# Non Squared Cubes

Posted by SEUTec
 Non Squared Cubes October 22, 2016 06:45AM Registered: 5 years ago Posts: 8
Hi all:

I am getting non squared cube while X Y scale calibration test for an 80x80x2 mm cube.

See attached fig. For M579 X1.00 Y1.00 Z1.00

I can adjust X Y scales by

M579 X1.066 Y1.074 Z0.98

But the vertical lines remains unparallel.

How can I solve this ?
This is a Firmware Bug ?

Regards.
 Re: Non Squared Cubes October 23, 2016 01:35AM Registered: 6 years ago Posts: 188
When I printed a con model after six factor calibration, it printed the deformed con in rightside.
I found it the six factor cablibration introduced excessive angle variation of towers which deform geometry of any model.
To check the current delta configuration, run M665 in gcode console on DWC.

The calibration assumes all 12 ball joints are square and in parallel, three towers are perpendicular to the bed.
And it will create best delta configuration to compensate the deviation based on probe points.

I learned the critical steps that impact calibration later on are the assembly of ball joints in carriage and effector.

A pair of diagonal rods require parallelism:
- the width of two ball joints is within 0.25 mm both carriage and effector side.
- length variation of two diagonal rods is within 0.1 mm.
(It will unlikely happen because of the laser cut precision.)
 Re: Non Squared Cubes October 23, 2016 04:22AM Registered: 5 years ago Posts: 8
Hi:

I have tested 4 points test and I don't get enough good autoleveling.
I need to test 7 points to get a good autolevel.

But I don't have parallel arms. I will correct this and see what happens.
I have a difference of 4.2 mm from effector to carriage.

I have changed a lot of pieces in my Fisher and this may the problem.

Thank you very much, for your help.

Regards.
 Re: Non Squared Cubes. Problem Solved. October 24, 2016 04:24PM Registered: 5 years ago Posts: 8
Hi all:

I have reprinted a dimensional adjusted effector to return to the original Fisher arms axis of rotation and reestablish arms parallelism.

Now works better. I have reduced an error of 2mm to less than 0.5mm.

Thank you very for your help janpenguin.

Regards.
Sorry, only registered users may post in this forum.