Definitive CoreXY Calibration Guide.

Hi All.

I want to start designing a definitive CoreXY calibration guide. To that end, I have created a document on Google Docs that everybody with the below link can modify to improve:

CoreXY calibration guide draft on Google Docs

Please give me links to the best existing guides and test objects, and feel free to modify the Google Document and or make suggestions here.

When suitably completed this should be posted somewhere that it can be permanently hosted. Please suggest where....

Best,
Patrick

Apart from motor configuration in firmware and wiring, how would calibrating a CoreXY be any different from any other gantry system?

It just seems like you are trying to re-invent the wheel.

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!

If you have a guide that is already made for gantry systems you can point me to I will eliminate the steps for that and just list the guide for all applicable steps. However, so far I have found that information is found in many different places, and I have yet to find a full calibration guide that also lists links to all the appropriate test prints.

My idea is to create a very good guide with all steps and necessary insights and print files all in order and in one place.

The extruder should be calibrated before any other calibration is attempted, and the calculated average value of the measured filament diameter should be used when slicing or printing a calibration cube. I like Triffid Hunter's calibration guide in the RepRap Wiki, especially the extruder calibration routines. Here's my cartesian calibration cube and a companion spreadsheet that allows alignment of the axes to true orthogonality and to set the X, Y and Z steps per mm.

The relatively large size of the cube means it is easier to measure accurately because measurement error is a smalller percentage of the measurement than it is when measuring a much smaller object. If you set X, Y, and Z step/mm using a 20mm cube, and then print a 100mm cube, the errors in the settings at 20mm will multiply by 5 in each axis. If you calibrate using a larger object, say 100 mm on a side, if you print a smaller object, the error in the calibration will divide and get smaller with small objects.

I would be even better to print a cube at the maximum size the printer can handle, most people don't have access to any instrument that can measure a large object very accurately, and few printers can print anywhere close to the edges of their beds.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Thanks Dentist.... I knew you would have figured out a good procedure list. I will make the mods to the doc later today.

Quote
patrickrio
If you have a guide that is already made for gantry systems you can point me to I will eliminate the steps for that and just list the guide for all applicable steps. However, so far I have found that information is found in many different places, and I have yet to find a full calibration guide that also lists links to all the appropriate test prints.

My idea is to create a very good guide with all steps and necessary insights and print files all in order and in one place.

Ah, I thought you meant a calibration guide specific to a CoreXY gantry (like there is for a delta).

I think the one Dentist linked to is considered the "Standard". I never thought about considering the size of the object for calibration, but that definitely makes sense.

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!

I too like the idea of printing a larger calibration object. Never considered that before.

Thanks for sharing.

There are some differences in calibration for CoreXY from other printers. 1. both XY motors turn equal revolutions for any pure X or Y move. which of the 4 directions is determined by direction of rotation between the two motors.

2. Many of the existing CoreXY printers have belted Z axis movement with a cantilevered bed. Most printers in Triffid's guide have single or double motor screw drive Z axis. Will probably show 400 step motor belt drive cantilever, a single screw canilever, and a single motor dual screw with dual support bed. Will probably recommend against dual motor z axis.

I too used a large object to test XY accuracy, but did not yet test Z height accuracy mainly because the calculated XY values were exactly correct so I assumed I got it right for a belted Z axis too. Bigger things to measure is generally better. I tested my extruder by extruding 3 meters of filament with the nozzle removed and measured length to .5mm accuracy.

I am going to do the whole calibration again too using this guide as a test, and hopefully will eliminate a few niggling print setting issues in the process... mainly I think I need a systematic and comprehensive retraction calibration. my 250 micron layer spiral cube prints look like the ones DDentist generates so I am pretty sure I am on the right track.

Actually, my 100 micron prints look really good too.

Edited 2 time(s). Last edit at 12/09/2015 03:40PM by patrickrio.

OK, I put some information on the document, but will now be taking a 5 day break because I have an office move. Feel free to post more existing guides and test objects to include in the document. Thanks!

I'm curious. Why do you recommend against a dual motor z axis?

dual motors can get out of sync which will cause the bed to lose its carefully set level. Don't get me wrong, you can make it work. However, this is a calibration and selection guide for creating a highly accurate CoreXY printer. the possibility of losing a step or even a microstep in one motor and not the other makes the dual motor system a less accurate choice in my opinion and the opinion of many others who have worked toward a high precision printer.

Single motor driving multiple screws through a belt drive will work much better. Pick a bigger motor and drive all the screws at the same time with it. Once you set it up, it will not lose level. If the motor in a single motor system loses a step, it will auto correct with the next z- endstop check by your controller. A lost step with a single stepper in a dual stepper z system will not be auto correctable by the z endstop.

Note, we are talking about z axis with screws above and not belted movements without screws.

Edited 3 time(s). Last edit at 12/16/2015 09:52PM by patrickrio.

Hello everyone.

I recently built myself a corexy system. For some reason when I move the X axis the Y axis will also move. The Y axis moves fine. It is only when i try to move the X axis that the Y also moves. When I for example tell the X axis to move 100mm it only moves 75mm and the Y axis moves the other 25mm.

