Alternative positioning idea.

[ynneb]

I am not sure where to put this post so please forgive me if its in the wrong forum.

I have two ideas that I have wondered if others have done or considered.

The first idea was Laser positioning feed back. By this I mean the traditional way of knowing where the router/cutter is, is by either stepper motor drivers counting the pulses, or servos using encoders to report the position. My idea ( It may be already thought of) is to have a laser that is in a fixed position. It shines on the gantry (for arguments sake) and refelcts back to a reciever. From this it calculates the exact position of the gantry.
I am aware of lasers that can be shined upon glass and record conversations in a room for espianage work. This tells me that if the laser feed back can decode this information from the small movements of the glass vibrating, then surely it could accurately record the postion of a ganrty. Surley it would be more accurate than any current method?

I can think how it would also solve many current issue that haunt our current methods. There would be zero issues with backlash because the position of the gantry would not be dependant on where the driving belt/screw was. It would rely on the actual position of the gantry or whatever needed to be monitored.

EDIT: I just thought about the limitations of computer speed would also not be as much as an issuse as before. At present to get great accuracy you must have a high pulse count. With my method the whole count issue is redundant. All the computer would have to do it know where you are by the laser reporter. I can imagine this would need new software to be written.

Because it relys on real positioning technology that would mean the current stepper/servo technology would now be redundant. We could have very cheap AC or DC motors doing the work of their expensive dinosaur cousins. It would make ball screws redundant because the position would not rely on them. A cheaper driving method could now be used. Just imagine something like traditional sprocket and chain driving our machines. I can imagine this sort of technology turning our industry upside down. (Sorry now I am getting excited)

Has this been done b4? If not, I know I dont have the technical knowledge to exploit this idea. If any one wants to run with this idea just remember me and lavish me with a few freebie positioning sensors. he he ( I am serious though )


Now for the other idea ( not so exciting )
I am building a 8 foot long rotuer table. The only way I can think to drive this is with timing belt. This is still not as cheap a method as it should be. I have an idea that harks back to the old days of the transitor radio. Remember how when you used to turn the dial tuner and it would position the needle pointer to show where you were tuned? It was done with string. The reason the needle didnt slip was because it was fixed to the string. Also the knob that was turned had a spiral drum that the string was wrapped around. The string was also fixed in the middle of the drum so as to avoid slippage. Ok Ok I know you know what I am going to suggest. Well with all out fancy technologies like kevlar etc, surely we could use a similar setup to drive our machine tools?
Lets say a 5 mm diameter piece of kevlar string has a breaking strain of 1 tonne, just imagine the possibilities. Again if you want to run with this and its a winner remember me. and give me some freebies.


Your feedback on this would be appreciated.

[ballendo]

Hello,

Got some bad news for ya, I'm afraid...

Using the laser IS done, but for setup only as the machine dynamics create too much movement and vibration for it to work during cutting(interferometry), and the reflective type you mention is no good at longer distances...

Next, cable drives have a long history, and are still used in the French equivalent of the "shopbot". In other words, I've found that most hobby cnc in any given country seems to be largely similar, and the French "version" of "hobby cnc" uses roller bearings on square shafts and flat surfaces, with cable drive...

(The early shopbots also had cable drive, but are now rack and pinion.)

Early plotters have cable drive as well, and there is a small 3axis pcb drill by milford instruments that uses it. BUT...

Cable drive is the least repeatable of many methods and gets worse over time and over long travels. The French "get it right" and do some of the necessary things to make it as good as it "can be". But it will never match chain, rack/pinion or screws.

Hope this helps,

Ballendo

[ynneb]

Oh well. It was good for 10 minutes thinking I might have poineered some great new idea.

You mentioned chain as a method of drive. I havnt seen that mentioned in any of the posts ( I am very new here )
Is chain ok? Would that be viable on a 8 foot table? How would that compare to timing belt? Im my mind i like the idea of chain better, been steel and all.

[NeoMoses]

Although some of your thinking on position feedback is sound, it does not eliminate the need to minimize backlash in an assembly. When machining, cutting forces vary and any backlash will allow the tool to 'wander' thus cause at a minimum gouges in the workpiece, possibly broken parts and/or tools, depending on the amount of backlash.

If you have elminated virtually all of the backlash in a system, the next logical step to improve its accuracy is to put the position feedback on the table, not the screws. If you look at most of the professionally made machines, they are done this way, usually with linear position encoders that are very accurate and very pricey.

As for the drive on your 8' router... a tensioned timing belt should work fairly well, but don't expect to be machining at 200+ IPM with it, the accel/decel will definately cause some stretching over that length. Other options, such as screws, are available, but at a higher cost. Take a look at the Cad Cut CNC router, they have a good drive method, too. Basically, with such a long span, you'll want the cross sectional are of any cable or belt to be maximized to reduce the stretching. Do some quick F=m*a calculations to see what your max acceleration forces should be, then size your cable/belt accordingly to give you the accuracy you desire.