We could curtail the motion in software to a square within this range of motion. The design motion is 6 degrees, and the legs are 65mm high, so that is 6.8mm each way from centre so we would be reducing the motion to smaller than the design motion.
  • We could measure the range of motion, either generically as the microscopes are of known geometry or measure as part of initial calibration, and then apply limits in software that reflect the design 6 degree deflection and the additional limits of the parts touching.
  • We should at least make the hardware limits in places where driving the actuator further does not put stress on the flexures. Practically that means that the limits should be at the top of the actuator column casing and in the foot.
  • We could adjust the leg or z-axis design to ease those limits to be beyond the 6 degree design limit.
  • We should re-design the sample clips so that they are not a limit.