PDA

View Full Version : X,Y home position



Keith Colson
06-16-2014, 3:41 AM
I am trying to do some really accurate cutting by placing rectangular objects right up against the rulers and use the 0,0 as my reference. I have found that my laser is out by 3.7mm in x and 3.5mm in y using my measurement microscope.

Is there a simple way to calibrate the 0,0 position on the my VLS6.60 ? It must be something obvious I am missing.

Cheers
Keith

Dan Hintz
06-16-2014, 7:03 AM
Loosen the rulers and readjust them to match up with your (0,0) point.

Mike Null
06-16-2014, 7:55 AM
Isn't there a way within the software to do that--such as positioning the red dot pointer at 0-0 then setting home position to that point?

Scott Shepherd
06-16-2014, 8:53 AM
I don't recall there being anything in the control panel settings but I could be wrong. A quick look around and I didn't see it. Two things to do, one is what Dan said, and just to simply it, you take the rulers off, then put masking tape or painters tape down where the rulers were. Then draw a line in your graphics software from X0,Y0 to the end of the X travel, then do it again and draw it to the bottom of the y travel, giving you 2 lines, essentially right where the rulers are located. Then laser through the tape, peel part of it away, revealing where the rules go, then butt the rulers against the tape and tighten them in place. Then your rulers will right to your machine. I've done it that way myself.

Second choice is to contact Universal. There is a hidden set of settings that I can't recall how to get to. I've only been in the screen 2 or 3 times over the years and the last time was probably 4 years ago, so I don't recall how to get there. There may be a way to adjust them in those settings, but I honestly don't know. I know I've never done it. It might not be possible, but I don't recall what's available to change. I accessed it via a tech support session and they told me the keys to hold down and what to type in and then I rebooted and it was over. Shoot them an email or give them a call. Maybe there's a way.

Brian R Cain
06-16-2014, 9:05 AM
Dan is correct.

The easiest way to set the rulers is to first remove them, then lay down a strip of masking tape in each axis such that the centre line of the tape runs roughly along the zero line. Next take a cut through the tape along the zero position in each axis. Remove the sections of tape where the rulers need to be the butt the rulers up the remaining tape.

If you're using Corel, it doesn't allow you to print lines that are exactly along the 0,0 positions, you need to bring these inside the page slightly, say 0.05mm, but this will be as close as you will be able to set the rulers anyway, short of attaching a dial guage to the carriage and spending a lifetime tweaking the rulers to perfection.

Keith Outten
06-16-2014, 10:33 AM
You can tape a piece of sheet material to the bed about a quarter inch from both rulers.
Then engrave a 1" square on the sheet at one inch from both rulers.
Then measure the actual distance from the engraved square to the rulers.
The distance you are off is the adjustment you make on the rulers.
You may have to do this a couple times to get the adjustment perfect.
.

Bill Stearns
06-16-2014, 11:57 AM
Keith - and All ...
Not sure this info fits here, but maybe? I engrave quite 'few plates. (i.e. 3" x 1", 3-1/2" x 2", etc.) I used to place them in the upper/left corner against the rulers - was never accurate - off just enough to be noticeable - 'least to me. Plus, careful as I was, a plate, or two, would sometimes slip underneath the ruler a smidge! (Believe the Epilog techs once told me I might try adjusting what they called the "tickle". ? ) Anyway, what I've done now is: cut out plate shaped templates of the most popular sizes on a sheet of lasermax. (designed taking into consideration holes 'n clipped corners.) I use this template sheet with confidence, now, on the single plate jobs. I've cut additional templates allowing for multiple plates of the same sizes. Now, I don't cringe when someone asks for a plate to be engraved! (Hope this helps?)

Bill

Kev Williams
06-16-2014, 9:41 PM
I'll add to the confusion... ;)

What I do when I need a high-precision home position is just make a new home position...

tape some scrap strips of laminate to the table against the stops, inch wide or so, 12" long if that works-

Now run a cut 1/2" from the stops. (Me, I set the cuts at .503, which usually accounts for my laser's kerf nearly perfectly)

Remove the scrap, and test your new stops for accuracy. Put in a 1/2" top and left margin in Corel to serve as your new zero points. Make a couple of test vector cuts exactly 1" from the new "zero", then measure. If you measure say, 1.003, then move the appropriate margin out .003". If you measure .997, move the margin in .003", etc...