Any help?

Quote
JeandreCoetzer
Hello everyone.

I recently built myself a corexy system. For some reason when I move the X axis the Y axis will also move. The Y axis moves fine. It is only when i try to move the X axis that the Y also moves. When I for example tell the X axis to move 100mm it only moves 75mm and the Y axis moves the other 25mm.

Any help?

Sounds like a firmware issue to me. What firmware are you running?

---

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
JeandreCoetzer
Hello everyone.

I recently built myself a corexy system. For some reason when I move the X axis the Y axis will also move. The Y axis moves fine. It is only when i try to move the X axis that the Y also moves. When I for example tell the X axis to move 100mm it only moves 75mm and the Y axis moves the other 25mm.

Any help?

Check your belt tension.

You can also play around with the orientation of you motors. I know I had to try every configuration before I finally found the one that moved the carriage the way I wanted it to move.

---

greghoge.com

HUGE 3D PRINTER PARTS SALE!!!

I wonder what progress you have made on this guide - no mention of differences in belt tension between the two belts in the corexy gantry?

Not a criticism, I started writing a guide on proximity sensors and got a lot less far with it than I had hoped for.

This [forums.reprap.org] seems to have the cricitcal corexy bit in it, the rest of your guide is very useful info but is not corexy specific, its general 3d printer calibration information.

Edited 3 time(s). Last edit at 06/12/2016 05:05PM by DjDemonD.

---

Simon Khoury

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

Quote
patrickrio
dual motors can get out of sync which will cause the bed to lose its carefully set level. Don't get me wrong, you can make it work. However, this is a calibration and selection guide for creating a highly accurate CoreXY printer. the possibility of losing a step or even a microstep in one motor and not the other makes the dual motor system a less accurate choice in my opinion and the opinion of many others who have worked toward a high precision printer.

Single motor driving multiple screws through a belt drive will work much better. Pick a bigger motor and drive all the screws at the same time with it. Once you set it up, it will not lose level. If the motor in a single motor system loses a step, it will auto correct with the next z- endstop check by your controller. A lost step with a single stepper in a dual stepper z system will not be auto correctable by the z endstop.

Note, we are talking about z axis with screws above and not belted movements without screws.

What can cause one of the two motors to loose a step?????????
My newly build D-Bot has 2 Z motors and I have not detected any problems with the bed loosing it's mechanical level.

---

appjaws - Core XYUV Duet Ethernet Duex5
firmware 3.1.1 Web Interface 3.1.1
Ormerod 1-converted to laser engraver, Duet wifi
OpenSCAD version 2020.07
slic3r-1.3.0, Simplify3D 4.1.2, Cura-4.4.1

Well here's the theory recounted to me on a dark stormy night in an inn by the docks over a rum...

The two z motors when powered-up can stop at any micro-step position, but when powered off will snap to a full step position. If both motors are at exactly the mid point between two full steps (so for example step 100+8/16ths) and the power goes off, one might go one way to rest on step full step 100 and the other on full step 101. Now they are out of sequence and your axis on one side is a fraction lower than the other side.

I've always thought that whilst this is plausible, it can't happen that often and surely averages itself out over time even if it does... Could probably be counteracted by firmware before issuing a motor off command parking the two motors on a full step, might be a more legitimate use for servo-stepper/closed loop systems that seem to be going around at the moment.

This is the explanation given for (mainly i3's) bed's going out of the level in x direction over time and the reason why a 3-screw z-axis is better run from one stepper with belts (possibly with reduction gearing to enable lower currents/smaller motors).

Personally I use this belt driven z-axis (http://www.thingiverse.com/thing:1417053) on my corexy, which is accurate, and using a 0.9 deg stepper and 16T pulleys has a theoretical resolution of 5 microns, I can print at 0.1mm layer height no problems. The reason these are not more popular is just perception, we perceive screws to be more accurate for a z-axis, but we are perfectly happy to accept belts on x and y. Sure the bed goes into free fall if the motor is powered off, so I fitted a counterweight, now I can use lower motor current.

Edited 2 time(s). Last edit at 06/13/2016 11:33AM by DjDemonD.

---

Simon Khoury

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

sorry folks. have had a new contract since january that has sometimes taken more than 100 hour weeks.... sorry I have dropped the ball on this....

Quote
DjDemonD
Well here's the theory recounted to me on a dark stormy night in an inn by the docks over a rum...

The two z motors when powered-up can stop at any micro-step position, but when powered off will snap to a full step position. If both motors are at exactly the mid point between two full steps (so for example step 100+8/16ths) and the power goes off, one might go one way to rest on step full step 100 and the other on full step 101. Now they are out of sequence and your axis on one side is a fraction lower than the other side.

It's actually worse than this. The current when the motors are powered up forces the motors to snap to a full step position that matches the way the motor phases are energised. Only 1 in every 4 full step positions will match. So the motors will move up to 2 full steps each, possibly in opposite directions.

Quote
DjDemonD
I've always thought that whilst this is plausible, it can't happen that often and surely averages itself out over time even if it does...

