Does Switching lenes cause miss alignment?

[Joseph Tovar]

Hi all,

I have a question about swapping lenses while doing a single project. For example, I use my 2.0 lens to raster acrylic for making a jig. When finished, I put in my HPDFO lens and want to raster the substrates that go in the jig. When I do this, I notice that the job is off by a few mm.

I just installed the HPDFO this weekend so I'm wondering if I did something wrong, or this is normal when switching lenses. If I create the jig with the 2.0, then do the job with the same lens, it lines up. I haven't tried making a jig with the HPDFO, then doing the job with the same one, but I assume it would line up also. It's just not working when I make the jig with one, and run the job with the other.

Anyone else ever seen this?

[Dan Hintz]

It sounds like your HPDFO may be in need of some adjustment... at the lens, both should line up perfectly. Several mm is huge from an alignment standpoint when you're that close to the lens (especially a 2.0FL lens)... something is really wonky. Are you sure the beam collimator is aligned properly?

[Scott Shepherd]

It's normal. I have two of the 2" ones. When I swap them out on the same job, it's does the same thing (I learned the hard way). I had a large job and one lens was brand new. I didn't like the amount of debris it was kicking up, risking my brand new lens. I thought I'd stop it, switch the lens and do another pass. Like you, it was off more than 1mm.

I don't know why it's not closer, but it's not. Dan, there's no adjustment on the collimator than I know of. It slides in, gets screwed into place and it's done.

[Richard Rumancik]

It does not surprise me - I mentioned the tolerances on optics mounting and axis errors in another thread recently. I don't know how well the manufacturers align the lens into a holder - ideally the holder is held on a datum surface and the lens positioned such that the beam "points" in an axis exactly perpendicular to the datum (holder) - then bonded into place in that position. But my guess is that they make a lens, make a holder, apply some glue, and bond them together, thereby locking in any errors of parts and process. (By the way - I have mounted my own lenses and did it basically that way -as I don't have the necessary equipment to do better. For the most part, it works well enough.)

Also keep in mind that the center of the lens can have some tolerance - ideally, your incident beam hits the lens exactly in the center but that won't always happen due to tolerances upstream. So it is not inconceivable to me that one setup could have an axis error of 1 degree one way and another have an error 1 degree the other way. This would be roughly equivalent to 2mm deviation over 2" focus. Adding the extra optics adds another tolerance.

For the most part we don't need to change lenses in the middle of a job so I suppose the manufacturerers don't see it as a problem. And usually if you did cut the outline with one lens, and raster (say text or graphics) with another, you might not even notice the misalignment. You have found a weakness that you probably just have to live with. Some of us are trying to use laser engravers to make precision parts but they were really designed for the awards industry. So you have to be aware of the limitations and find suitable workarounds. Laser engravers are not like machine tools which are built rigid, heavy, with defined mating surfaces between components, clamping of components to datums, etc.

[Joseph Tovar]

This kind of sucks. I was hoping their was something that I could calibrate to get both lenses to work together. I was told that the HPDFO is better for marking and getting small detail. I was hoping to use the 2.0 to cut the jigs and make the fixtures, then use the HPDFO to do the detailed rastering on the items. Since they don't line up, I can't do what I was planning...at least not as easily as I was thinking.

As for aligning the collimator, I followed the instructions....Remove the lens, put tape over the hole, adjust the screws in the collimator until the beam appears in the center. It says to get the beam as close to the center as possible.

BUT...If I change the alignment of the collimator, the change would affect both the 2.0 and the HPDFO lenses the same, so they'd still be out of sync.

[Dan Hintz]

Dan, there's no adjustment on the collimator than I know of. It slides in, gets screwed into place and it's done.
Steve,

It should have some adjustment screws on the collimator for alignment... you may not have noticed them as your was installed by a tech. Ask Paul for a copy of the install instructions and you'll see what I mean.

Is there any play in the mirrors/lenses on the carriage insert? Knowing how they use countersunk screws, I'd say probably not, but some gentle filing of holes would allow for some adjustment in the X-axis... Y-axis adjustment would likely require minute shims. It's possible, but it will be really fiddly.

[Rodne Gold]

Unless you have to change some other optics in the way of the lens when you change em and your general alignment is spot on , the only thing that should change is focal length unless one lens optic is cocked to another IE - \ (exagerated)
Where you might have an issue is if the lens slips into a holder and the position it slips into is higher or lower than the other lens , you would get a beam that is NOT hitting dead centre IF the alignment is off like this

\
---
.|

vs

\
.\
---
..|

[Scott Shepherd]

Dan, a tech didn't install our collimator, I did :) It does have adjustment screws, but this isn't a collimator problem. It's the lens assembly. It's a piece of aluminum, then they used countersunk holes for mounting it to the carriage, then mounting the mirror and the lens. Countersunk holes are horrible for aligning things. It forces the screw into that one place and allows for no adjustment. So in order for it to all work out, the countersinks would have to be within .001" of an inch on location (and concentricity-something countersinks are not well known for, especially if they used a single flute countersink), and then the holes in the brackets that hold the mirror and lens would all have to be within .001" of an inch as well. My guess is the tolerances on all of those items are +/-.005", or maybe +/-.003" if you're lucky.

