Lead Screw usage for X and Y axes

Hello Everyone!

I'm designing an all-lead screw printer (see here), and that's what got me started on this topic. The RepRap wiki page doesn't really adress X/Y lead screw usage, so I thought it should be filled in. Here are some of the questions I personally would like answered on such a page, but there's plenty of other interesting engineering questions related to quick lead screw moment.

- How to make it cheap?
- How to make it wear resistant?
- How to lubricate?
- Could we rely on rotating lead screws for straightness?

Basically, I'm asking after any knowledge about quick movements using stepper motors and lead screws...

Summary of my research so far:
It works, we should do controlled tests and optimize it.

Successful builds
BetterHalf's video,
SamP20's video of the ScrewRap,
Kyo Shinohara's Andamanite video,
Kyo Shinohara's Fina First Print video
Y-axis of the Son of a MegaMax instructable

Previous related discussion
Thread ending up talking about string drive,
samp20's thread,
nicholas.seward's Lisa Simpson thread,
Short thread recommending to use two nuts,
Thread containing warnings that lead screws might be slow and expensive

There's also this Google Groups thread where user Buzz does a build, has some insights:

- Take steps to avoid entire lead screw moving in end-to-end direction
- 5 starts is adequate

... and this Openbuilds thread about Kyo's Adamantine build.

Related Wiki Pages
ScrewRap wiki page,
Leadscrew wiki page (this is the one I want to improve)

If you don't feel like an expert on lead screws, please share your thoughts anyways!
Cheers!

Edited 5 time(s). Last edit at 01/09/2016 04:43PM by tobben.

---

torbjornludvigsen.com

I would add two further questions:

- How to make it fast enough? Possible answer: use leadscrews with course pitch, motors with low inertia (short ones), and electronics and firmware capable of generating high step rates and using 24V or greater power so as to achieve high stepper motor speeds.

- How to avoid play in the leadscrew end bearings and between the leadscrew and the "nut"? I don't know the answer to that one.

Edited 1 time(s). Last edit at 01/09/2016 05:21PM 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].

My printer has screw drive in the Y and Z axes. Just like the printers in the videos you linked, it's pretty noisy. Ball screw drives are great at transferring vibrations from the motor shaft to the bearing blocks, printer frame, nut, and whatever the nut is connected to. The moving mass of the screw adds to the torque required and you will need higher torque motors than similar belt driven systems.

The main advantages of the screw drive is a complete absence of print defects like ringing and rippling in the surface (assuming you have no backlash).

I used the motor sizing tools at the Oriental Motor website to calculate the torque required to drive the Y axis in my printer, then bought a 425 oz-in NEMA-23 stepper. It seems to perform exactly as calculated.

Lead screws are like everything else- the cheap stuff isn't as good as the expensive stuff. My Z axis uses two 1/2" lead, 4 start, teflon coated acme screws that turn in plastic (delrin?) nuts. The Y axis uses a precision ground 10mm lead ball screw with a backlash-free nut. X and Y axis bearings are fully supported linear guides.

You can see my printer design by following the link in my sig, below.

Edited 1 time(s). Last edit at 01/09/2016 05:56PM by the_digital_dentist.

---

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

Quote

use ... motors with low inertia (short ones)

Wow, that's a solution I hadn't thought of. So a short NEMA23 is better than a longer NEMA17 assuming all other specs equal. How great is this effect? Just making up an example: How much acceleration would a 42mm NEMA23 gain compared to a 60mm NEMA17 just from being short (depends on lead screw of course, say 10 mm per rev or something)?

Quote

I used the motor sizing tools at the Oriental Motor website to calculate the torque required to drive the Y axis in my printer

Thanks, that's an excellent resource! I'll add the SoM y-axis instructions to the first post, its really similar to what I'm after myself. Thanks for sharing.

---

torbjornludvigsen.com

I've been working on a printer like this for a while.

