OK, i have faith since I have done it before, but please think about resynching the preview bed camera image and the job file after snap mark registration. It’s nerve-racking to press the button on a hard to obtain material item, when it looks like it’s 0.5" off (I know the magic will work… but not great UX)
i almost always score the snapmarks on the lightest possible setting just because, even though i’m sure it will work, i still can’t trust it.
Leaps of faith are a challenge for me as well. I trust, but verify with the light score as suggested.
I’m on the fix the UI side of the fence. Since the GF already takes a picture of the snapmarks to align to, why not update to that and fix image alignment.
Not that simple. The current capability is to deliver an preview image that approximates the bed image through a distorted fish eye to rectangular dewarping process. That’s good enough to send the head near the Snapmark positions. The head camera, which has a more narrow and planar image capability, then takes a detailed picture to accurately determine the actual Snapmark position. So at this point the unit knows precisely the number of steps to move the stepper motors for accurate placement of your design.
But to update the UI there would still need some fairly complex and machine specific calculations. The UI preview is not a linear image. It’s off a lot toward the edges and fairly accurate directly below the camera. You can’t just move the preview image a specific amount. It’s the same problem as dewarping the original Preview image.
And before someone says it… Yeah I know a fixed move might be a much closer match but it will still be wrong. Might come close to matching one Snapmark but not the other.
Well, supposedly the units were mapped at the factory (i.e. they know the warp factor?) which means the math is pretty simple actually. You know the precise location of part of the image and then move on the image. Now I am assuming they truly do have a warp map for each unit (one of the reasons I did give up my PRU was to get that dewarp at the factory; admittedly we haven’t seen a whole lot done with that)
Not saying they shouldn’t have done it by now. I’m only saying that the Snapmark accuracy is separate from what it takes to fix the Preview. Still the same math and H/W mapping solution.
Calculations shouldn’t be much more expensive than dewarping the original when the lid was closed but not with 2 known good locations. But My point was they should have already processed at least one new image to get close to the snap marks might as well at least pull that in and refine over time as better calculations become available.