I've aligned a lot of precision stuff in my life time and countersinking holes is not a precision alignment system.

You could correct it if you were so inclined.

In fact, when I got my new lens assembly, it wouldn't even fit into the carriage holder part. The lens and mirror were so far out of whack, it wouldn't slip into the slots they are supposed to go in and have the screws line up.

[Richard Rumancik]

. . . I was hoping their was something that I could calibrate to get both lenses to work together. I was told that the HPDFO is better for marking and getting small detail. I was hoping to use the 2.0 to cut the jigs and make the fixtures, then use the HPDFO to do the detailed rastering on the items. Since they don't line up, I can't do what I was planning...at least not as easily as I was thinking. . .

Well, I wouldn't give up yet. I suppose there are two approaches - you could fiddle with the mechanics, changing screws, shimming, etc. but quite honestly I think it will be too iterative and an exercise in frustration.

The other approach is to see if the offset between the two lenses is repeatible, and compensate in the drawing file. Does the collimator go in only one way? If not add some markings so it is always installed the same. Then what I would do is try to see if you can verify repeatibility between the 2" lens and your HPDFO setup. You could do this with some kind of test pattern eg plot crosshairs with the 2" lens, replace the optics, then plot crosshairs with the HPDFO.

Then measure the offset between the two setups. Suppose you find the HPDFO to be offset (x,y) = (+.065, +.020). What you would have to do in CorelDraw is compensate by shifting the graphic (only) -.065, -.020.

This may seem like a kluge but if you have no other choices it may allow you to do what you want.

In actual fact, I tweak my raster alignment fairly routinely when I dust off an old jig (oops - I mean fixture, not jig) for re-use. I can't really be bothered to try to set 0,0 accurately on my machine as it is done optically with the red-beam, which may or may not be coincident with the laser spot. So it is a crude reference at best. And with time (0,0) drifts; I may need to remove a rule and re-install, etc. So when I re-install an old fixture, I still butt it against the rules and screw it down, but I assume that the raster pattern may need tweaking. I use my trial part with masking tape on it and plot the graphic lightly. If it is not spot-on where I want it, I tweak the image in Corel. It is a whole lot easier than trying to move the fixture or trying to establish the true (0,0).

So if you wanted to establish the exact offset and tweak in Corel, that works, but for many graphics situations it is not that critical. I would just install a sample part in the fixture and use the actual graphic (or else a test pattern aligned to the graphics) and do a quick correction based on visual results. Takes only a few minutes for me to re-align an old job.

Of course, this strategy is based on having repeatibility. But I would think that even if the alignment between the 2 setups is not accurate, there is a good chance is it precise.

[Mike Mackenzie]

Joseph,

I doubt that the collimator is the issue, I believe that the difference you are seeing is due to the focus and spot size. Even though the HPDFO optics use a 2" focus distance they are not the same as the 2.0 Inch lens and where that focus is. You also have the difference between a 0.005 spot size and the 0.0015 spot size.

I would just score a square with the 2.0 inch lens and then put the hpdfo in and re focus then score the same line in the same position. Then look at it under magnification. If there is an off set with the vector lines then I would guess there is an issue with X or Y. If it does offset try scoring one line with one of the lens and then re home X&Y then change to the other lens and focus and run again.

If the vector lines don't match up then the problem exists with the motion system. (belt, bearings, Idler, motor).

Run the test in several locations on the table it would be best to use a painted aluminum or something like that so you can see any offset.

[Dan Hintz]

Mike,

Are you saying the HPDFO and standard 2" lenses have different focus heights (and different gauges)? Why would ULS do something like that? Even so, a difference in focal length wouldn't account for an X-/Y-axis offset.

Also, did you miss a "don't" here?

If the vector lines don't match up then the problem exists with the motion system. (belt, bearings, Idler, motor).
I hope this was a missed word, as the meaning is drastically changed (not to mention counter-intuitive).

[Mike Mackenzie]

Yes there are two different focus tools and they are quite a bit different this is mainly due to the grinding of the lens. You can calibrate both optics and then just switch between them in the UCP.

And yes I corrected my mis wording thanks.

[Scott Shepherd]

I'd have to disagree with you on that one Mike. How can an issue with the motion system cause lens misalignment from one lens to the next? It's 3 thumb screws and slide it in. It's an alignment issue with the lens and the mirror on that plate that they are mounted to. I scrapped a large job because of it. My swap out was a 2.0 lens for a 2.0 lens. It was off about .040-.050" left to right. Up and down was pretty close.