Here's a render of what i have so far
Render

and here's the bill of materials:
BoM

- How to make it wear resistant?
I used bronze leadscrew nuts on precision ground leadscrews from mcmaster carr. bronze is very wear resistant, plus these can be machined into any shape that you want.

- How to lubricate?
powdered graphite might work, or a light machine oil.

- Could we rely on rotating lead screws for straightness?
no, not directly. you would want to rely on your linear rails for straightness. because of this, they need to be more rigid than your leadscrew. your leadscrew is for transmitting force, the rails are to make sure it goes in the right direction.

- prevent play in the leadscrew:
I designed it so that a spring would keep tension between a flanged bearing and a collar. but i think i might need to rethink that one. it doesn't seem to work very well in practice.


I can keep you updated if you'd like.
Hope it helps!

~alby

Edited 2 time(s). Last edit at 01/09/2016 05:13PM by albatroopa.

Quote
tobben

Quote

use ... motors with low inertia (short ones)

Wow, that's a solution I hadn't thought of. So a short NEMA23 is better than a longer NEMA17 assuming all other specs equal. How great is this effect? Just making up an example: How much acceleration would a 42mm NEMA23 gain compared to a 60mm NEMA17 just from being short (depends on lead screw of course, say 10 mm per rev or something)?

Both inertia and torque increase with stepper motor length, in about the same proportions. So if most of the inertia is due to the stepper motors and not the leadscrews/bed/carriage, then these effects roughly cancel out and both long and short motors will produce the same acceleration if run at their maximum current (or at the same proportion of their maximum current). However, the longer motor will have higher inductance and higher back emf due to rotation, so it will need a higher drive voltage for the same speed. Using leadscrews, you are likely to need less torque but more speed than for a belt-driven X or Y axis, so short motors will be better unless the carriage or bed is very heavy.

Unless you are building a very large printer, don't use Nema 23 motors - they need more drive voltage than Nema 17s and have higher inertia, and as I have already said, motor torque is unlikely to be a problem (because of the gearing provided by the leadscrews). You can do the maths to work out what torque you need.

Edited 3 time(s). Last edit at 01/09/2016 05:31PM 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].

Thanks alby, you bet it does!
So much concrete insight. I'll bake it into the wiki tomorrow.

Quote

I can keep you updated if you'd like.

Yes, that would be really interesting

---

torbjornludvigsen.com

Hm... You really know the trade offs between Nema 17 and Nema 23, dc42. I think I understand most of it. Short Nema 17 is what we want then.

Quote

Using leadscrews, you are likely to need less torque but more speed than for a belt-driven X or Y axis, ...

Nice! With no experience of lead screws from before, I was expecting the opposite.

---

torbjornludvigsen.com

NEMA-17 motors top out at about 100 oz-in of torque. If you need more than 100 oz-in, you need to start looking at bigger motors. You won't know how much torque you need until you know all the moving masses (rotational and linear) and the losses such as stiction and friction. Any money spent on a "short NEMA-17 motor" at this early stage is likely to be wasted- short NEMA-17 motors don't produce much torque.

---

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

Quote
tobben
Hm... You really know the trade offs between Nema 17 and Nema 23, dc42. I think I understand most of it. Short Nema 17 is what we want then.

Quote

Using leadscrews, you are likely to need less torque but more speed than for a belt-driven X or Y axis, ...

Nice! With no experience of lead screws from before, I was expecting the opposite.

It all.depends on the microsteps/mm that you get with your leadscrews, which depends on their pitch and number of starts. Using x16 microstepping, belt driven printers usually give 80 to 100 microsteps/mm. Short Nema 23 17 motors are adequate for these printers if Bowden extruders are used to keep the carriage mass low. In saying that you need less torque but more speed than a belt driven printer, I have assumed that your chosen leadscrews will give you greater steps/mm and that other things such as carriage mass wil be the same as for a belt driven printer. Do the sums before you buy any motors.

