Amount of Z-axis movement is unstable

General Discussion
1

When I printed in spiral mode, Although the flow rate is constant, the movement amount of the Z axis is uneven.
As a result, the width of the print changes periodically.

IMG_2752
IMG_27521612Ă—1948 486 KB

the gcode file is made by Cura.
This issue does not occur with marlin on same printer same gcode file.
TXXPRO_test_c_spiral2.ZIP

As strange point ,Looking at mainsail on printing.
Flow : stable
Z-position : increase unstably

Are there any factors that change the amount of increase in the Z axis?
or Is it truncated by some calculation?

My printing environments
Tronxy X5SA Pro + Orbiter v1.5
Cura v4.13.1
Klipper /mainsail up to date

2

Additional information
printer.zip
klippy-2.log.zip

3

Hello @boyon !

I had similar issues. A PID tuning of the bed fixed it.

4

After updated bed PID ,but my printer is not fixed.
Bed PID seems to be correct.
Amount of Z-axis movement is unstable.

image

5

There are only three factors that can influence your Z values:

  1. The GCode produced by your slicer
  2. Your bed mesh compensation
  3. Some fixed Gcode offsets, e.g. by using SET_GCODE_OFFSET

For me this looks like a mechanical issue with your Z axis as it is nicely periodical. Most likely the axis is binding or has extreme wobble.
I do not know the Tronxy printer but also your Z axis speed seems a bit optimistic for a lead screw:

[printer]
kinematics = corexy
max_velocity = 300
max_accel = 5000
max_accel_to_decel = 3000
max_z_velocity = 25
max_z_accel = 30

Try max_z_velocity = 8

6

I tried max_z_verocity=8. but not good.
but you gave me good hint.
max_z_accel =30 seems too small.
now I am trying “100” , and better than used to.
Thank you .

7

There is definitively something going on with your Z-axis. 100 is much too fast for a lead screw. You should mechanically check this axis.
You could try a higher accel but really lower speed. Something along:
max_z_accel: 120

8

Sineos-san
I got some good tips from you twice.
The cause seems to be this.

image
image413Ă—552 42.2 KB

I taped the coupling and it worked more correctly.
Should I replace it with a rigid coupling ?

9

My pleasure, boyon-san

What I would do:

  • Remove the lead-screws
  • With removed lead-screws make sure the bed moves freely on the linear guides without binding or bumping (these linear guides and rods often are really crappy)
  • If it does not move completely freely, then try to align the rods. With 4 rods the alignment must be pretty perfect to avoid binding
  • Install the lead-screws and make sure to align them perfectly perpendicular to your bed. Avoid any skew

This would need perfect alignment. I would go for flex coupling or even use a so called Oldham coupling.
Flex coupling:
https://www.aliexpress.com/item/1005001469408944.html?spm=a2g0o.store_pc_groupList.8148356.2.5c4c5066Q3NmEo

Oldham Coupling
https://www.aliexpress.com/item/1005001525548983.html?spm=a2g0o.order_list.0.0.21ef18029z33wP

10

Question from the image you have posted: What is the purpose of the belt? Doesn’t the Tronxy have two independent Z-stepper?

11

Thanks for Links.
I’ll purchase one of them.
X5SA Pro has two independent Z-axis.
but the stock do not have the belt.
It’s optional parts.(but it’s quite necessary)

a lot of garbage … :laughing:

image
image922Ă—528 69.9 KB

12

I do not know this printer and I thus do not know why it is necessary but from a mechanical point this looks like a very bad idea:

  • The belt has a certain tension thus trying to pull the lead-screw inwards
  • To compensate for alignment errors the lead-screw has this flexible couplings (which makes sense)
  • As a result you will inevitable get a skew on the lead-screw especially on the upper end as it is a quite long lead-screw

If you want to do something like this, it would IMO absolutely need some bearing that takes the force of the belt tension. Actually I would not be surprised if parts of your problem comes from this.
And an Oldham coupling would also not work in such a setup.

13

Hi Sineos-san
The belt is useless for in printing.
It’s not tight and not too loose.
It has the effect of preventing slippage when not printing or push the bed unintentiaonally.
(I’m not good at English.May be not correct)

by the way
I purchesed flex coupling.
I can’t wait to arrive. :grinning:

14

It does not really matter why the belt is there. The way it is attached will lead to problems and it will not get better with a new coupling. If you want to solve these print artifacts then you will need to:

  • Improve the mechanical alignment of your rods and make sure that the bedrun freely
  • Improve the mechanical alignment of your leadscrew and make sure there are no forces skewing this lead-screw

For the second point you have two options:

grafik
grafik1378Ă—1374 117 KB

This image shows a similar setup but done right: The leadscrew has a bearing on the bottom and one above the pulley. The forces of the belt cannot skew the leadscrew.
This is the only really correct setup

Second option:

grafik
grafik413Ă—552 47.1 KB

Move the pulley down to the axis of the stepper. Then the stepper axis will take the load, which is not ideal but much better than having the load above the flexible coupling. Of course you will need new pulleys with the right dimension.

15

I made sense your opnion.
First pic looks very strong !
It’s difficult to move the pulley down.
I’ll check what happens when the pulley is removed.

This is the pulleys I bought.
Tronxy Z-axis timing belt adjuster for X5SA 3D Printer with Z axis synchronous wheel + belt (tronxyonline.com)

16

You would just need to get pulleys with the same amount of teeth and a 5mm bore (for 6mm belt). Something like:

Quite easy to find actually.

Edit: Just to remind. Check the alignment of the rods and of the leadscrew. This will have a really huge impact

17

Hi Sineos-san ,
Finally the parts have arrived.
image
but It’s wrong teeth size.I need 24T.
now I’m contacting to the seller for exhange.

Should I replace the anti-backlash nut ?

18

You can try, boyon-san. In doubt I would wait until you can modify both, saves you some work.

When you do this modification, I strongly encourage you to really check the alignment of both the lead-screws and the linear bushings. Especially the bushings have a high chance of binding if not aligned perfectly and this will almost surely heavily affect the print quality.
You can only check the quality of the bushing alignment when the lead-screws are removed.

19

Sineos-san
I replaced the parts today. (including anti-backlash nut )
has been good shape!
Thank you for the many tips.

image
image519Ă—617 50.5 KB

20

Looking great boyon-san :+1:
Awesome that you worked it out.

1 Like