I build ski lift operator panels, which are machined from the reverse side because I have to mill bosses for blind studs. Since the panel is built in reverse on a "top-left" machine (my IS7000), this means the ENGRAVING 0/0 position is the top RIGHT side of the panel. Since the material is rarely identical in width from piece to piece, I use the above method to create a top-right zero point in my lasers. Once I determine the the new 0/0, the engraving always lines up perfectly with the machining.

Henri Sallinen
06-17-2014, 2:31 AM
Keith - and All ...
Not sure this info fits here, but maybe? I engrave quite 'few plates. (i.e. 3" x 1", 3-1/2" x 2", etc.) I used to place them in the upper/left corner against the rulers - was never accurate - off just enough to be noticeable - 'least to me. Plus, careful as I was, a plate, or two, would sometimes slip underneath the ruler a smidge! (Believe the Epilog techs once told me I might try adjusting what they called the "tickle". ? ) Anyway, what I've done now is: cut out plate shaped templates of the most popular sizes on a sheet of lasermax. (designed taking into consideration holes 'n clipped corners.) I use this template sheet with confidence, now, on the single plate jobs. I've cut additional templates allowing for multiple plates of the same sizes. Now, I don't cringe when someone asks for a plate to be engraved! (Hope this helps?)

Bill

This is my way of doing it. Although it is good to have your home position where it should be at 0,0, but when I need to do something precisely in the middle etc. I cut a jig from corrugated cardboard that I have laying around quite much. This way I can be sure that my file is set dead on to my materials and everything cuts / engraves out just nicely.

It would not be a bad idea to calibrate your home position even if you would use the jig way.

Keith Colson
06-17-2014, 4:37 AM
Thanks for all the feedback.

I have looked into the calibration methods and see that the Universal laser recommends cutting the tape on the engraving table and butting the rulers up against it. I can see two errors in this method, correct me if I am wrong. The laser cut width is about 400 microns in width. So doing this method precisely put your rulers in an extra 200 microns plus my unit will not cut on the 0 line, I need to step it to 0.03mm to cut. That's a 230 micron error stack before my eyes/microscope even come in to play. Its seems a bit silly, with my 3D printer I can just type in the offset and be done with it.

I am familiar with the nesting method where you cut a nest to place your part, this is currently what I do but it's extra work and another step I can make a mistake on, not to mention more cost for my customers. My Universal agent is finding out if there is a home tweak but he assures me that there is not but he is checking anyway.

If I get no luck with Universal I will try an idea I have. Having a precision laser cutter with ruler stops means I must get the rulers set to with in 10 microns or so for a fusspot like me to be satisfied. If I stick something in the laser I want it to cut exactly where I say within my careful placement error. My $500 home built 3D printer gets 10 to 20 microns accuracy so I expect my much more expensive laser to do the same.

The other thing that worries me, if I go to all the trouble of getting the rulers right on the money and then I adjust/maintain etc. I will have to go through the process again. So I need a quick repeatable process.

Cheers
Keith

Scott Shepherd
06-17-2014, 8:03 AM
Having a precision laser cutter with ruler stops means I must get the rulers set to with in 10 microns or so for a fusspot like me to be satisfied. If I stick something in the laser I want it to cut exactly where I say within my careful placement error. My $500 home built 3D printer gets 10 to 20 microns accuracy so I expect my much more expensive laser to do the same.



Your expectations are way off here. 10 microns is .0003" of an inch. You can't hold .0003" of an inch with machines that lasers. The beam deflection and angle of the beam on various materials and thickness would make that tolerance impossible. I do think you are way off on your expectations. Moving the table up and down will completely change that too. The machines aren't aligned with the kind of accuracy you're talking about. You could get it dead right at Z0 and it could be off 100 microns at Z 1".

