Frustrating calibration issue
Posted by Moriquendi
|
Re: Frustrating calibration issue October 20, 2017 01:59PM |
We have a wiki page on this at [duet3d.com]. Obviously effector tilt won't affect your Z probe reading if you are using a piezo to detect nozzle contact with the bed, but it is still indicative of geometric errors.
If you were using a Duet, you would have the benefit of least squares auto calibration permanently built into the firmware, and a graphical display of the height map.
Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod
Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].
If you were using a Duet, you would have the benefit of least squares auto calibration permanently built into the firmware, and a graphical display of the height map.
Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod
Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].
|
Re: Frustrating calibration issue October 20, 2017 05:45PM |
The pattern is quite symmetrical. Lower at the towers. Higher between them. I would not expect symmetrical pattern from a tilted bed.
Here is an example of an error similar to what Moriquendi has:
The above error is caused by errors in both diagonal rod length and delta radius. Make sure you have diagonal rod length precisely entered. Calibrate endstops and delta radius (tower positions).
Here is an example of an error similar to what Moriquendi has:
The above error is caused by errors in both diagonal rod length and delta radius. Make sure you have diagonal rod length precisely entered. Calibrate endstops and delta radius (tower positions).
|
Re: Frustrating calibration issue October 21, 2017 04:54PM |
My experience is that printer errors are mostly smaller than geometry errors after calibration. For example, the hight error maps I posted in my previus message simulated geometry errors of 1 mm. Note that this resulted in bed level error less than 0.0.3 mm (because the errors partially compensate each other).
Here is an example of 1 mm error in x coordinate of Z tower:
Notice that 1 mm tower position error resulted in about 0.4 mm bed height error.
Here is an example of endstop error of 1 mm.
Notice that the resulting bed error is almost planar. That means that endstops can quite well compensate for small bed tilt. The problem is that if you use this feature then your printed parts will be as tilted as the bed is.
executivul: 0.1 mm endstop error is too much. Either adjust the endstop, or (if its error is varying) replace it.
Here is an example of 1 mm error in x coordinate of Z tower:
Notice that 1 mm tower position error resulted in about 0.4 mm bed height error.
Here is an example of endstop error of 1 mm.
Notice that the resulting bed error is almost planar. That means that endstops can quite well compensate for small bed tilt. The problem is that if you use this feature then your printed parts will be as tilted as the bed is.
executivul: 0.1 mm endstop error is too much. Either adjust the endstop, or (if its error is varying) replace it.
|
Re: Frustrating calibration issue October 22, 2017 03:28PM |
I don't want to hijack the topic, but I'll keep posting here maybe moriquendi has the same problem
1. I've rechecked the machine is square and mechanically adjusted the endstops untill I get:
2. Ran the LSQ smoothie branch to calibrate and G29 P1 I17 to map the height, I get 0.16mm of difference:
3. Probed and activated grid compensation and reprobed with G29, I get 0.0877mm of difference:
Calibration grid had 15 points, probing 17, so points don't overlap and I get real values.
4. Tried printing a circle aaaannnndddddd:
1. I've rechecked the machine is square and mechanically adjusted the endstops untill I get:
Recv: Calibrating Endstops: target 0.030000mm, radius 155.000000mm Recv: set trim to X:0.000000 Y:0.000000 Z:0.000000 Recv: initial Bed ht is 480.934326 mm Recv: center probe: -0.0022 Recv: T1-0 Z:4.8831 Recv: T2-0 Z:4.9082 Recv: T3-0 Z:4.9019 Recv: trim already set within required parameters: delta 0.025085 Recv: Calibrating delta radius: target 0.030000, radius 155.000000
2. Ran the LSQ smoothie branch to calibrate and G29 P1 I17 to map the height, I get 0.16mm of difference:
3. Probed and activated grid compensation and reprobed with G29, I get 0.0877mm of difference:
Calibration grid had 15 points, probing 17, so points don't overlap and I get real values.
4. Tried printing a circle aaaannnndddddd:
|
Re: Frustrating calibration issue October 26, 2017 07:29PM |
Quote
mikes3ds
What you guys using to plot your height maps. I am thinking about using R, but it seems there are some ones out there already?>
On the Duet the height map plot is built into the standard web interface. Example:
You can rotate and zoom it, and hover over the dots to see the height error at each probe point.
The rendering is done in the browser using JavaScript, which takes data read from the height map file which is in CSV format. The credit for implementing this belongs to Christian Hammacher, not me.
Edited 2 time(s). Last edit at 10/26/2017 07:34PM by dc42.
Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod
Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].
|
Re: Frustrating calibration issue October 26, 2017 10:12PM |
My zprobe info. Graph using R (in mm)
Pretty bad, going to see what the cause is.
Edited 2 time(s). Last edit at 10/26/2017 10:15PM by mikes3ds.
Sorry, only registered users may post in this forum.