DRO For moving the head

The resolution will kill you. You are essentially measuring angular rotation as determined by movement around a circumference. For any angular movement, the movement on the circumference is proportional to the radius of said circumference. A sensor mounted on a 2.5" radius will have to be four times more accurate than the accuracy at the spindle, 10" out.

Linear scales do not use a single point for sensing. They have a ruled scale with regular spacing and the sensor picks up a change from light to dark, low capacitance to high capacitance, etc. to generated pulses which are then accumulated to determine position

I have designed an optical system for homing my CNC mill which is repeatable to +/- .0001" and could be modified for your needs. I used an Omron OPB829DZ optointerupter, combined with regulated emitter LED current, comparison to a regulated reference voltage via an LM311 comparator, and well designed optics. This provides consistent positioning to within +/- .0001", the resolution of my Tormach DRO.

The Omron optointerupter has a narrow window for better collimation of the led light beam. My design used a sharp knife edge to block the beam, generating the signal. This resulted in a steep voltage ramp as the knife edge moved to block the beam. This voltage was then compared to a stable reference voltage to generate the trigger signal.

You still have the issue of magnification of error so your ultimate resolution may only be +/- .0005" at best and possibly worse.
 
mickri said:
I have been reading about encoders. Rotary, linear, absolute, incremental and on it goes. For someone who knows nothing about these gadgets your head spins rather quickly. The bottom line is that encoders shine a light through glass that has transparent and opaque sections in known positions with a light sensor on the other side of the glass. As the light source moves the sensors pick up the movement and somehow calculate how far the light has moved. Most interesting to me was that absolute encoders can tell when you are back to the starting point. It seems like someone should be able to make one of these things that can tell when the light source is lined up with the light sensor. Should be easier than trying to calculate how far the light has moved.

Looks to me that this would work. Someone just has to make it.
Click to expand...

An incremental encoder like the Omron E6B2 series has reference point for obtaining absolute position but you must rotate to that position to re-reference the encoder. An absolute encoder has an additional set of scales so it always knows its exact position. The benefit of the absolute encoder is that you can remove power and repower without having to home the encoder.
 
Just had another thought. I know, it is always dangerous when I have a thought.

How about attaching a laser pointer to the top of the machine head, have it pointed across the room in the longest direction possible. Put a mark on the wall where the dot hits, now when you move the head just bring the dot back to the same point on the wall, and the head should be in the same position it was before in both rotation and Z axis.

Depending on the quality of your laser pointer and how well you mark the spot on the wall, you may actually get pretty good resolution, especially if the wall is a long distance away.
 
Thanks RJ for sharing your knowledge about this. I now understand the concept but no way could I ever make something like this. This is one of those things that is simple to imagine but may be difficult to achieve. Or it may not. I don't know. My idea doesn't measure the distance that the head may be moved. It tells when two things are lined up. When they are lined up all that needs to happen is for a light come on.

Flyinfool people have been using lasers like you describe for years. For that to work the mill/drill has to be fixed in place. My mill/drill is on casters so a laser won't work. It moves around. Currently I use a DI on the spindle like I posted above. It works. I was just thinking that if you have a DRO it shouldn't be a big deal to add a couple of sensors for realigning the head. And it could be a stand alone unit for people who don't have a DRO.

And it depends on the level of accuracy you need. Lot easier to achieve .001 than to achieve .0001. If all you need accuracy wise is .001 you could achieve that with a vernier scale. One part fixed to the head and the other part clamped to the column. Line them up. Move the head. Then bring the head back to where they are lined up again.

This was just a thought on my part.
 
Flyinfool said:
Just had another thought. I know, it is always dangerous when I have a thought.

How about attaching a laser pointer to the top of the machine head, have it pointed across the room in the longest direction possible. Put a mark on the wall where the dot hits, now when you move the head just bring the dot back to the same point on the wall, and the head should be in the same position it was before in both rotation and Z axis.

Depending on the quality of your laser pointer and how well you mark the spot on the wall, you may actually get pretty good resolution, especially if the wall is a long distance away.
Click to expand...
Expanding on this thought. I did the following. I dulled the edge on a utility knife blade and set it so the blade was collinear with the axis of my RT at a distance of 12 ft. I focused my laser center finder at that distance and mounted it in a vise on my RT. I then rotated the RT, observing the angle at which the laser reflection from the dulled knife edge was brightest. The angle was repeatable to about 2 minutes of angle.

2 minutes of angle equals .58 mrad ( mrad is 1/1000th of a radian which is 57.295º. The angle in mrads times the radius of a arc subtended by that angle is equal to the length of the arc in mils and for small angles the length of the chord of that arc. This is the error in linear distance in the x direction at the spindle axis. For a 10" distance from the column center to the spindle axis, the .58 mrad angle would correspond to .0058".

This was a quick and dirty experiment and could possibly be improved upon with some refinement. I would expect that getting down to a .001" repeatability might be difficult though. Laser beams have inherent divergence so little is gained by moving the target further away. A laser beam at 30 ft has about twice the beam width it has at 15 ft.
 
You must log in or register to reply here.
Back
Top