It's not ever going to be that accurate and it's going to move around over time as well. Not a lot, but it will move.

Keith Colson
06-17-2014, 9:38 AM
I should have clarified my goal. The item I am working with is paper thin so Z does equal zero. I also only want the accuracy in the corner around the rulers. I want to align to about 10 microns or as good as I can get with the hope that I can stay well under 100 microns of error for some period of time, but starting with 230 microns by following the manual does not work for me. Yes, once the z is moved there are many other errors that stack up, table movement, beam angle etc. It is an interesting challenge and there is plenty for me to learn on this journey.

Richard Rumancik
06-17-2014, 9:48 AM
I was very surprised when members suggested that the only way to move the origin on thsi machine was to move the rulers. But I pulled up a Universal manual and it seems like that is the prescribed way. It seems to be a very crude method - it is odd that they have not yet patched their firmware to allow you to enter actual measured offsets.

Given that moving rulers seems to be the only way to adjust home, I would go with Kev's method. I think this could be automated somewhat. Keep in mind that if you are cutting rectangular sheets you do not need to cut full (offset) rulers. You just need enough points to lock the material in uniquely. A tooling engineer would say this only needs three contact points (two on the long edge and one on the short edge.) You might elect to have a couple more if it works better for you. But lets say you just attach some small "coupons" to the table along the rulers in select locations. They should be screwed down or clamped. Then cut off each coupon to generate the required contact points. No real need to have full "rulers" if you are fixturing for rectuangular sheets.

You could probably do a setup like this in 5 or 10 minutes with no material cost.

When I think of nesting, it is for multiple parts. Each part is dropped into a pocket which has some clearance. This is fine if you don't care about exact position or rotation (eg name badges). But if you want better accuracy you want to fixture to edges or points (i.e. push the part into the corner, rather than let it float in a pocket.)

When the job is done you just remove and discard the coupons and install new ones for the nest run. You can get surprisingly good accuracy and repeatibility using this method.

Scott Shepherd
06-17-2014, 9:55 AM
I should have clarified my goal. The item I am working with is paper thin so Z does equal zero. I also only want the accuracy in the corner around the rulers. I want to align to about 10 microns or as good as I can get with the hope that I can stay well under 100 microns of error for some period of time, but starting with 230 microns by following the manual does not work for me. Yes, once the z is moved there are many other errors that stack up, table movement, beam angle etc. It is an interesting challenge and there is plenty for me to learn on this journey.

That's not even remotely realistic on these type machines. I've seen machines move .010" from day to day. They just aren't that accurate. They position accurately, but they aren't dead on repeatable to the level you are talking about. Not by a long shot.

10 microns is splitting a human hair 10 times, and it's only one of those remaining pieces.

Kev Williams
06-17-2014, 10:48 AM
This reminds me of a nightmare job I had, laser marking an anodized go/no-go jig used by a medical parts mfr. to measure needle tubes of various sizes. The jig measured roughly 14" x 7". There were nearly 100 vertical hash marks on this jig, .188" tall x .060" wide, and centered to within +/- .0015" of the dim spec. From dim input to the jig, about 10 marks were dead on, but less than half of all marks were within tolerance, the remaining marks were out slightly to over .030" in a couple of places. In the end I had to test run and measure each mark individually. I used painters tape for the mark, and my measuring equipment consisted of a 7x eyepiece and 12" Harbor Freight digital calipers. Once I was satisfied I etched the mark and moved to the next one. I definitely underbid THAT job! ;)

Shepherd is correct, there's no high-precision to be expected from these machines. Belts stretch, belt teeth wear, focus distance changes, laser beam alignment is never perfect...

Brian R Cain
06-17-2014, 12:33 PM
Shepherd is correct, there's no high-precision to be expected from these machines. Belts stretch, belt teeth wear, focus distance changes, laser beam alignment is never perfect...

