Online least-squares delta printer calibration calculator

I have provided a delta printer calibration wizard at [escher3d.com]. This uses the same least-squares algorithm that I implemented in RepRapFirmware. Instructions for using it are at the top of the page.

If you try it, please let me know how well it works (or doesn't work) for you, and which printer firmware you use. In particular, I am not sure that the sign of the M666 endstop corrections is correct for all firmwares.

Also available is a wizard for generating a bed.g file for RepRapFirmware, at [escher3d.com].

---

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].

Quote
dc42
I have provided a delta printer calibration wizard at [escher3d.com]. This uses the same least-squares algorithm that I implemented in RepRapFirmware. Instructions for using it are at the top of the page.

If you try it, please let me know how well it works (or doesn't work) for you, and which printer firmware you use. In particular, I am not sure that the sign of the M666 endstop corrections is correct for all firmwares.

Dave

Like the Bed.g generator however 2 things that I can see

1. There are no point designators in the result (P0 thru P12 for the 13 point output)
2. The probe offset is applied for all points which for example if a offset in Y of +25mm is given and a probe radius of 140 (For a 250 dia bed say) it generates a probe point in the Y- direction that is outside of the Print radius

See attach snip of the output I get hope this is helpful


Doug

Quote
dc42
Also available is a wizard for generating a bed.g file for RepRapFirmware, at [escher3d.com].

Thanks Doug, I'll attend to those issues when I have finished testing my new firmware release.

---

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].

I have corrected those issues with the bed file generator. A few people have used the delta calibration wizard and not found any issues with it. I have confirmation that the endstop adjustments do have the correct sign when running Marlin firmware.

Edited 1 time(s). Last edit at 12/08/2015 03:40PM 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].

Quote
dc42
I have corrected those issues with the bed file generator. A few people have used the delta calibration wizard and not found any issues with it. I have confirmation that the endstop adjustments do have the correct sign when running Marlin firmware.

Looks good now thanks

Doug

The output for Smoothieware can have an invalid positive number for the end stop correction.

The M665/M666 output doesn't match the Smoothieware documentation at [smoothieware.org]



delta_tower1_offset
delta_tower2_offset
delta_tower3_offset
delta_tower1_angle
delta_tower2_angle
delta_tower3_angle

alternatively the values can be changed live (and saved with M500):-

M665 Axxx Bxxx Cxxx Dxxx Exxx Hxxx

where A is delta_tower1_offset, B is delta_tower2_offset, C is delta_tower3_offset
where D is delta_tower1_angle, E is delta_tower2_angle, H is delta_tower3_angle

etfrench, thanks for your feedback.

I have updated my online least squares delta calibration calculator at [escher3d.com]. The main changes are:

* Fixed three bugs that caused the calibration to be less then completely accurate.
* Added a separate firmware option for Rich Cattel Marlin, which uses different M665/M666 parameters compared to standard Marlin.
* Added M665 parameters D E H Z when the selected firmware is Smoothieware
* Normalise the endstop corrections to be negative for Smoothieware

If you already used this calculator to calibrate your machine, you should get more accurate results with the new one. Please let me know how well it works for you.

Edited 1 time(s). Last edit at 01/01/2016 03:30PM 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].

Hello, Im been using the webtool to calibrate my micromaker, but Im unsure what value is the one I need to change with the "New tower position angle corrections: X Y Z" Im using Ramps 1.4 with repetier.

Repetier config tool includes the tower position angles, but they call them ABC instead of XYZ. You need to add the tower position angle corrections to the default tower position angles in Repetier.

---

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].

Hi DC42

Do you have any results comparaisons with your calculator ?
I use it to calibrate my delta printer, and it's very long !!!
My problem : i'm not sure of the arm lenght, because I use magntic joints : [www.youmagine.com]
The balls are on the effector and the carriage, not the diagonal rod.
So I don't konw what is the start point and the end one to do the measure.

