Ball screw vs lead screw

Hi
I am working on a 3d printer design and I was wondering if lead screw or ball screw would be better for my project?
I am wondering which is more accurate? Are there any other advantages to one over the other?
I was looking to use it for the z axis of a core xy design.
What would be a good pitch to use?
Thanks
Tanner

It's mostly a cost thing. Lead screws are cheaper, ball screws are (arguably better) more expensive.

As for some more specifics as to what the performance differences are between them, lead screws are less efficient and have more friction. This means a couple of things.
-You need more motor torque to move them (not a big issue, NEMA 17 can handle that)
-There is more wear (you'll need to replace them/the nut more often, but not a big issue again, you won't need to replace them often to begin with)
-They can't be back driven, this means that if you're motor dies or stops, your z axis wont move or drop under it's own weight.

They also tend to have more backlash, but this is not really an issue for the z axis as it doesn't change direction much and has a pretension downwards due to gravity. When it does change direction it is rather slow and small movements and gravity helps get rid of any backlash. If you're paranoid you can make/buy an anti backlash nut, basically two nuts separated by a spring but that can increase the friction, I'd say not worth it.

Ball screws are more efficient and have less friction. They need less torque from the motor to drive them which means you can run your motors are less current (small but I guess noteworthy point), and there is less wear usually because of the lowered friction. Due to their higher efficiency they can be back driven, but I wouldn't really expect your z axis to drop if you cut power to the motor.

I think most people would recommend ball screws if you can afford it, but it is a bit of a price hike, and there may not be any noticeable print quality improvement. Most high end machines use lead screws and print fantastic.

...what he said
+ ballscrews can be driven faster, so homing after a tall print doesn't take ages. Also z-hop during print is faster which reduces oozing time.
the pitch of ballscrews usually is higher, so it is good practice to use bigger pulleys to have a gear ratio between 2:1 or 3:1.

Quote
Trakyan

They also tend to have more backlash, but this is not really an issue for the z axis as it doesn't change direction much and has a pretension downwards due to gravity. When it does change direction it is rather slow and small movements and gravity helps get rid of any backlash. If you're paranoid you can make/buy an anti backlash nut, basically two nuts separated by a spring but that can increase the friction, I'd say not worth it.

I disagree with this, an unbalanced load will cause backlash on one side and not the other. For example, a typical prusa style X axis with the motor on one side will have a 'see saw' like effect across the axis. Anti backlash nuts solve the issue.

Edited 1 time(s). Last edit at 06/26/2017 03:37AM by Origamib.

Fair point, however he's building a core XY and the Z axis here shouldn't suffer from the see saw effect. Good to know, I'd never considered the weight of the hot end and extruder shifting having a 'see saw' effect on my I3 style set up. That said, for this 'see saw', doesn't the extruder+hot end always have a greater torque than the x axis drive motor regardless of position? At worst, both have an equal distance from the lead screw as a pivot with the carriage at x=0 (realistically I think the carriage is a bit further away), and at any extension the carriage definitely has a larger lever arm, and higher mass. Wouldn't this keep the 'see saw' consistently tipped to one side? This slight tilt would be accounted for in bed leveling. Correct me if I'm misunderstanding what you mean by a see saw (X axis rails are the armature, lead screw between x carriage and x motor as the pivot and the carriage and x motor as the respective masses on either side of the pivot.

Question on backlash. With autoleveling the lead screw changes directions constantly. Doesn't this cause backlash?

It doesn't cause the backlash, but it may be cumulative if the backlash isn't equal in both directions and the same along the full length of the screw. Ballscrews as well as acme and trapezoidal screws have backlash with ball screws generally having the least amount. Inexpensive Chinese ballscrews can have as much or more backlash as well made acme or trapezoidal screws.

p.s. Ballscrews are also leadscrews. Leadscrew describes the function not the tooth form of the screw.