Absolutely. Add to that the machine has a conflict of interests between achieving high-speed engraving and accurate cutting. For any machine to achieve the degree of repeatable accuracy that is being asked for, it requires being built from robust parts with lots of mass for stability and to dampen vibrations. Compare the rigidity a high-precision milling machine or grinder is built to with that of a laser engraver, which is primarily what these machines are designed to be. The fact they can be used for cutting as well is a bonus, but you can't expect high-precision from lightweight parts.

High speed engraving requires the motion system to have as little mass as practically possible so it can be accelerated and decelerated quickly without overshoot. They use belt drives for this rather than high-precision ball screws because of speed and cost. If you were to take the X-arm off the machine and examine the wall thickness of the extrusion, you might be surprised by how little of it there is.

I had an instance where a machine was reported to be cutting sometimes with rough saw-tooth shaped edges on the parts and at other times smooth edges. On inspection I found the machine had been wedged into an alcove alongside a compressor, such that the compressor was touching the machine cabinet. Every time the compressor started up it was vibrating the cabinet.

In another instance, someone bought a machine for cutting intricate stencils and imagined that because he had lots of power and was only cutting thin film he could run it in vector mode at 100% speed and get good results. Quite simply, it's the wrong machine to do that.

As for re-location of the origin, this was possible on earlier ULS machines but ditched when they made the move to have them controlled in real time by the computer. The control software gives the ability to re-locate the job position after it's been printed and to be moved to a point located by the red dot pointer, which is a real plus when you need to align a cutting file with a print. I can see no reason why they couldn't add a re-locatable origin to the software if they find people require it, but I expect they imagined that job re-location would be more useful. If they pick up this thread, they might add it to a future update.

Dan Hintz
06-17-2014, 4:57 PM
My $500 home built 3D printer gets 10 to 20 microns accuracy so I expect my much more expensive laser to do the same.

Clarifying expectations...

The typical low-price 3D printer only nets you 10 micron resolution in the Z-axis, mainly because the platform doesn't move during processing (i.e., it only moves in one direction after a layer is done, and gravity helps to get rid of gear slop), and movement is over a very small area (e.g., 6" square). Now, compare that to a laser's 75-150ips moving carriage that floats above the table on rails 24-48" long... something has to give.

Brian R Cain
06-17-2014, 5:42 PM
Clarifying expectations...

The typical low-price 3D printer only nets you 10 micron resolution in the Z-axis, mainly because the platform doesn't move during processing (i.e., it only moves in one direction after a layer is done, and gravity helps to get rid of gear slop), and movement is over a very small area (e.g., 6" square). Now, compare that to a laser's 75-150ips moving carriage that floats above the table on rails 24-48" long... something has to give.

Very good point. Further it isn't the same question being asked, which as I understand it, is about user placement of material rather than machine resolution. I have yet to see a 3D printer where the issue of rulers exists.

It would be quite difficult and time consuming to achieve the sort of accuracies being asked even when using a robust machine capable of it. The material isn't clamped, a speck of dust or a burr can throw it out and it depends what is being cut. I've had occasions when cutting small pieces of 3mm acrylic close to the rulers move as a consequence of the waste part bending away once the stress is relieved. You can never tell with light materials like paper and card whether the extraction will move them.

My best suggestion is you build a fixture to hold the material in, something that can be secured in place on the bed, then run a test every time you initialise the machine. I doubt that the initialisation mechanism is anywhere near that repeatable, although once established, carriage positioning relative to the home position (not the ruler zero) isn't too bad. Measure the error you get and compensate for it in the Re-locate View of the Control Panel.

Scott Shepherd
06-17-2014, 6:24 PM
I've had occasions when cutting small pieces of 3mm acrylic close to the rulers move as a consequence of the waste part bending away once the stress is relieved.


Oh, if I only had $1 for every time that's happened. To take it one step further, when you're cutting some parts out of that material, by the time it gets to the end of the cut, the part has moved inside the kerf, being pulled in one direction from the cutting. You can end up with mismatched cuts at the start and finish, based solely on that happening, nothing to do with any slop in any of the drivetrain.