My resultt (6 tries) with 10 points :
Printable bed radius: 150 mm

Before the test :
M665 R168.51 L347.00 H487.40 B150.00 X-0.48 Y-0.65 Z0.00
M666 X-0.61 Y0.54 Z0.07


Results : (P0 to P9)
0.05
0.00
-0.10
-0.10
0.00
0.05
-0.05
0.15
0.05
-0.10

After the test :
Factor 6
M665 R168.47 L347.00 H487.39 B150.00 X-0.52 Y-0.61 Z0.00
M666 X-0.63 Y0.57 Z0.06
Factor 7 :
M665 R168.59 L347.38 H487.40 B150.00 X-0.52 Y-0.61 Z0.00
M666 X-0.63 Y0.57 Z0.06


Before : 0.08
After : 0.07


The #7 point sucks... and I don't why this result is so high....
So what is a good result with your calculator ?

Thanks

Edited 1 time(s). Last edit at 09/09/2016 10:07AM by shofman.

The diagonal rod length you should use is the distance between the centres of the balls. If you have any spare balls then you should be able to measure that.

For 7 factor calibration to work at all well, you need to probe between each pair of towers as far out as you can. Otherwise there is insufficient info in the height errors for the least squares calculator to get an accurate value for the diagonal rod length.

However, your 6 and 7 factor results are remarkably similar apart from diagonal rod length and delta radius being slightly different, so the 7 factor calibration may be working well in your case.

Any RMS error better than 0.1 is good. You appear to be measuring the height errors to the nearest 0.05mm, and you will need greater measurement accuracy to do much better.

---

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 I use the center of the balls, the diagonal rods ends will be about 349mm :
Carbon rods : 307mm
2 magnetic : 15x2 mm
2 1/2 balls : 2x6mm

And the delta radius is about 170.4mm : from the center of the balls to the center of the over. Can you confirm this is good ?

Ok I will try with 0.01mm height error.

another question : do you calibrate the printer in hot conditions ?
For example :
Hotbed 60° C
Hotend 180° C

Yes, I heat the hot end and bed to operating temperature before running auto calibration. It affects the Z=0 height slightly.

---

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].

OK, my results in hot conditions (bed 60° and hotend 180°C)
Nothing better than 0.08 ...

Before (3 tries before that)
M665 R169.24 L349.00 H487.08 B150.00 X-0.47 Y-0.69 Z0.00
M666 X-0.50 Y0.42 Z0.08

G1 X0 Y150 Z30 F5000
G1 X129.9 Y75 Z30 F5000
G1 X129.9 Y-75 Z30 F5000
G1 X0 Y-150 Z30 F5000
G1 X-129.9 Y-75 Z30 F5000
G1 X-129.9 Y75 Z30 F5000
G1 X0 Y75 Z30 F5000
G1 X64.95 Y-37.50 Z30 F5000
G1 X-64.95 Y-37.50 Z30 F5000
G1 X0 Y0 Z30 F5000

Measures :
-0.07 (Z tower)
-0.01
-0.12 (in front of Y tower... strange ?)
-0.01
0.03 (X tower)
0.07
-0.07
0.14
0.10
-0.10 (Center)

deviation before 0.08 after 0.07
Propostion :
6 factors :
M665 R169.19 L349.00 H487.07 B150.00 X-0.51 Y-0.64 Z0.00
M666 X-0.59 Y0.44 Z0.15

7 factors (after 0.07 too)
M665 R168.82 L347.83 H487.06 B150.00 X-0.51 Y-0.64 Z0.00
M666 X-0.59 Y0.44 Z0.15

I have a great difference between the 3 center points (the 3 last)... So the first layer is not good at all. Brim is awfull.
What do you think about this results ?

The pattern of height errors doesn't suggest anything in particular to me, so I think either your print bed is not flat or you have some of the typical geometrical errors in your delta build. Try putting a bull's-eye spirit level on the effector and watch it for signs of tilt as the effector moves. Also put a straight edge across the bed to check is flatness.

---

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].