The OP is talking about using screw(s) in the Z axis. Gravity keeps the nut's threads meshed with the screw's threads at all times, no matter which direction, up or down, the bed moves. There is no backlash unless the mechanism is so poorly constructed that it binds. If that happens, you should fix the mechanism so it doesn't bind, not add an antibacklash nut.

If you use screws for horizontal motion, gravity doesn't keep the threads in contact so you have to be concerned about backlash and use antibacklash nuts.

---

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

Gravity only stops backlash for accelerations under 1 G, above that you can get backlash, it's why I mentioned that movements are small and slow. It's not a super noticeable thing but it can allow faster z hops which can shorten print time a little. That being said, the z axis doesn't move much so this wont be a big change, it may not even be noticeable or there may be other limiting factors. Z hopping seems the biggest issue regarding backlash I can think of.

As for backlash during auto level, I guess it can happen but you're only measuring the bed level as your bed is coming up, during the change in direction (after your probe has triggered), you are neither printing nor trying to measure anything so the backlash does not affect anything. Again, if you have accelerations over 1G, you can have the bed end up in 'free fall' or forced against the upper surface of the thread rather than resting on the bottom surface as it tries to catch up to the screw itself. This can cause backlash which will affect bed leveling, but I don't see why you'd need or even want such high accelerations when bed probing, most sensors want a relatively slow approach to the bed to give them time to trigger at their intended distance.

How does acceleration in either direction ever exceed 1G? And even if it did, you don't extrude plastic while the Z axis is moving, so how will acceleration >1G affect the print?

---

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

I don't know what someone may choose to set their acceleration settings to. And like I said it may effect bed leveling if you do it stupid fast. On a moving bed it can cause vibrations in the print due to sudden, juddery movements of the bed. But you're right, acceleration over 1G isn't very likely.

My design for Bateson uses 12 identical leadscrews in groups of 4 for X, Y, and Z, in a cube, and suspends the cube on the diagonal to eliminate backlash as noted for accelerations below 1g.

This has the peculiar effect of changing the base of replication from 2 to 12. That is when the machine replicates x generations there are 12x axes, not 2x machines. Not quite the desired 12^x basis for replication but listen on.

The cube arrays indefinitely and beyond 5 x 5 x 5 there are *fewer* "surface" cube joints requiring complex octahedral machining than mere cubes joining the cubes that make up the (125x and larger) array. It's built to scale. The exterior cubes receive an octet truss bracing that rigidifies the array.

You'd add robot arms to pass work in and out of the factory array to every cube. Still one cube is of value and one axis is of value.

The central cubes in an array would benefit to an unstudied degree from the bracing around them; milling should be possible in the central cubes; forming may be possible.

Just rambling on. Start with simple components, multiply them, let the synergy happen.

Don't over think it.

Doug

Bateson's leadscrews can be threaded rod, precision threaded rod, Acme threaded rod, precision Acme threaded rod, ball screw rod, precision ball screw rod, or unthreaded precision rod with spiral ball nuts; it doesn't matter. Those are the classes of materiel offered by my suppliers. Bateson isn't specified by what is built but by how it's built.
One assembles a single X axis and uses it to run of 12 X axes; those are assembled into a cube with secondary rails between pairs of electrically synchronized axes. They're electrically synchronized because the cube can later be arrayed with other cubes to make an array factory and cog belts would prevent that happening. The secondary rails define the position of tertiary rails which cross in a final block. From 12 primary axes there are 6 secondary rails, 3 tertiary rails, and 1 block. The shape of the build volume depends on what is attached to the block and to a mount point opposite the hang point.

I should open a page for Bateson instead of threadjacking the OP on ballscrews.

Doug

What on earth (?) are you talking about?

---

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

Quote
the_digital_dentist
What on earth (?) are you talking about?

I thought it was only me, who understood narda. But maybe that was the intention? Making us curious and search for Bateson.

If it works as well as the 4 axis self replicating mill, I'd be interested in seeing it.