Brian R Cain
06-17-2014, 7:09 PM
By the way, Scott, you've reminded me that you and I had a misunderstanding a week or so ago that I intended to clarify, but when I read your reply It was way past my bedtime here in the UK and the next time I looked in, the thread wasn't hot to remind me. I'll see if I can find it to better explain that my comments weren't in reply to yours as you'd imagined.

Scott Shepherd
06-17-2014, 7:33 PM
By the way, Scott, you've reminded me that you and I had a misunderstanding a week or so ago that I intended to clarify, but when I read your reply It was way past my bedtime here in the UK and the next time I looked in, the thread wasn't hot to remind me. I'll see if I can find it to better explain that my comments weren't in reply to yours as you'd imagined.

No problem Brian. Don't sweat it.

Keith Colson
06-18-2014, 3:01 AM
Thanks for all the feedback, lots of great ideas and stories!

I decided to go and measure the repeatability on my VLS 6.60 today. I carefully placed A4 paper in the laser against my rulers and cut a cross-hair at 5mm,5mm on my uncalibrated rulers. I repeated this 8 times. I then measured the edge of the paper to the cross-hair using my measurement microscope in both x and y. My measurement resolution is 37.5 microns, half a graticule mark. It was all done at 0 Z.

Here are the results. You can see that I placed sample 7 badly but the rest of the numbers look like reasonable results. X was typically +50 to -70 microns. Y was +125 to -80 on 6 out of 8 samples. I had one large error on y, not sure why, maybe placement again. I will do some more testing. Here is the spread sheet of the results.

291459

I had another idea for better ruler placement which I will try that should take up the error I noted earlier. Cut some blue painters tape at 1/2 a laser width in from 0. Paint the etched line with a marker pen. Butt the rulers up against the edge of the black line and then test to see if it worked as expected. Maybe I can work out a trim value for this technique. I do like the idea of adding some anchors to the table too.

I have lowered my expectations but I am still keen to see how good I can get Z 0 to work around home. Thanks again for all the feedback.

Cheers
Keith

Scott Shepherd
06-18-2014, 7:44 AM
Keith that's about .001" of an inch and that's repeatability. The issue is that once the machine is powered on, it's pretty repeatable. However, when you turn it off and turn it back on again, it can move. Your test needs to be a lot more than just one sample while the machine was turned on, on one day. If you want to see the real world expectations, you need to turn it off and on, move the table up and down, and do it for various days across a longer time frame. Then you'll see what we're talking about. You're creating lab type experiments and those of us who have been using them for a while are saying that's not a real world result so don't expect it.

Your machine is also brand new so it's about as good as it's going to get (unless you bought it used, in which case, it can be better than it is right now, quite possibly).

You mentioned putting paper down as a test. How square was the paper? Was the paper square within 10 microns? If not, then how can you say the laser didn't repeat? It could be the paper. I've never seen paper cut that square.

Robert Walters
06-18-2014, 10:32 AM
Don't forget about thermal expansion between...
Night and Day,
Summer and Winter.

Steel body/fasteners,
Aluminum table/rulers/guides,
Nylon belts,
Glass mirrors,
Plastic rollers,
they all expand and contract at different rates.

Dan Hintz
06-18-2014, 12:17 PM
Don't forget about thermal expansion between...

You got to it before I could...

Keith Colson
06-19-2014, 4:30 AM
I had another fun afternoon playing with the laser today. I calibrated my rulers using the masking tape and cut, painted the line and put the ruler on the edge of the line (using my dentist loupes). I got some quite good results. My Y axis I managed to get within 50 microns. The X I only got to 100 microns. With careful paper placement I could repeat these numbers +-50 microns. Here is a picture of the paper x-line which is cut 90% through. The center of the laser line is 1mm in from the calibrated ruler.

291539
291534

It is quite tempting to put some micrometer stages on the rulers but the honeycomb gives small clearance to place it.

I put the table up and down and power-cycled the machine a few times and did more cuts. This yielded a 250 micron shift in the x axis (thanks Scott! I might have missed that) but y stayed pretty good. Now I have this easy measurement method I can raise the table (at the given ambient temperature) do a couple of calibration cuts, record and use the offsets.Then I can do my work accurately for a few hours or so.

