Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[TSTwizby]

I wasn't sure whether or not to put this in with creative projects, and decided not to since I'm still a ways away from actually trying to build this.

I've got an idea in my head to design a laser cutter, with the major distinction between it and most others being that rather than using an X-Y table to direct the laser it would instead embed the laser in a sphere which would rotate in two directions and raise/lower itself. The main reasons for this are to make it more compact and to make some operations (rotations and scaling) much easier, from a computational standpoint.

The basic idea here is a laser in a hampsterball. It's sitting on a ring filled with ball bearings to keep it from falling, and there are a pair of motors which spin it around the vertical axis and which change the angle the laser points to the vertical axis, and one which slides the whole assembly up and down a pole. So far the problems I've thought of which might come up mostly involve the motors slipping, or the wires from the laser tangling up or something like that, but when I described this to someone they said that there would be problems with keeping the laser collimated at different incident angles. I've looked into this a little, and haven't been able to find anything relating how well collimated a laser is to its incident angle, but I'm not especially familiar with lasers and it's possible I misinterpreted what I read. Does anyone have a comment on this, or any other aspect of my design which they think would be problematic?

[Osmosis Jones]

Are you searching for the correct things? The term I think you're referring to is spelt "collimated" not "columnated".

Also, from what I understand of lasers (warning, not much), over the distances you're likely to be working, a gas-based laser should be fine. Apparently diode lasers have a much lower degree of collimation than gas due to the size of the lasing cavity, but I doubt you're using one of those given this is for cutting.

[TSTwizby]

Ah, sorry. I've been searching for the right, thing, the correct spelling just slipped my mind for a moment.

I'm still not sure exactly what kind of laser I will use, mostly because I'm having trouble finding any information on the subject. I don't need it to be especially powerful, though of course the larger the range of available powers the better. I'd be happy with something that can cut paper reasonably quickly.

[Osmosis Jones]

Hmm, if it's paper then you probably don't need to be concerned about collimation at all. I imagine the only reason you'd need to be worried about it is if you're trying to achieve an extremely high heat at the cutting point; necessary for metal, not so much for paper. Hell, a simple 1W pen-sized blue diode laser probably has enough oomph to cut paper; step that up to a small bench top laser and you should be fine. Burning might be an issue though.

Also, what is your planned working distance (how far from laser tip to surface), and how precise do you need the cut (+1mm, 0.5mm, what)?

[TSTwizby]

Mostly, I'm just building this to test the principal. I'd like for it to be able to etch/cut paper because that would make it useful to me as well, and of course the more things it can cut the better, but if it would take excessive work/money to do so then I'd be happy with the basics.

As for working distance, with the scale I'm currently looking at, it would need to work at maybe 10-50 cm. I'd like the cut to be as precise as possible of course, but again at this stage I'd be happy with an error less than a millimeter or so.

[Osmosis Jones]

Okay, assuming you're using something like the CO2 laser at the top of this page, for your laser, you're looking at a 1.5mm diameter spot, with a divergence of ~1 milliradian.

Taking the numbers you've given me for working range, assuming 100mm is minimum distance, we have a beam divergence of ~ 0.2 mm increase in the rated spot size from vertical illumination (laser is perpendicular to the surface). This goes up to ~1mm at 500mm, again vertical illumination.

If we're assuming that the 10cm is the height above the surface, and the 50cm is the maximum possible hypotenuse distance (so the laser will be scanned up to a maximum angle of 1.37 rad or ~43.6^o from the vertical), thats a little trickier to calculate. My original attempt at a calc for this gave weird numbers, so we'll try something else:

Simplifying it to a 45^o or pi/2 radian angle, you're going to get something like an outer radius of the angled spot being sqrt(2)*radius of the flat one. It'll be less for the inner radius, but we'll assume they're equal as the worst case scenario. So, a 1.5mm diameter spot will swell to ~2.1mm even if there is no beam divergence whatsoever. With the listed divergence, we're looking at a spot size as large as 3.5mm diameter, basically doubling the spot size compared to the vertical case.

As the angle increases, so too will your spot size. While divergence will have a larger effect in absolute terms with increasing angle, I *think* the fact that you're taking an angled slice will always have a greater influence, *especially* at higher angles.

