Printable View
What about the emission point accuracy.. That kind of says the beam isnt guaranteed to exit 100% perpendicular
Attachment 40576
Just a thought.. Everything has tolerances( my parts are +/-0.05mm ). I suspect this issue is just all the tolerances added together..
True. It could be. But there's only one way I can find out. About the tolerances, anyway. About the emission point, that's an angular measure, so it amounts to the wider deviation being a half-angle limit on what can be acheived for width/divergence compromise. Its proportions to total angle are fairly large, but the implications are that translation across axis alone won't fix it, you'd need to adjust pitch and yaw of the diode too, and I dread to think what that might mean for thermal coupling, not to mention alignment hassles. >:) Basically, I'm just hoping I find a compromise between performance and fast easy setup for a new diode, that I can live with.
"it seems unlikely that significant off-axis variation exists between diodes" , nope. Unfortunately they vary quite a bit. My largest combine is 20 diodes they all needed to be tweaked. My co-linearity between lens threads (tight threads) and diode seat is .0005" or .0127mm, I really sweat that operation and I still got quite a bit of variability.
chad
Well, Dan did say he had a sample size of one. :) If I have to deal with it I will, my main concern (apart from the compromise I mentioned) is that I don't want to add error to error cos that just makes a hard task harder still. I aim to create one point where most change occurs in one movable optic, where possible. If translation across lens axis becomes unavoidable, we might as well separate mounting of diode and lens and be done with it, that way we can get the angular correction done too, without compromising thermal coupling and having to do so much trial and error adjustment.
What is funny about all this is that the system I'm making my test rig from was a thing I once part-built to put the lens mount totally separate from the diode mount! I'd rejected it as a bit of stupid overengineering with too many spurious demands for accuracy. The funny bit is that at the far end of this experiment I may yet go full circle and do what I intended at the start, for that design. (And it is no accident that I chose that old platform for these new tests..)
Edit:
Chad, what method did you use to tighten the tolerance gap between threads? Dave mentioned teflon tape, and it's a good idea I hadn't thought of, but my method was previously to use a spring and a bit of kilopoise grease. When it settles (which is annoyingly not instant for fine tuning), it is accurate because the shape of the thread is self-centering in this condition. Given variation in the lay of teflon tape it likely isn't.
I machine my own threads and adjust the pilot hole to be a little smaller. The frustrating thing is all of the lens holders are a little different sized. It seems there isn't a standard m9x.5 standard they are all a little different when you are going for close fit. I also use a thick grease similar to what you use. Mine is vacuum rated and doesn't out gas and is good to a little higher temp.
chad
What name for that grease? I think Kilopoise (PFPE) does outgas a bit. I also think it oxidises given time, and goes yellow. It's good but I always thought there was better. About the M9P0.5, definitely, I've only seen a few and that fact jumped out starkly. But so long as diode mount hole and lens thread hole are cut with no change in centering, it reduces the number of unknowns a bit.Quote:
Originally Posted by chad
The other big source of error is in precise position of lens in barrel (especially when slightly different lens sizes are supported by one barrel design), and I'll be making tiny adjusters out of acetal rod and piano wire so I can see the beam constantly, uninterrupted, while rotating the barrel. Looking between adjustments takes a lot of deduction, but seeing it in real-time motion is an extremely fast way to gauge error sources. It will save far more time than I'll spend making adjusters. :)
can you please describe what you are doing with acetal rod and piano wire? :confused:
Thanks!:)
Quote:
Originally Posted by The_Doctor
Probably not a lot. :) With the acetal, anyway. Too much like hard work. Instead I'll use a bit of thin walled brass modelling tube about an inch long. It's already got the hole through the middle, and it's bigger, lets wider beams through. The idea now is to use quarter inch OD or the next size up if the fit will be better (doubtful), then bend a bit of piano wore around a 5mm drill bit to get a sort of hairpin shape with parallel straight parts that fit the tube OD with enough grip to hold it while I solder it to the outer wall on both sides, with the bight crossing the axis just beyond one end. After soldering, use a small fast grinding disk in a Proxon drill to cut and deburr the wire so it sticks out only 1mm beyond the tube end. The ends of the wires left behind will engage the slots in the lens barrel so it can be turned without interrupting the beam. Chamfering the tube rim to prevent strong reflections, and maybe painting the whole thing black will help too. Also wise: don't run lasers at full power while adjusting focus with it. Keep it under half a watt, or half power for any laser less than 1W power.Quote:
Originally Posted by LaserCo
My old mounts, sold by LasIrVis had a little focussing thing like this, all in one bit of plastic, but the above method seems to be the easiest DIY method that will get a good result. A similar trick using a metal washer might make one that gets into tight spaces.