My statistics is a little rusty, but from memory I think you can expect the difference in motor positions to increase approximately with the square root of the number of times you power it off and on again. See [en.wikipedia.org].

---

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 for the info.
dc42, is it possible to have a facility whereby the power OFF/ON routines ensure that the motors are at a full step? Could this be accomplished in firmware for startup and perhaps a macro for shutdown ?

---

appjaws - Core XYUV Duet Ethernet Duex5
firmware 3.1.1 Web Interface 3.1.1
Ormerod 1-converted to laser engraver, Duet wifi
OpenSCAD version 2020.07
slic3r-1.3.0, Simplify3D 4.1.2, Cura-4.4.1

Quote
appjaws1
Thanks for the info.
dc42, is it possible to have a facility whereby the power OFF/ON routines ensure that the motors are at a full step? Could this be accomplished in firmware for startup and perhaps a macro for shutdown ?

It would certainly be possible to add a "park" gcode that moves the bed down slightly such that the motors are at an exact multiple of 4 full steps from the starting position. On the Duet WiFi there is a power monitor, so it would be possible to do this automatically when the power goes down - assuming that when the power down is detected, there is sufficient power left to move the motors. This will be much easier with 24V power than with 12V power.

---

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].

Thank you, do I need to create a request for this facility to be included in the firmware?
It would certainly help users with 2 Z motor's .

Quote
dc42

Quote
appjaws1
Thanks for the info.
dc42, is it possible to have a facility whereby the power OFF/ON routines ensure that the motors are at a full step? Could this be accomplished in firmware for startup and perhaps a macro for shutdown ?

It would certainly be possible to add a "park" gcode that moves the bed down slightly such that the motors are at an exact multiple of 4 full steps from the starting position. On the Duet WiFi there is a power monitor, so it would be possible to do this automatically when the power goes down - assuming that when the power down is detected, there is sufficient power left to move the motors. This will be much easier with 24V power than with 12V power.

---

appjaws - Core XYUV Duet Ethernet Duex5
firmware 3.1.1 Web Interface 3.1.1
Ormerod 1-converted to laser engraver, Duet wifi
OpenSCAD version 2020.07
slic3r-1.3.0, Simplify3D 4.1.2, Cura-4.4.1

I know this is an old thread but I should mention a few things when others run across this thread.

Yes, you should use longer model to measure, longest you can accurately measure, for less percentage of error. Although this isn't the best way or the most accurate. Everyone seems to forget about shrinkage. Every filament brand and type will shrink at different percentage rate. So if you measure for one filament it will be off on another. Example. I printed a 200x10x5mm rectangle on X axis. I measured it right after it was done printing, leaving it on the hot build plate, and it measured 199.83, after I let it cool down it measured 199.43, sure it's not much but also how much did it shrink while it's printing.

The most accurate, and the hardest, way would be to measure the actual movement of the head or carriage a certain distance and you don't have to worry about the extruder being calibrated. It will be accurate, or more accurate, faster since you don't have to wait for the print to finish and you don't waste filament. Then you just need to figure out the shrinkage percent for the filament you want to use, when you need it to be more accurate, and adjust the model size in the slicer.

For the Z motors issue: I solved it by using 4 motors, that are individually controlled, using Duet 3 board using "True bead leveling" that uses a BLTouch testing at each Z attached location and it adjust all 4 corners to make it level. You can also do that with a 3 axis system also.

Happy printing everyone.

It's an interesting observation that the print is larger while it's hot on the machine, but most parts get used well away from the printer, after they have cooled, so calibration is usually best done based on measurement of a cooled print. 3D prints don't exhibit the same shrinkage that bulk plastic does because only a small part of the print, right behind the nozzle, is hot at any moment.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

That's fine if you use the same material from the same manufacturer. Obviously people use the printed part when it's cooled. And there's a big difference in shrinkage factor say from PLA to ABS. PLA is around 2% but ABS is around 8%, 4 times more shrinkage, and Nylon around 1.7% (Couldn't find a definitive site to quote on the actual percentages). Now your PLA calibrated machine and you print something in ABS that's 100mm wide is now off by 6mm. Not to include other factors, like heated build plate or not, and temperature of build plate, if using an enclosure or not. Different manufacturers use different formulas for the same class of material, like PLA, which can change the shrinkage percentage.

Doesn't really matter if you don't need dimensionally accurate part like a model of baby Yoda.

You're quoting bulk shrinkage numbers. That's not how 3D prints behave. Bulk shrinkage numbers apply to injection molded parts (and probably not exactly) where the entire object starts as molten plastic and cools as a single object.

Yes, ABS shrinks more than PLA and PETG, but no, a 100 mm ABS print will not measure 92 mm after you remove it from the printer unless your printer calibration is off by -8%.

In 3D prints, the plastic does 90% (guesstimate) of its shrinking by the time the nozzle is a few mm away (guesstimate) from the plastic it just deposited.

You don't need to calibrate the XYZ steps/mm for every different material you print.

Edited 1 time(s). Last edit at 03/04/2021 09:34AM by the_digital_dentist.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]