So, in short, your beam divergence (which is what collimation is all about) shouldn't be a problem here. The big problem is gonna lie in the simple fact that you're taking conic slices (cylindrical in the ideally collimated case) at increasingly sharp angles. If you want to limit yourself to smaller angles though, this might work.

[TSTwizby]

That's one of the benefits of being able to move the thing up and down. By moving it up, I can reduce the angle needed to reach the edges of whatever I'm aiming at. I'm pulling these numbers mostly out of nowhere, but let's say I have a 30x30 cm base that I want to be able to aim at. If I have the laser sitting at 10 cm above this, then I'm going to need to hit angles up to the order of 65 degrees, which gives a pretty bad amount of spreading, by a factor of 2.4 or so. But If I raise it up to 50 cm, that cuts it down to about 25 degrees maximum, which is a much more manageable number (the spot increases by a factor of about 1.09). Over smaller distances, say a 10 cm radius circle with the apparatus centered on it, the size increase at the nearest distance is by about 1.23, with the increase at the largest being negligible. The only reason to move that close would be to do very detailed small-scale stuff.
So using the same laser you mentioned, we have a spot of about 3.6 mm in the absolute worst case, about 1.7 mm in the case with max height and greatest angle, with a range between about 1.5-1.9 in the small-area case with minimum distance. I'd say this is acceptable, though I'm not sure I actually know what I'm talking about.

For what it's worth, I wouldn't let this thing go at angles higher than 45 degrees in actual use. Too much of a risk that it would get jostled or something and end up shooting the laser across the room.

[Osmosis Jones]

Hmmm are you sure of those numbers? They might be better than you think... 

30x30cm base means a square with a max radius of 15*sqrt2 cm (~21.2cm) in the diagonal corner. At 10cm you're looking at an angle from the vertical of
tan^-1(10/21.2) = 0.44 rads or ~25^o by my calcs.

Also, don't forget to isolate the workbench from vibration. In a vertical burn, it's effect will be negligible, but if you're going up to 45^o, vibration will play havoc with the precision.

[TSTwizby]

I might not be right, but I'm sure you're wrong. Think about it; you have a right triangle whose base is 20 cm-ish, and whose height is 10. The laser is definitely not making less than 45 degrees to the vertical.

Thanks for reminding me about vibration, I hadn't thought of that.

[Osmosis Jones]

Herp a derp. You're right; I had tan as adjacent over opposite, not opposite over adjacent, which is 64.7 degrees. I blame the caffeine wearing off.

[TSTwizby]

Ah, it's late. Don't worry about it. I'm going to head off and get some sleep soon, thanks for your help!

[Osmosis Jones]

No worries, I needed an excuse to brush up on my trig any way

[RedKing]

After reading this thread, I just keep picturing this. 


EDIT: Link fixed. Stupid wandering cursor.

[Osmosis Jones]

A broken link? (You somehow inserted a into the url). But heh, dethklok are awesome.

[palsch]

I'd guess you could do it, but the loss of precision and consistency would be significant over a basic horizontal moving arm style setup. I'd worry that changing the angle and height would mean a very inconsistent energy delivery over a pattern, which may or may not matter that much depending on the material and overall laser power. Off the top of my head a 45 degree rotation would be a roughly 30% drop in power at the extremes, assuming the distance change is negligible compared to the angle effect. You may need to compensate with more sophisticated control than usual for your burn time, and that would be hard given the motor setup.

For any sort of precision (and to avoid cable problems) I'd certainly go for rotators rather than a ball and external motors. It's more guaranteed that there won't be slippage and it's easier to calibrate.

Not to mention, purely from a safety POV, that a free floating hamster ball would be a huge risk. If the laser is even capable of coming to the horizontal (or worse, above it) you would want the system well contained on top of basic precautions (effective eye protection, no reflective objects in it's firing range). The chances of it firing outside of your expected area means you need to make the entire room laser safe, or at least absolutely any theoretically possible firing angles. Having it more directly connected to a motor setup with a hard maximum elevation would vastly reduce the hazard.

You can probably build around a lot of these issues, but they are complications that reduce the viability of the build in my eyes.