I actually had not read your post when I typed the “don’t whine” comment. So was in no way directed at anything you have said. Posted in hopes of heading off hundreds of people that expect to lay organic material on a flat surface and for it to stay that way. Personally have never had that experience from any manufacturer, Glowforge, Inventables, and certainly not the residential lumber providers. But have never used material in an especially controlled environment.
Will multipoint run the head all over the piece first then cut, or does it scan while it’s cutting?
I’ve been cool. I’ve been calm. I’ve not tried to get too excited. But now I am so stoked to get my GF! It’s really coming together! Thanks Dan and team!
We designed for a large margin of error. When we get multipoint autofocus, we’ll be able to run the head faster as we can reduce the MoE.
We also have a lot of secondary things that help: high quality optics reduce scatter and improve spot size, beam quality, the single-point measurement is deadly accurate (within 0.1mm), etc. Allows a few things to vary and still let you be successful.
Material thickness does two things. Primarily, it provides the dewarp specifications for the lid image allowing more accurate placement. The second thing it does is populate the focal height field for each operation.
However. The scanning process when you hit print is measuring the material height to set autofocus. It will use the measured autofocus height unless you have changed the focal height in the job operation manually, in which it will use that value.
That material height setting auto-populates the focal height setting for each job operation on the left with the same value you entered into the material height box of the uncertified material box.
At this point, the possibilities fork:
A. If you leave the focal height in the job alone (unmodified), the actual laser focus will be set by the scanning/red dot measurement. So if you entered a material height of .240” but the material was actually .300”, and you left the focal height in the job operation at the auto populated value of .240”, the scanning measurement will override this and focus the laser at .300” (or the closest available figure to that value)
B: you change the value in the focal height box under the job operation in the thumbnail. The red dot measurement will be overridden by your manually entered focal height.
I assume the logic behind this is that if you enter the material height and leave the focal height alone, it’s assumed that you want to focus on the material surface. And the scanning process (and subsequent autofocus) basically saves you from making a mistake in measuring your material. Or maybe you guessed. Whatever.
And the further assumption is that if you manually changed the focal height, you did so with a specific intent to focus above or below the surface of the material, so those wishes are honored.
I’m ok with the assumptions because I measure my material and if I focus somewhere else, it’s by intent.
What can be a little frustrating and could be confusing is that the focus only changes by steps. So if you had .240” material and put .230” into the box - the actual focus would be the same. I’m not so worried that the focus is the same but if you’re trying to focus below the surface, you have to find where that next step in focus change occurs.
It appears that the current possible focus heights are:
0.000, 0.028, 0.056, 0.083, 0.111, 0.139, 0.167, 0.195, 0.222, 0.250, 0.278, 0.306, 0.334, 0.361, 0.389, 0.417.
I believe it does round to the nearer value, but I can’t say with certainty. @palmercr did a bunch of testing on this, and can probably give a more definitive answer.
I think it rounds. to the nearest 0.0278 step but I haven’t tested the boundaries. I only tried every 0.0139 step and looked the the number of Z steps. Certainly 0.09 would use 0.835, i.e. 6 steps. It never does odd number of steps for some reason.
As it converts to metric it might round before or after that, who knows. I did notice a filename uses a metric conversion of 0.0139" to mm that is itself a rounded conversion from metric value derived from the metric gears. I.e. less accurate than its number of significant figures suggested due to a double conversion. So I don’t think there is any numerical rigour.