What kind of sensors could I use to determine if two plates were parallel to each other and secondarily the distance and angle between the center points of each plate? The plates can move similar to a hexapod or Stewart platform. The distances between the plates will be 50mm to 300mm.
What accuracy is required?
What's the budget?
Four rangefinders on the corners?
"The plates can move similar to a hexapod or Stewart platform"
Got a link? Is the middle-to-middle distance fixed?
Budget: <$300 for the sensors.
Accuracy: If the plates were 150mm square, then parallel within .01 mm from side to side would be a good target.
Similar to a hexapod, the middle to middle distance will not fixed. The biggest difference is the arms holding the platform will be more flexible than a hexapod. The device doesn't exist yet, and nothing closer than a hexapod exists. The movements will be relatively slow, so the sampling rate can be low.
I'm mostly interested in discovering which technologies might work.
With precision 10 micron, probably, only laser, but you have to make plates reflective.
If the plates are not joined together then I think you will struggle to do this with the level of accuracy you want.
Assuming the center of the plates are the center of rotation, that is the centers are always perpendicular:-
The best approach might be a laser bouncing off one plate back into a linear CCD on the other plate. Then you see which cell of the CCD is illuminated to get the angle between the plates. You will need five of these, one in each corner and one in the center. The center on gives the distance which is then used to calibrate the returned angle from the corners.
Practically I don't think this is doable.
What about two lasers in adjacent corners aimed at a reflective target in the centre of the other panel.
The beams bounce off the target and land on linear CCDs or some such on the opposite corners of the originating panel. The location of the brightest CCD cell gives you the angle of the panels.
EDIT: Dang, Mike beat me to it.
Rob
The center of both platforms will be occupied by the payload so bouncing lasers off the center won't work. What kind of accuracy are other technologies capable of?
The two plates might act as a capacitor and the force is a function of the distance, but I doubt if that would be precise enough.
What about gyroscope sensors ? - http://www.sparkfun.com/tutorials/167 - (second half )
The center of both platforms will be occupied by the payload
Maybe the same idea but along the edges.
Rob
elf:
What kind of sensors could I use to determine if two plates were parallel to each other and secondarily the distance and angle between the center points of each plate?
If your intention is to move/control the platform with steppers, I would not look for sensors, but rather base it on a model of the platform dynamics. Positional accuracy and awareness would then be limited only by stepper angle resolution. Working out the math is not trivial, but then again this would be a core challenge and most likley a requirement of such a project.
If you need to relate the platform to life coordinates (gravity) then a 3D accelerator could be used. They typically have a 12-bit resolution across 180 degrees (approximately 0.1 degree).
In the precision optics field $300 is the loose change you find down the back of the sofa.
What is the projected cost of the entire project including hexapod hardware?
At this stage of the project, there is no budget. I can build at least 90% of it with left-overs from other projects. I haven't used any position sensors before, so don't really know what all of the possible technologies and their tolerances are.
robtillaart:
The two plates might act as a capacitor and the force is a function of the distance, but I doubt if that would be precise enough.What about gyroscope sensors ? - http://www.sparkfun.com/tutorials/167 - (second half )
Navigation gyroscopes might work but rate-gyros (which is what MEMS technology electronic gyros all are) won't work as they only measure rotation rate, not angle. The Sparkfun article shows ignorance of real gyroscopes (which hold an absolute direction in space to great precision over long time periods). Real gyros are very expensive.
For low accuracy 3-axis accelerometers on each plate with work if they are stationary when readings are taken.
For higher precision some sort of physical measurement probes are probably needed to measure a number of distances between the two planes, possibly with linear encoders.
Thanks, everyone. I'll experiment with accelerometers to see if they will be accurate enough.