OK, so pretty familiar with optical alignment and set focus> I was doing a job without the crumb tray (and while I wrote the tutorial on crumb trayless measurement, decided to be lazy and use set focus. There are 2 things I note:
One the image of the board is not square to the machine (I squared it via a machinist square to the rails, which had better be square to the machine since they are the machine!!) (you will notice the hazy outline in the image which I put there specifically for lining up to the outside of the sign), and then “Set Focus” errors out (and yes I have refreshed, restarted my browser (actually my entire mac) and my GlowForge got rebooted recently as well.
And before any critics jump in, I didn’t name the damned rabbit, this is a job for someone. The rabbit is one of those massive Giant Welsh Rabbits so it’s a big hutch and a big sign.
I’m going to guess that you’ve done the camera calibration? Wondering if this is just the effects of aggressive camera/fisheye correction being applied to the wrong bed surface…
Yep. Normally I get dead on alignment. Hanve’t seen this, particularly the rotation issue (I mean I know it’s a cheap machinist square but still against my precision 2-4-6 block is true to tenths over it’s entire length. That is way, way worse than tenths… (well it might be tenths of an inch rather than .0001")
I’m sorry that you’ve run into a persistent issue with Set Focus. I extracted the most recent logs from your unit to take a closer look. Based on what I can see, it looks like this isn’t related to your browser.
Will you please take a photo of the underside of your unit’s print head and either share it here or email it to us at email@example.com?
I figured we’d eventually get to the lid camera so threw that in too. I cleaned the head with a Zeiss wipe. Not a lot came off.
Oh and then it gave this error:
Now when I measure this manually with a set of calipers it is 1.775" off the aluminum floor of the machine, making is .2775" thick as far as the head is concerned if the tray was in place…
I see that the image looks deeper at the left than it does at the right. This looks to me like a distortion of the camera view due to either the material not being parallel to the plane of the machine (crumbtray if it were present) or the blank being tapered, which I doubt is the case.
The image is not so much rotated as it is tapered.
Edit - just looked at it again. Is the hazy outline part of the metal?
"Hazy outline…Grassy Knoll. It’s a conspiracy!
No, it’s just a bounding box I put in the illustrator file (and it’s gray because I set the layer to ignore) that I put on to align the wood to the piece It is possible it’s not level (although those are machined blocks of aluminum it’s sitting on, so not the usual precarious stack of scraps).
Trust but verify:
OK so I just checked the focus operation while filming it to make sure it is actually trying to focus. And in fact it does move out onto the part and fire the red-laser, then retracts back to the home position. But then you get the error of part is too close/far as I noted above. Again with a micrometer the part is 1.775 inches above the aluminum plate.
Thank you for the detailed video and photo of the underside of your print head. I was able to locate the logs for the set focus attempts which returned an error and will share notes with our team so we can look into this further. I’ll follow up as soon as we have more information.
It seems a wild shot - but could it be the room light messing with it? Silver & white both reflect a lot… Can you throw a towel over the lid while it focuses?
So I swapped the aluminum blocks out for steel 1-2-3 blocks. Now the difference is clearly not the material since they’re under the wood. It’s more likely the slightly smaller height (high precision 1” rather than extruded 1” stock) so clearly being lower by a few thousandths made the difference.
Thank you for letting us know how your most recent test went. After extracting and reviewing your unit’s most recent logs, it looks like this might be an issue we’re unable to resolve remotely. I’m going to close this thread, and I’ll be in touch shortly via email with more information and next steps.