Precision Cutting

Pushing the Limits on Plexiglass Cutting

Once again I used Inkscape to design a symmetric stencil, this time using spirals and mirror reflecting the copy of the spiral while superimposing it. I then used a line tool to connect the very short (< 0.03 in.) disconnects as the spirals are very hard to exactly superimpose. This is a big test for the laser, as it will have to precisely cross over cuts already made to get the interior objects of the spiral to fall out. How cleanly can it leave very thin and sharp corners and will the shapes fall cleanly out?

I used painter's tape on the back of the plexiglass to reduce cleaning time for debris, cracks and char marks.

The photos show that the crescent shapes are intact even at the thin tips. All interior spiral shapes fell through cleanly except for one which pivoted upwards. The laser had to be stopped or else it could collide with the protruding plastic. The laser does have some issues picking up smoothly after a pause command to the gantry, leaving some rough edges. Next time I will not use any center support bars and this should fix the problem.

The laser power is so strong after a full cleaning of mirrors and lenses that next time I will be looking to drop the power down from 30%. The full run time for the job is less than 6 minutes.

The original Inkscape design

The painter's tape protects the edges of stencil cutouts 

and the outer circle.

All of the cutouts, in good condition. 

Precision Cutting

Today I started a series of tests to see if a quantitative relationship could be developed to understand how a nominal dimension cut size in software translates into actual dimensions in flat wood. In other words, how much material is lost due to burn and charred dust that makes the cut dimension LONGER than the size indicated in the software design. I found that for a 1 inch block, the gap inbetween the cut-out block and the frame cut around it is about 1/32 inch. Knowing that the laser cut performance varies in the vertical and horizontal directions, I designed a series of 5 blocks for each dimension. The dimension being varied changes by 0.01 inch incrementally (1.39 , 1.38, 1.37, 1.36, 1.35 inches) for a nominal 1.4 inch horizontal. The verticals are varied also by 0.01 inches starting at 1.14 inch to 1.10 inch for a nominal 1.15 inch design dimension.

First, I tested various combinations of power and speed to see whether (a) the cut makes it through wood 3/16 inch thick, and (2) how much char burn dust is created along the cut edges.

This is a Boss Laser LS-2436 150 Watts and 4 inch focal length.

#1 35% power and 25mm/sec speed cut

#2 30% and 25 mm/sec no cut

#3 25% and 25 mm/sec no cut

#4 20% and 15 mm/sec cut

#5 15% and 15 mm/sec cut

#6 12% and 15 mm/sec no cut

#7 12% and 10 mm/sec no cut

#8 12% and 5  mm/sec cut

I assessed which combination of speed and power processes cuts with the least char debris by visually inspecting the sides and wiping all four edges of each cut-out block against the cotton terry cloth pictured below.

To my eyes the 15 mm/sec and 15% power produced the best results.

So, when I set my parameters to 15/15 and ran my series of 10 test blocks, I was surprised to see that none of them had full cut-throughs! As you can see from the photos though, burn lines are evident to different extents on the backside of the wood, suggesting that the power only needs to be increased a bit to achieve full cut-throughs. 

Front and back sides of a series of unsuccessful cut-throughs on 3/16 inch wood 

Click to see the video of the cutting action

What could possibly explain this? Two things that I can think of:

1. There is warping/unevenness in the wood and the metal bricks holding the wood in place are successful to different extents in keeping the wood perfectly flat.  Differences in flatness will affect the ability of the auto focus to work properly and thus affect the cutting strength. I will try to use more bricks to keep the wood flatter.

2. I made a mistake and did not match the laser frequency in the g-code files between the single (125 kHz) and multi-block designs (20 kHz). So I went back and resaved the multiblock deign with a 125 kHz setting and will cut the test series again.

Precise Wood Cutting

I ran a set of 10 block cuts at 20% power and 15 mm per second speed on the 4 inch laser, with block dimensions set to the 1/1000 inch.  The two small stocks did not cut all the way through on one corner.  I attribute this to either a local warping in the board or an anomaly in the wood grain or density of the wood.  With the exception of the block of dimension at 1.135 inches by 1.40 inches, 1.133 inches by 1.40 inches,, and 1.373 by 1.15 inches all of the openings accommodated the nominal 1.40 inch by 1.15 inch Plexiglas block with little to no resistance.  The test stock was able to push all the way through.  The variations in length across a series of blocks cut into the wood were in intervals of 2/1000 inch.

Use painters tape wrapped around your finger to lift cutout blocks from the board without lifting or moving the board, or having the block fall through to the debris catch pan. 

The conclusion is that the 4 inch a laser cannot be trusted to precisely differentiate cut lines at precisions at 2/1000  inch.  However based on the prior tests earlier in January, the 4 inch laser certainly can be trusted to cut lines as good as one-100th of an inch. These findings appear to be true in both the X and while Y directions of the laser chamber.  The cut-off mark between true precision and lack of precision for the 4 inch laser under these operating conditions is somewhere between 9/1000th and 3/1000th of an inch. See the photos below and also this wood cut album for more examples with captions.

incomplete cuts due to wood anomalies

opening is too small