Ok I will do this.
Do you have a checklist of the typical geometrical errors in a delta build ?

Quote
shofman
Ok I will do this.
Do you have a checklist of the typical geometrical errors in a delta build ?

Rods not all the same length between bearings
Bearing spacing not quite the same at the top of a pair of rods as at the bottom
Carriage rotated about its face so that the two bearings on it are not at the same height (bearings on the effector being at different heights will have the same effect)
Carriage rotated about the Z axis so that the rods attached to it are trying to twist the effector a different amount than the other carriages
Play in the joints, which is taken up in different directions depending on effector XY position, due to the Bowden tube exerting sideways forces in different directions

---

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].

Is it important that the hotend is perpendicular with the effector ? I think I have a problem with this (play between them ).

Quote
shofman
Is it important that the hotend is perpendicular with the effector ? I think I have a problem with this (play between them ).

If there is play between the hot end and the effector, you definitely need to get rid of it.

---

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].

Hi DC42,

I have made many improvments in my printers...
1/ V-slot carriage modification : Back plate
2/ Hotbed plate
3/ endstop NC (before NO)

So the results are good I think : 0.03 !!!
I must do a 7 factors configuration to have this value.
With only 6 factors, I cannot do better than 0.11.

My results :
Before / after 0.11 with 6 factors
M665 R164.74 L349.00 H470.86 B150.00 X-0.15 Y-0.83 Z0.00
M666 X0.36 Y0.31 Z-0.66
0.16
-0.19
0.08
-0.14
0.14
-0.11
-0.02
0.05
0.04
-0.08


Before 0.11 / after 0.03 with 7 factors
M665 R166.96 L356.55 H470.90 B150.00 X-0.14 Y-0.79 Z0.00
M666 X0.36 Y0.27 Z-0.62
0.02
0.00
-0.02
0.02
-0.01
0.02
-0.05
0.04
0.02
-0.06

What do you think about this results ?

Those results look very good to me. The real test is to measure the height errors at points you didn't feed in to the calibrator.

---

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].

Thanks. Do you have any formula or example of points to test?

Extra points around the edge are good, so are points half way between edge and centre.

---

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].

I have tried these points :
G1 X0 Y150 Z30 F5000
G1 X129.9 Y75 Z30 F5000
G1 X129.9 Y-75 Z30 F5000
G1 X0 Y-150 Z30 F5000
G1 X-129.9 Y-75 Z30 F5000
G1 X-129.9 Y75 Z30 F5000
G1 X0 Y75 Z30 F5000
G1 X64.95 Y-37.50 Z30 F5000
G1 X-64.95 Y-37.50 Z30 F5000
G1 X0 Y0 Z30 F5000

So these kind of points are ok for you ?
X-32.5 Y18.8
X-32.5 Y18.8
X0 Y-37.5

I typically generate a few hundreds of points in a regular grid over the hole bed surface.
You are going to print over the whole bed. So you can just as well test the whole bed too.

Ok I will try...

David, do you have the opportunity to add simple M665 and M666 command parameters parser to fill initial parameters of calculator? Just field (like commands field at the end of calculator) where user can enter M665 and M666 commands from his config to fill initial values.

Also simple list of G1 commands to move effector close to each probe point would be useful too. The user will be able to copy command directly to console to move effector to next probe point.

Quote
roboprint
David, do you have the opportunity to add simple M665 and M666 command parameters parser to fill initial parameters of calculator? Just field (like commands field at the end of calculator) where user can enter M665 and M666 commands from his config to fill initial values.

Also simple list of G1 commands to move effector close to each probe point would be useful too. The user will be able to copy command directly to console to move effector to next probe point.

That's all possible to do, however I have to implement an urgent requirement for additional axes and IDEX support in RepRapFirmware. So it's not likely to happen any time soon unless someone else contributes the code.

---

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].

Hi David, any chance of the least squares online calculator being updated to do 8 and 9 factor calibration, and to auto generate more than 7 points?

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions