Thread Rating:
  • 2 Vote(s) - 5 Average
Z Axis Height Calibration
#1
So there has been some speculation on the inaccuracy of the Z height for prints that are attached directly to the build plate and since I've had some free time the past few days, I've decided to delve into this issue first hand.  

There has been another use in the FB group who did a nice little write up in regards to the spring in the build plate as well as the spring washers internal to the stepper and I wantd to have a solution that was non-intrusive to be messing with anything like that.

Sooooo, here are my findings...

I created a simple 20mm calibration cube and printed it.  Checked it's z height and measured 19.6mm, so a 0.4mm height difference.

Now, since I didn't feel the need to be pulling apart steppers and taking apart the build plate to remove a spring, I dove into the software side of things and unlocked the Z=0 menu option.  Now since my particular test print was exactly 0.4mm short, this was easy.  Simply, rehomed the build plate through the menu, then set the 0.1mm option and tapped the up arrow 4 times to equal the 0.4mm difference (19.6mm cube + 0.4mm offset = 20mm), then I went back one menu option and tapped the Z=0 and confirmed it.  This has now set my Z axis home position to 0.4mm above the z axis stop.

I ran the calibratin cube again and low and behold...  measured exactly 20mm on the z height Smile

Right now I'm running a 5mm, 10mm, 15mm and 20mm cube at the same time to be able to verify multi-height accuracy and that this wasnt just for the 20mm cube size.

I've also conjured the way to be able to offset in values that are outside of the printers 0.1mm, 1mm and 10mm presets, but I'll address that shortly as well Smile

So for right now, no need to mess with taking anything off the printer to be able to adjust for z height accuracy...

Update...

Pulled the set of calibration cubes from the printer and damn, all are good, so that shows that this works for any height of print Smile  Now I just have to package everything and do the writeup for the "How To", hopefully will have that done and up later this evening.

Update...

Here are the files and walk though instructions needed, ENJOY!!!  I'll work on getting a video together next to also show the process...

Update...

If your offset is upwards of 1mm or more, then you will most likely need to mod the stepper with the washer or add an additional spring washer, as this means the stepper probably has a single spring washer installed instead of 2.  Typical offset would be in the range of 0.3mm - 0.5mm.


.zip   Z Height Accuracy.zip (Size: 286.45 KB / Downloads: 1,404)
Reply
#2
Clever solution!  I'll be re-leveling/zeroing my build plate ASAP..  Thanks for sharing.
Reply
#3
(10-17-2019, 04:46 AM)CPatrick Wrote: Clever solution!  I'll be re-leveling/zeroing my build plate ASAP..  Thanks for sharing.

Awesome!!!

Please post your results afterwards Smile
Reply
#4
Does this dimensional accuracy also hold on a taller model - say 20,40,60 tower?
Reply
#5
(10-17-2019, 03:16 PM)marsman Wrote: Does this dimensional accuracy also hold on a taller model - say 20,40,60 tower?

Yes, this will address accuracy for the entire z axis.  
I originally just tested with a 20mm cube, but after making the change, printed a 6, 10, 15 and 20mm cube and all were accurate, so should cover the entire z axis Smile
Reply
#6
Bonjour à tous
Je viens de découvrir ce forum et comme j'ai reçu mon imprimante depuis moins de 24 heures, je me suis tout de suite rendu compte du défaut Z.
Pour moi je pense que le problème vient surtout de la vis trapézoïdale parce que travaillant dans la mécanique je sais qu'il y a toujours du jeu
dans ce type de vis et c'est normal cela s'appelle le jeu fonctionnel.
Je pense qu'il aurait plus judicieux a la marque ELEGOO de nous proposer ce produit avec une vis a billes parce-que sans jeu du tout
quitte à payer l'imprimante un peux plus chère.
Je suis en train d'imprimer un cube de 20x20 mms pour faire le test mais ou doit on intervenir pour le Z dans le logiciel ou sur l'imprimante
directement?
Vous l'aurez compris je suis de FRANCE vous n'aurez qu'a traduire comme je viens de le faire.
Merci à tous et bonne journée.
Reply
#7
salut,

Malheureusement, le problème n'est pas créé par la vis de type trapézoïdal. J'ai également testé en enlevant l'écrou anti-jeu et remplacé par un écrou standard TR8x2 qui avait un jeu nul et obtenu les mêmes résultats.

le correctif est en fait assez simple et se fait en modifiant le fichier de configuration gcode des imprimantes.
Reply
#8
Bonjour à tous,


Pour ma part le correctif logiciel proposé par "Admin" ne fonctionne pas et le décalage du Z0 pose des problèmes d'adhérence des impressions (mon décalage sur le cube de 20mm était de 0.8mm avant modification du Z0 !).

Pour vérifier que le problème persiste après la calibration proposée, il suffit d'imprimer un escalier avec des marches de 1mm à 10mm et de contrôler la hauteur imprimée de chacune des marches (fichier TEST.STL joint).

Les différences de hauteur entre le modèle STL et le résultat imprimé varient suivant les marches et de toute évidence, ce type d'erreur est impossible à corriger par une simple modification du Z0 (plus l'épaisseur est faible, plus l'erreur est importante).

Je propose donc à Admin d'imprimer le fichier STL joint et de nous donner honnêtement les hauteurs mesurées pour chaque marche de son impression...

Je pensais pouvoir utiliser l'imprimante pour réaliser des pièces techniques - un peu plus complexes que de simples cubes  Big Grin - malgré la très bonne qualité visuelle des impressions, je suis déçu par le manque de précision de la machine.

Je pense, comme beaucoup d'autres, que ce problème est bien lié à un problème de précision de l'axe Z sur l'imprimante.

Bonne soirée à tous.


Attached Files
.zip   TestStl.zip (Size: 792 bytes / Downloads: 15)
Reply
#9
Actually, I don't understand how this modification of the Z-axis home position affects dimensional accuracy.  No matter the Z-axis home position, the build plate must be "leveled" within 0.1mm of the vat's FEP film.  The build plate is in this location at the beginning of each print job, and always moves up the Z-axis from here; correct?  It's the z-axis stepper motor settings that need to be modified to correct z-axis accuracy, in my opinion.

This GCODE line: M8010 S0.000625 ;DO NOT CHANGE

The calculation formula: lead /((360/1.8)*16)

In the code you provided, S0.000625 would equal a "lead" of 2.

Maybe I'm whacked.  Let me know.
Reply
#10
Reason is some play/air in the z-axis with springs and such in all the mechanics it seems.
Elegoo is providing modified z-axis-motors on request on goodwill and very friendly.
Workaround!

Change the settings in the machinesettings gcode https://drucktipps3d.de/wp-content/uploa...eset.gcode
from M8083 from I0 to I1 to make the set-zero-button working.
Level with the level-button and still loose platform.
Lower the platform 1 or 2 millimeters and level your plate.
Lift the platform some 2 millimeters and lower it agai in 0,1 mm steps
Set zero once you get your desired zero, paper barely moving.
Remove G28 Z0 in the chitubox Gcode settings.
For that navigate into the istallation folder and find the elegoo settings and set show gcode from 0 to 1.
Delete your printer and read it in anew. Then the gcode settings will be shown...
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)