It’s just measuring the distance between two known and calibrated points. That’s the whole reason for the encoder plate.
Each QR code is a position identifier. It then measures how far it thinks it traveled vs how far it actually traveled.
The perfect placement of the encoder is not necessary.
But do we even think the calibration plate works in Z?
I assumed it was just X and Y and that Z was handled by the normal bed probing.
If the encod erplate did do Z, wouldn’t it be thrown off as soon as you removed it for printing?
No offense in what follows, it’s hard to convey tone and intent online.
The entire point of the encoder plate is to provide a calibrated and known distance measuring capacity for the camera. Each code is a know datum for the camera to reference. It does not make any difference if that plate is slight nudged in any other direction.
The reason for this is simple, it’s using the calibrated distance on the encoder to compare its distance traveled.
The take away point is this… it’s like a ruler for the printer to read.
A ruler doesn’t change just because you moved it.
Are you responding to me?
If so, I think you missed my point, since you repeated the same information as your previous post, but nothing I said contradicted it.
But the bed has alignment values as well where the head understands its initial position. If that is how the head homes/zeros itself, then plate alignment would matter if in relation to that origin.
I assumed you would have gathered an X/Y calibration doesn’t affect the Z calibration. It’s not included or altered in the encoder calibration.
Tool head homing position is not the same as a motion system accuracy calibration. It’s accounting for distance traveled and not used for homing/bed position.
I’m running the motion calibration test now with it shifted to the left, and will do the same with it shifted to the right and see what the results are.
Average should be much less than the original, I would assume.
Decided to get some samples.
Running the motion calibration again (just ran it yesterday)
196 avg, 393 max
Re-ran without touching anything
202 avg, 404 max
Shifted the plate to the right
198 avg, 387 max.
Seems the plate alignment does not impact the results.
Also seems like the 50micrometer accuracy is a bit exaggerated.
Perhaps they should run the test multiple times to get a gaussian distribution and just calibrate based on the average.
I’d also be interested to know what is being calculated and adjusted after the calibration. A map of the points measured would be nice.
If it’s like my CNC milling machine, it compares how much it was told to move with how much it should have moved based on the encoder plate readings, and then applies a scaling factor to compensate for the error.
My mill also has a separate backlash compensation function.
Presumably the H2D does not have that since it’s not mentioned, so I wouldn’t expect it to be able to correct for backlash.
I think this is what I am not understanding. what does avg/max mean? What is being changed with calibration? How far off is a factory printer going to be and how long will it stay that way? How often is recalibration recommended? What tasks does calibration apply to/benefit from? Is calibration from the factory recommended to achieve the presented accuracy?
Reading the marketing and this thread covers many of the questions but not all of them for me.
I misspoke; I meant to say “it compares how much it did move with how much it should have moved based on the encoder plate readings.”
Hah… that’s a strange looking pony.
I knew it was just time until someone did it. You just a rebel eh? The plate says dont do it, and you just had to! 
Nah, he clearly had a… vision!
someone’s rolling in their grave over that one
Did it leave any marks 25
I was just being silly, I didn’t really do it. The plate came in the mail today and I had posted on another thread “challenge accepted” when someone said you couldn’t print on it.
Ow I looked at the brim on the prime tower and thought it looked to nice to just be placed there
