Laser power variance within cutting area, and over tube life



I was thinking about the comment from @fablab_elpaso in another thread:

You also lose power as the beam passes through the air from one end of the bed to the other. So, maximal power is usually at top left of the work area (where beam path is shortest) and drops off significantly at bottom right (where beam path is longest).

Is this variance likely to be of an amount that will make a visible difference in a piece of work, and if so, can it be automatically taken into account by the Glowforge software when figuring the tool paths?

And on a similar note, one would think that as output power may decrease with tube life, some sort of calibration would be required to get consistent results between pieces engraved on a new machine vs. pieces engraved on an older one.


Never really thought about it since my laser efforts have always been limited to a small place on the bed. Interestingly because the tube is mounted on the moving gantry the beam only has to travel maybe 60% of the maximum distance than that of a typical laser with a similar sized bed. (one axis is fixed and on a typical laser cutter the tube is WAY in the back) So if there are noticeable effects it should be less on a Glowforge.


Yup… They probably managed to eliminate some of that power drop by moving the laser with the gantry. Really cool design feature! Not sure why other lasers don’t do that, but might have something to do with: 1) accelerating/decelerating more mass; 2) not wanting to move a breakable piece of glass all over the place at high speeds; 3) safety concerns with flexing a cable carrying kVs!. In any case, if GF have figured out how to improve the flying gantry config then it should be a much less noticeable drop in power across the y-axis of the bed. And if they’ve managed a fully enclosed optical path instead of only enclosed optics, then it should pretty much eliminate that draw back of these kinds of systems.


Can be done on the user side by changing your cut parameters when you notice you’re not getting through the material anymore. Fixing it in software would be awesome but no two tubes will age the same and modeling that power drop as a function of cutting time would require burning through a LOT of tubes.


Lasers do not suffer from the inverse square law, and distance is essentially pointless.

Yes, there are particles in the air, but not many, and most not relevant at this wavelength.

So, distance across the bed does not matter in terms of cutting power. Unless you have a TON of smoke in there, which the ventilation really should make impossible. But smoke is not something which can be compensated for in the software (in case you do find some application which reliably floods the chamber with pea-soup levels of vapor).

It is true that even lasers will suffer from the inverse square law eventually. So if you search around on the internet to try and verify it may get confusing. The point at which the inverse square law “kicks in” for a laser is when you are so far away from the origin that the light is no longer travelling in nearly parallel paths. ie - the beam is diverging just like a normal light source would.

Take a flashlight, point it at your hand a few inches away. Really bright. Now point it at a wall a few feet away, not as bright. Now do the same with a laser pointer. On hand, very bright. On wall… just as bright. We are using the laser pointer, and working in this range of distances.


It’s not just a laser. It’s a laser bouncing off no less than three mirrors and possibly passing through multiple housings. And yes, the relatively small amounts of smoke and soot do matter. That’s why the enclosed optical path is so appealing. Anyone who has used a flying optics laser cutter can tell you there is some power drop across the bed and will plan their cuts accordingly (like using deeper cut parameters for designs spanning the entire bed vs. restricted to the near side of the bed). It’s a real (albeit minor) issue to deal with. Hopefully GF has solved it!


Maybe that has more to do with design (mirror alignment and focus) and not the power. If you are planning for it, it sounds like it is a known thing for a given unit (design flaw).

I have seen tests proving that is has very little to nothing to do smoke and everything to do with bad mirror alignment and focus.


That could very well be the case! Clean well aligned optics are the biggest factor by far. And I heard mention somewhere that mirror alignment would be much easier on the GF than your typical hobby desktop cutter. One way to get around power drop is by brute force. Typically you want to cut as fast as possible with the lowest power that will get through the material. There are various speed/power variation tests that can be automated to get the ideal settings. Then you just bump the power up well above what is needed to get through at the near end, so that it cuts well anywhere on the bed. But that needlessly wastes power and may drain your tube faster. Of course, with GF moving a lot of the speed/power parameters to tried and tested values raining down from the cloud, it will probably be a moot point if you’re using GF sourced and approved material.


This is being discussed in another post with some pics of quick and dirty test results.