So what thickness/height measurement is being used where?

[raymond]

Hi All!

I’ve just taken delivery of my Glowforge and I’ve been extremely impressed! Having spent many hours messing around with other more conventional laser engravers, it’s so nice to see so many things automagically taken care of.

There is one bit of the software though that I’m trying to wrap my head around. I think I’ve got it but I was hoping someone could confirm that my understanding is correct.

The Glowforge needs to know the height of the surface of the material that it is engraving. It needs this for two reasons.

The first is to perform perspective correction from the camera so that when I place a job, it matches with the picture. If this is wrong, cuts and engraves end up in places that don’t match the interface.

The second is to focus the laser onto the surface of the material. Get this wrong and the laser power won’t be focused, resulting in a poor cut, wide engrave, larger than expected kerf and so-on.

Am I correct in my understanding that these distances do not come from the same source? Reading around the forums and docs, it seems to me that:

The first distance (for camera perspective correction) comes from the Proofgrade spec (as scanned by the QR code) or the thickness specified when you go “Use Uncertified Materials”. As far as I can tell, this distance is never independently checked by the Glowforge’s sensors. This means that engraving onto a warped bit of Proofgrade that isn’t sitting flat on the honeycomb will result in the engraving not lining up where the software shows it will.

The second distance (for focusing the laser) normally comes from the single point distance measure made by the distance sensor in the laser head when it says “Scanning material”. Crucially, this happens even for Proofgrade materials (for which the Glowforge already knows the “theoretical” height) so the focus will be (locally) correct despite any warping. The only time that the distance measure isn’t used is if a “Focus Height” is specified in the manual mode.

Does this sound right?

And on that note, is the focus distance measure always taken from the exact middle of the bounding box of the current job? This is pretty important for stock that is either warped or oddly shaped as we would need to make sure that either the middle of the job is the same height as the actual bits of the job to be worked on or that the height is overridden. For example, engraving onto a doughnut shaped object, we’d either need to put something in the hole to bring it up to the same height as the rest of the object for the distance sensor, or we’d need to specify the focus height manually.

Cheers!

• Ray

P.S. I hear mentions of ‘continuous autofocus’ … I’m assuming this is still a work in progress?

[kennethclapp]

@raymond

P.S. I hear mentions of ‘continuous autofocus’ … I’m assuming this is still a work in progress?

Yep.

I’ll be honest, I’m not sure what the thickness measurement taken by the laser is for other than to make sure the material isn’t so think the head will run into it. I’ve been confused by that one as well. Maybe someone can clear that up for both of us. From what I’ve been able to tell, both the visual interface, and the focal point of the laser are determined by the material with as either provided by Proofgrade QR or by us in the UI.

I’m sure I’ll be corrected in the areas I’m wrong

BTW, Congrats on your Glowforge!

[markwal]

@raymond

If this is wrong, cuts and engraves end up in places that don’t match the interface.

I believe your understanding of the two reasons to be mostly correct. I’m being a little pendantic here, but the cuts always match the interface (ie the rulers in the UI) (well unless you’ve lost registration by hitting something or your machine is broken in some way). The cuts don’t always match the bed image of whatever you have there because the dewarp algorithm has to compensate for the shape of the fish-eye camera as well as how far away the material is from the camera.

@raymond

never independently checked by the Glowforge’s sensor

Today this is true. I suspect they’ll eventually either measure it automatically or allow you to ask it to measure.

@raymond

always taken from the exact middle of the bounding box of the current job?

Maybe. Anecdotally, I don’t think so. I haven’t been able to determine the algorithm, but it doesn’t seem like it as simple as this. But yes, if you get unlucky and it takes a measurement from somewhere anomalous or missing from your material, it’ll get it wrong.

[markwal]

@markwal

I’m being a little pendantic here, but the cuts always match the interface (ie the rulers in the UI)

To expand on this a little more, if you put some painters tape on the bed and cut it with lines at as close to 0,0 as the UI will allow, then you can place things relative to that cut and get repeatable results.

There are still a few things that reduce precision with this approach:

• the kerf is non-zero on the painters tape so pushing something against the edge of the tape will still be off by about a half a kerf.

• the UI won’t let you place a cut at exactly 0,0

• the UI doesn’t let you set or see the precise location and the visual inspection of the guide vs. the rulers can only be as precise as the resolution of your browser not to mention the selection frame is offset from the object.

[dwardio]

Short version:

Height/thickness (whether Proofgrade or set manually for “Unknown” materials is only used for focusing the lid camera view so that you can place materials and files. The GF detects the actual height in one representative spot (pending continuous auto-focus) and uses that or the actual operations.

The only exception to this is when the user overrides the focus depth setting for an individual operation (cut, engrave, or score).

That’s it.