[Mike Mackenzie]

Scott,

With your scenario it is possible due to the lens mounts, Mirror mounts etc. It also could have been a focus issue between the two lens. I always check the focus tool calibration when changing lens. (replacing) You would be surprised how much different they can be.

[Scott Shepherd]

The fact that the "new" lens wouldn't even slide into the mount and allow the screws to line up with the holes told me there's not a lot of "alignment" being done on them out of the factory. My guess is they put the screws in, let them tighten up wherever they do and put it in a box and ship it. It caused me a lot of grief as all my files with fixtures are now .04" off where they used to be.

[Mike Mackenzie]

Scott,

Did you just change the lens or the entire plate? If you changed the entire plate you could try taking the lens and mount that to your existing plate to see if that corrected the offset. I still think it is important to double check your focus when changing out a lens.

[Scott Shepherd]

Yes Mike, I got the entire assembly- the plate, the lens, the mirror, all mounted. I just pull one out and put the other in.

I do agree, when you change lens, you should verify focus.

[Richard Rumancik]

. . .My swap out was a 2.0 lens for a 2.0 lens. It was off about .040-.050" left to right. Up and down was pretty close.

I don't know how easy it is for a manufacturer to "aim" a lens - for lack of a better term. In your case you have a deviation of about 1 degree of "aim" between the two lenses. So if the first lens was pointing 0.5 degrees left and the new one was 0.5 degrees right that would give you a .040" offset (1mm). It is not routine to change lenses in the middle of a job, so your situation was unusual. I expect it is just manufacturing tolerances between the lenses.

It might be more common to change from 2" to 1.5" FL lens in the middle of a job - but this thread suggests that if one tries to do that, you may not align. In other words, each lens has its unique "home" or (0,0) position. It would be nice if they could be manufactured so accurately that this did not happen, but laser engravers are not made to that level of accuracy.

In a machine tool like a mill, if you broke a milling cutter you could not just pop in a new one and resume. Each tool has it's own settings (size, length, etc) so these would have to be entered before running the new tool. This is kind of analogous as to what happens when you change a lens.

[Scott Shepherd]

Richard, I respectfully disagree again :) Now that I put in a new lens holder, supplied by the Manufacturer, it's screwed up every single file I've ever created. I do a ton of stuff that goes with stuff done over time. So I'm constantly adding something into a system that's full of engraved stuff. When I changed lens, it changed every file I have. Every file I create, 1000's and 1000's of them, is now off by .040". So every time I create a file or open an old file, I have to shift it .040" before I run it so I can know it'll line up with what I did last year. There should at least be an easy adjustment where I could shift my zero (there might be), and then the problem is gone.

I have a 4400 piece job sitting next to the laser right now. I made the fixtures 3 years ago. It runs 260 at a time, it's also about a 3-4 pt font. So I can't be off on them too much. So before I can even start the job, I have to figure out exactly how far I need to move my file before I scrap a batch of them.

[Richard Rumancik]

Steve, after re-reading your posts I'm starting to think that you got a misdrilled assembly. If the assembly would not mate with the carriage as you expected, what did you have to do to make it mate? Did you rework it? Maybe you should have rejected it outright as it did not fit. I'd be interested in the test Mike suggested: putting the old assembly back but using the new lens. Perhaps ULS needs to replace the badly-fitting metal bracket/plate. You might get better than .040 error, but its also possible that the error might not disappear altogether.

I do agree that countersunk screws are not good for aligning things. Each screw tries to force the plate around. To align two mechanical parts well, it would be better to use dowel pins and accurate holes, not threaded fasteners. But of course this adds cost.

[Joseph Tovar]

-Richard,

Your suggestion was exactly was I was going to do to remedy my issue. The offset seams to be consistant. It only appears that it's off on the x axis, while the y axis seams alright. I can make the fixture with the 2.0, then offset the graphic before I put in my HPDFO lens and raster.

I happen to be at the ISS show in Long Beach today and ULS has a booth. They had a tech from one of their Southern California distributors working the booth and I asked him the same question about swapping lenses. He said they should be lined up and it was a problem with my system. Then I asked him to run his demo unit with a 2.0 and HPDFO lens. He did and the results were like mine....off. His lenses were off on both the x and y. He changed his answer to "it could be off because he didn't calibrate his system before he ran the test. Then re recalibrated them both and ran again with the same result. The answer then became. "Each lens is made with the same process, but they(lens/mirrors) can slightly be off because the lens that are glued in can be misaligned, or the holes can be drilled a bit off, etc"

Anyways, I was happy to see that they had they same result as my system, but sad that they don't align as I wanted them to. Attached is a picture I took a few minutes ago of the test they ran today. They made a simple box and ran it on a piece of aluminum with each lens. You can see that the 2 rectangles are off a bit.