When i first received my Glowforge I made some Gifts of good Measure for friends and that was about it.
More recently I have had more projects suited to the machine and am using it quite a lot. I built a box with tabbed corners for a fan I wanted to use for the exhaust, and was amazed all over again with the dimensional accuracy of the machine.
One aspect of that project was just a bit less than perfect; as I was assembling the box i noticed it would not lie flat on the table. That made me question whether it was building with a twist in it. So out to the granite surface plate in my shop and sure enough the faces of the assembled box were twisted. As a cabinet maker I knew that if dimensionally accurate panels are assembled into a twisted box - assuming the stock is flat and not warped or twisted - then the panels are almost certainly out of square ie. parallelograms. Note: My proofgrade Draftboard was dead flat - not warped or bowed
Sure enough, all the pieces were out of square. About 1mm in 200mm. Checked against a Mitutoyo machinst square.
One design choice made by the Glowforge team was to use 2 stepper motors to drive the Y axis (move the gantry). Great choice. The gantry has quite a bit of mass and could use a bit more motor to accelerate it and travel. Unfortunately the two timing belts driven by the steppers are not mechanically connected to each other. The gantry itself is somewhat self aligning on the rails and wheel it travels on. On my unit the gantry can be deflected off square about 3mm along its length without very much force. If you move the gantry by hand (power off of course) pushing near one end and not in the center it also easily goes off square the same amount.
Now as users and readers of manuals and forum posts, we have been told to gently move the gantry to the rearof the machine (power off), and square it initially by contacting the the back of the case. I have done this frequently - it makes sense. The problem is that doing this on my machine does not result in a square gantry. It’s close, but not square. Once the steppers are powered this alignment is held by the snychronised steps of the two motors. I do not believe that my gantry is wandering off this initial alignment during cutting.
If you are still reading you are probably very patient! LOL.
I have made a crude but effective test arrangement to see whether the alignment of the gantry can be maintained by the addition of some pulleys and non elastic cord. Picture attached. I have not tested this extensively yet, but early on it seems to provide a significant improvement to maintaing square. I estimate that the misalignment potential has been reduced to about 25% of what it was before. I will update when I know more.
In order to prevent readers from misunderstanding I will summarise:
The procedure to square the gantry as posted in mauals and on this forum is very sound. It certainly creates a repeatable initial alignment.
The above procedure may not be initializing the gantry perfectly square on all machines. Mine was close but not dead on.
I love my Glowforge, and do not want anyone to think it is substandard in design or build.
If you are finding that sometimes parts you cut do not fit well, check that your machine is cutting square. Perhaps adding some shims at the back of the machine could correct initial square if your unit is a bit off.
My arrangement is to satisfy the designer in me that always wants to improve things.
Hello every one,
I knew that this would be a hot topic. So I will answer to the above:
I agree that it is possible to square the gantry with a square. I just want to not have to do that. Some people are moving the gantry to a position near the center of the bed to promote the initialisation working better. When you do that you will probably “unsquare” the gantry.
Jules: Yes, mine was cutting “pretty” square too. It is the machinist in me that wants square to .005" per foot LOL. It is relative. For the more decorative stuff you could be way out and not notice it - I never did until I checked specifically. For inlays where you might be flipping a part over to match the bevels of the cuts it will magnify “off square” dramatically.
deirdrebeth; good comment, belts are tight and play the same note when you pluck them Believe it or not the frequency of a timing belt when plucked is a modern and very accurate way to check belt tension.
The focus of my post was “don’t assume your machine is squared”. BTW, the article that eflyguy kindly referenced is a bit inaccurate, as an out of square gantry will cut the lengths of the sides the correct length. “Rhombus” A rhombus has four equal length sides, but the corners are not square.
I think that you are right about .001" resolution stated by GF. I am pretty realistic about actual vs theoretical accuracy of a machine tool, for that is what the Glowforge is. I hope that you get my point about the lack of mechanical gantry alignment.
I am going to do the following test, just for giggles: Cut a 4" square out of the center of a 6" square piece of stock. Label the corners of the two pieces and then check whether the part fits into to hole in all 8 possible positions. If it does, then the geometry of the cut is accurate to the within the width of the kerf.
Yes, what I like about that method is that only the parts are needed - no measuring tools which may or may not be calibrated. Of course, the larger the square, the more demanding the test becomes. I realize that I am being picky, but a machine that can resolve movements to .001" also needs to back that up with accurate geometry. I make things like gears with my Glowforge, so geometry matters. If the gantry is not squared then circles are not round for example…
OK, I have my machine now cutting square enough to cut a 4" square and flip it to all 8 positions in the cutout. Happy with that. Now I will see how well that setting on my pulley arrangement holds over time & machine power cycles & gantry hand positionings.
I went through the squaring steps outlined in the link further up the thread and discovered that if I squared the gantry to the Right side rail it was not sqare to the Left side rail. Hmm - gantry not straight? No, gantry is straight compared to a straight edge. Rails not parallel? Yep. .018" wider at rear of unit than at the front. Seriously doubt that aligning the rails is feasible. So I split the difference with the square. I wonder what the effect of the gantry running back and forth on rails that are not parallel is? is the carriage restrained in the X direction at one end or both ends?
Once again I want to say “I love my Glowforge!” For the money it is excellent. My findings are interesting to me only as a machine designer. If I can cut with X & Y axis perpendicular to within .004" per 4" that is great!
The support trick never worked for me. Mine always goes back to not being square. I just live with it. Luckily I have other lasers that maintain a square though
Most people are finding their machines stay square, or they do not need very accurate perpendicularity between X & Y axis. I usually do not require it either. I want to see whether my pulley arrangement helps. Looks good so far…
