Need some help calibrating Kossel mini.

So I've been using M666 (endstops), M206 (Z-hight) and M665 (deltaradius) to calibrate bed level and is fairly pleased with it for now.

But my prints are dimensionally inaccurate so I have followed this guide: [delta-calibration.s3-website-us-west-2.amazonaws.com]
Printed a 50% version of the asterisk he links to on that page and from how I understands it you should measure each line starting closest to a tower and moves away from it in a straight line (might be wrong on this, english is not my native language)
So the lines come out like this:
X=102,1mm
Y=101,9mm
Z=101,6mm

So I add the lines suggested by the document, all of them, last under the "Delta Settings"-part with a line modified to compensate for the Z-inaccuracy:
"#define DELTA_TOWER3_CORRECTION 0.8 // back middle tower"

Flash it, performs a quick bed-leveling and hits print.

The print come out identical to the first one.
Went back to the firmware and increased the value, actually to 3.2, and then does another print.

Still identical....

What do do now??

EEPROM disabled?

Quote
sungod3k
EEPROM disabled?

Nope, as I wrote, I used M666 (endstops), M206 (Z-hight) and M665 (deltaradius) for bed-level calibration.

Or do you mean I have to disable it for it to work?
But then I would loose my bed-calibration, and the ability to use the M666 to calibrate it again...?

Edited 1 time(s). Last edit at 06/22/2015 12:18PM by Koenig.

Quote
Koenig

Quote
sungod3k
EEPROM disabled?

Nope, as I wrote, I used M666 (endstops), M206 (Z-hight) and M665 (deltaradius) for bed-level calibration.

Or do you mean I have to disable it for it to work?
But then I would loose my bed-calibration, and the ability to use the M666 to calibrate it again...?

I am not that great with the software so I will try to explain best I can.

The EEPROM is a way to save/backup/restore different sets of variables quickly. After entering in whatever settings you want to print with using your M666, M206, and M665 you should send the M500 command to save them to the EEPROM. There is also the command M502 which will load the settings from the eeprom. M502 comes in handy if you are trying different settings and want to go back to the old settings quickly.

I have seen many posts on here stating if you have different settings saved to your eeprom vs. your active settings it can cause problems.

Quote
jculver09

Quote
Koenig

Quote
sungod3k
EEPROM disabled?

Nope, as I wrote, I used M666 (endstops), M206 (Z-hight) and M665 (deltaradius) for bed-level calibration.

Or do you mean I have to disable it for it to work?
But then I would loose my bed-calibration, and the ability to use the M666 to calibrate it again...?

I am not that great with the software so I will try to explain best I can.

The EEPROM is a way to save/backup/restore different sets of variables quickly. After entering in whatever settings you want to print with using your M666, M206, and M665 you should send the M500 command to save them to the EEPROM. There is also the command M502 which will load the settings from the eeprom. M502 comes in handy if you are trying different settings and want to go back to the old settings quickly.

I have seen many posts on here stating if you have different settings saved to your eeprom vs. your active settings it can cause problems.

Yeah, I know of what the EEPROM is and wich commands I need to use to utilize it, that is the way I did my bed-leveling as I mentioned in my first post, bed is leveled and it sticks.
Now I'm trying to correct the dimensions of the prints and the only way I have found to do this is described in this guide: [delta-calibration.s3-website-us-west-2.amazonaws.com]

I'm trying to calibrate the dimensions of the prints, not bed-level or scale.

Increasing length of diagonal rod reduces printing size.
Decreasing length of diagonal rod increases printing size.

New_diagonal_rod = measured_length / target_length * old_diagonal_rod

For example, you print a 50.00 mm x 10.00 mm x 10. 00 mm cube.

Quote

1st iteration

Measured length of the cube is 49.80 mm in X when diagonal rod is 203.00 mm.
New_diagonal_rod = 49.80 / 50.00 * 203.00 = 202.19 (round off)

M666 D202.19
M500

2nd iteration

If the cube's X length comes out 50.15 mm.
New_diagonal_rod = 50.15 / 50.00 * 202.19 = 202.80 (round off)

M666 D202.80
M500

...

Print the cube again, measure it, apply the formula to get new diagonal rod length.
Keep doing it until you get the accuracy range.

Edited 1 time(s). Last edit at 06/22/2015 04:02PM by janpenguin.

Quote
janpenguin
Increasing length of diagonal rod reduces printing size.
Decreasing length of diagonal rod increases printing size.

New_diagonal_rod = measured_length / target_length * old_diagonal_rod

For example, you print a 50.00 mm x 10.00 mm x 10. 00 mm cube.

Quote

1st iteration

Measured length of the cube is 49.80 mm in X when diagonal rod is 203.00 mm.
New_diagonal_rod = 49.80 / 50.00 * 203.00 = 202.19 (round off)

M666 D202.19
M500

2nd iteration

If the cube's X length comes out 50.15 mm.
New_diagonal_rod = 50.15 / 50.00 * 202.19 = 202.80 (round off)

M666 D202.80
M500

...

Print the cube again, measure it, apply the formula to get new diagonal rod length.
Keep doing it until you get the accuracy range.

Yes, know about this as well, as I wrote in the post above, getting to scale after correcting tower positions.

To clarify, If I print a perfect square the sides come out at slightly different sizes, so I want to correct the DIMENSIONS not the scale, in other words I want all the sides of the square be the same length before I adjust diagonal rods for scale.

Anyway, found another way of doing it: [www.thingiverse.com]