Now I have my origin sorted its time to make a printed circuit board that has been on the back burner. Does anyone have any settings for cutting FR4 glass epoxy sheet 1.6mm thick on a 60 watt Universal?

Cheers
Keith

Brian R Cain
06-19-2014, 5:56 AM
"Does anyone have any settings for cutting FR4 glass epoxy sheet 1.6mm thick on a 60 watt Universal?"

I supplied a 30 Watt VLS2.30 for this very purpose. Don't imagine it's straight-forward and can be done in a single pass. You'll end up with a charred mess if you try that. The trick is to make loads of low power repeat passes but they can be quite fast ones, say 10% power 20% speed, 150 PPI. You'll get carbon on the edges but it's liveable with and most can be wiped off. The company doing this makes chips about 6mm square for use in navigation systems and for practical purposes they build 100s at a time on a PCB then separate them with the laser, so you'll get an idea of the quality they need and what's achievable. Although these chips are actually built on thinner FR4, the testing before they purchased was done on 1.6mm.

To reduce the amount of heat build up, turn off vector optimisation so you can decide for yourself the order that the cuts take place, which will be the order the vectors were created. In other words, work out a path that's efficient, then copy and paste the drawing on top of itself a number of times. The file will then run through the vectors as you drew them, also in the direction the lines were drawn, and repeat the sequence over and over again. If you leave vector optimisation turned on, the driver will sort the vectors into the fastest path to get through the job, thereby repeatedly cutting the same line until all the vectors for it have been completed before moving on to the next one and doing the same with that.

I wouldn't suggest this as a production method for PCBs unless they are very high value, which in this instance they were. If you're interested, prior to getting the laser, the chips were being separated with a diamond wheel but due to the small size, it was a difficult to manage and had a high failure rate. The laser method gives a high success rate. If I remember correctly, the circuits were built then covered with a cap and the cutting took place on the underside but not all the way through. This enabled easier management of the boards and the individual chips could be snapped of after the board were cleaned up.

Keith Colson
06-19-2014, 6:12 AM
Thanks for the info Brian. I only plan to do this for my own circuit boards, I can order nice ones cheaply for mass production. I wonder if I should use phenolic instead, its just not as common so harder to buy, plus I have FR4 in stock. I was hoping to drill the holes in the board with the laser (and chemical etch the copper side) but it sounds like that might be too hard. I guess I should just try it for fun. I will see if I can get my hands on some phenolic to try too as I am guessing it will laser more easily.

Thanks for the hints on the vector optimization and multiple cutting method. I will have to keep that in mind for all kinds of projects.

Cheers
Keith

Dan Hintz
06-19-2014, 6:19 AM
I was hoping to drill the holes in the board with the laser (and chemical etch the copper side) but it sounds like that might be too hard. I guess I should just try it for fun.

I would not waste your precious time on it. Cutting the edge is where your work should stop, and even then it will be a slow-going process, as Brian says. Expect post-processing if you go down the road of cutting out the boards themselves. Unless you're working with odd-shaped boards, the easiest (and fastest by FAR) method would be to set up a table with a Dremel grinding disc poking through the top. Set up a straightedge to press the PCBs against and run them through. Nasty dust, so work in a well-ventilated area, but you can cut close to 1"/sec by that method... the laser isn't going to touch that.

Robert Walters
06-19-2014, 8:20 PM
You got to it before I could...

Heh, you can have it Dan, I learned that the HARD WAY!

Did you know...
That a 0.500" rod does NOT fit in a 0.500" hole... DOH!

Dan Hintz
06-19-2014, 9:11 PM
Did you know...
That a 0.500" rod does NOT fit in a 0.500" hole... DOH!

But with a big enough hammer, you can sure give it the ol' college try!

Robert Walters
06-19-2014, 9:32 PM
But with a big enough hammer, you can sure give it the ol' college try!

5lb sledge, dead blow... didn't even budge.
Tossing rod in freezer and taking torch to hole does wonders!

Making the laws of thermodynamics your bitc... work for you! :D