[raymond]

Terrific, thanks for the replies everyone! OK looks like I’ve got my head around this now.

I’m also discovering how even a fairly small amount of warping (I’m using a piece of the proofgrade maple ply that came in the pack) can make a pretty big difference to positioning! >.<

[kennethclapp]

@dwardio

The only exception to this is when the user overrides the focus depth setting for an individual operation (cut, engrave, or score).

quick question, if this is the case (and I’m not really doubting you here, just connecting the dots) then why is the focus in the UI always the same as what you put in (the number it shows before you adjust the focus height. Maybe I’m just great at measuring (not) or maybe it just doesn’t bother to update that in the UI.

[raymond]

I vaguely remember reading in an update that one reason for a manual override on the focus height is if you want the laser to focus somewhere other than the surface of the material … perhaps the material isn’t actually flat (and you know this) … or perhaps you want to use a defocused spot for a particular effect (I accidentally specified an incorrect focus just now doing an etch on anodised aluminium, the color is quite different, this could be done as a feature rather than a bug like I did!).

[dwardio]

@kennethclapp

why is the focus in the UI always the same as what you put in (the number it shows before you adjust the focus height.

Great question. If you’ve entered a non-PG material’s height at the top left of the UI, it will populate each underlying operation with that as the “default” focus depth. Despite that, it still doesn’t mean anything to the GFUI, unless you change that value in one or more operation. The red-dot focus will be used for all operations unless you have manually changed one or more. The ones you change will be adjusted by the difference between the material height you entered and the focus height. Nope. I was wrong.

Personally, I’d rather the UI use a “focus adjustment” in +/- against the red-dot value, but I rarely adjust, so don’t have a problem with this as is.

This is, as always, my hopefully-helpful but non-authoritative understanding.

[palmercr]

@dwardio

The ones you change will be adjusted by the difference between the material height you entered and the focus height.

Are you sure about that? It still says scanning but I got the impression it is ignored. Oh for a manual!

[scott.wiederhold]

@dwardio

The GF detects the actual height in one representative spot (pending continuous auto-focus) and uses that or the actual operations.

The only exception to this is when the user overrides the focus depth setting for an individual operation (cut, engrave, or score).

I have verified by experiment (adjusting settings, and examining the resulting PULS file that is sent to the device) that this is correct.

If the thickness of the material (what is set by the material you choose, or automatically for Proofgrade) is the same as the focus for the operation, the device will use whatever reading it gets when it goes through the “Scanning” operation.

If you change it to something different in the operation, it will use that value instead.

[brycelives]

To be pedantic, “pendantic” is misspelled.

[dwardio]

If I enter 0.010 as the Unknown material height, every operation is pre-populated with 0.010, but the laser will be focused on the red-dot value. If I change an engrave to 0.012, the laser will be focused 0.002 above the surface as determined by the red-dot value. My PRU worked that way, as does my production Basic. I might be wrong, but I’m not uncertain. Nope. I was wrong.

[markwal]

Hmm… This doesn’t seem correct according to other threads on the topic. It also seems pointless since it could then just ignore the red dot value (the red dot reading factors out of the equation as you’ve given them).

[markwal]

I take it back. You said if you change the focus height in the engrave to 0.12". I read it too fast and thought you said change the material height to 0.12". Apologies. Now that I read it right it matches my understanding.

[markwal]

I wish they’d just do “Auto” in the UI.

[dwardio]

@markwal

(the red dot reading factors out of the equation as you’ve given them)

My thinking for this specific use case was something along the lines of:

Actual focus depth used by the GF (aFD)= redDotFocusValue + (operationFocusHeight -
unknMaterialThickness)

In my example,
aFD = rDFV + (0.012-0.010)
or
aFD = rDFV + 0.002

Nope. I was wrong.

[dwardio]

@markwal

Apologies. Now that I read it right it matches my understanding.

No worries – It’s good to be challenged from time to time, it makes me take a 2nd, 3rd, or 103rd look and check it a different way.

[jbmanning5]

I get the thinking behind it…

Presumably, one is going to input as accurate of a material height as they can muster though reality is it may not be very accurate. The red dot fixes that accuracy - and then presumably one is going to change that value when it defocuses based off of what they perceive the thickness is, so it generates an offset value to try and get to where the user wants to go.

But yes, I’ve never understood why the red dot/depth sensing isn’t used earlier in the process to populate the material height.

[raymond]

OK in that case I’m slightly more confused … perhaps the answer to this question will clear it up.

It sounds like you can offset the red dot depth sensor (by specifying a height, which is interpreted as an offset to what the user thinks the height is). But is there anything you can do if you don’t trust the red dot depth sensor (eg. because you’re using an uneven height material so you don’t know where it’ll hit) and want to just use a completely manual focus height?