Just setting up new h2c. Vision encoder calibration almost makes it look like build plate is skewed. Had only minor variance on my h2d… anyone think its a problem
1 Like
It’s the same on my printer lol, I tried to correct it by doing belt tension adjustments but that didn’t work. I don’t worry too much though.
1 Like
Did you read the text?
The deviation seen in the image is simply for illustration purposes to see in which direction the points were off. But it doesn’t give an indication how far the points were off.
It actually does have a fixed scaling. I tested it by manually increasing some restriction on the right side and only the points on the right hand side moved when compared with previous test result. The left hand side points stayed basically the same. Which means that the illustration is at a fixed scale and the further it’s away from ideal, the worse the results.
Yes. I know that. However the points are much more out than on my h2d so there must be some relation to how far out things are.
1 Like
you know what’s funny? Run the calibration AGAIN, right after 
and it’ll show the same… again.
Or at least it did for me.
1 Like
Probably because running the calibration resets to factory setting first.
2 Likes
The calibration is compared to a perfectly squared printer not the deviations from previous calibration
3 Likes
which would make sense, there is no gain to use the last calibration.
the head is commanded to a position, and the position read from the optical plate, or, more likely, the head then moves to the exact center of one of those patterns on the plate, knowing the error it had from there.
a table is saved, and likely a NxM point interpolation table is formed.
I wish they gave the scale, or numeric deviations, or some numeric merit figure, like RMS xy or d error.
One way would be to print an actual array of little single layer dots, uncalibrated (is there a way to ‘forget’ the calibration? and then redo calibrated (could be different color, on same plate still with all dots).