Edited 1 time(s). Last edit at 01/10/2016 09:50AM 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
the_digital_dentist
You won't know how much torque you need until you know all the moving masses (rotational and linear) and the losses such as stiction and friction.

The notion of stiction was new to me. It should be quite straightforward to test stiction for some different leadscrew/nut/lubricant combinations like this:

• Wind a cord around the center of the flanged nut
• Mount the lead screw horizontally
• Add weight to (a little bag or platform) the end of the cord until the nut starts rotating
• Weigh the little bag or platform

Following alby's advice, I'll test with a steel lead screw (1018 Carbon Steel?) and a preloaded bronze flanged nut with powdered graphite lubricant first. I don't know if stiction really is a major hurdle to overcome, but my interpretation of SamP20's excellent demonstration (video here) is that stiction issues might be significant. (Since his threaded rod is placed off-center, a "skew-forced drawer effect" in the linear bearings might be more significant than lead screw stiction in the demonstrated case, I don't know.)

Quote
the_digital_dentist
Any money spent on a "short NEMA-17 motor" at this early stage is likely to be wasted- short NEMA-17 motors don't produce much torque.

Quote
dc42
Do the sums before you buy any motors.

Thanks for caring for my budget guys. I think you're right that lead screws generally will have to rotate faster than belt pulleys since extremely aggressive lead screws aren't widely available. Retaining decent dynamic torque at high rotational speeds (~300 rpm) is a functional requirement then. Following advice on Stepper torque wiki page we want the motor to have low inductance and high rated current. (Also want good match with driver chip and driving voltage, as usual with steppers). My project has the budget to test out a couple of different motors, so I'll explore which cheap alternative works better (and why and how). My design project itself has the money it needs, but the design it is producing needs be as robust as a $500 (in parts) printer can be.

Does any of you know why extremely aggresive lead screws (> 14 mm lead) aren't more widely available?

I've found answers in other threads suggesting that there are lots of manufacturers of such lead screws around, but very few has online stores. I mean, belts and pulleys are easy to get online, leads screws have some advantages like less vibrations/ringing without being _that_ expensive. Are there disadvantages to aggressive threading that I'm missing?

---

torbjornludvigsen.com

Quote
albatroopa
Here's a render of what i have so far
Render

and here's the bill of materials:
BoM

Was just looking through your BOM and noticed that the planned lead screws use imperial units. Are you sure that won't be an issue?

I changed my imperial-unit Z-threaded rods on my first Printerbot+ because the millimeter/inch roundoff errors produced visible artifacts (every fifth layer was noticably thinner), that's why I'm asking.

Beautiful render by the way. What's inside the V-shaped end?

---

torbjornludvigsen.com

Just asking, why lead screws (metric of course, "inch" is a big "NO WAY" for me ) for 3D printers when cheap timing belt seems to provide enough accuracy and are so simple to use, low inertia, no noise, no play leading to backlash unless high quality ball lead screws, no lubrication and thus no dust lubricant mess etc...
OK, the 3D printer I selected uses two lead screws for the Z axis but here they have to hold the X axis hardware.

I guess I'm not in a position to answer this yet, but the successful builds linked in the first post shows some interesting properties:

• Very little ringing/underdamped oscillation/shadow wave patterns near corners.
• Clean and simple mechaincs
• Simple assembly?
• Durable?
• Low maintenance?
• Possibility of a lower unique part count?

The possibility of a straightforward modular design based on linear actuators that work the same on all axes is what motivates me to explore this.

---

torbjornludvigsen.com

For a given volume of wire and magnets: arranging them in a long thin form factor reduces the rotational inertia because it is proportional to the radius squared times the mass. That is why long thin motors exist.

For a given volume of wire, whether a motor is high voltage / high inductance or high current / low inductance depends on how thick the wire is. I don't think form factor affects inductance, it is proportional to turns squared. The number of turns depends on the wire gauge.

To avoid the rotational inertia and wobble / whip of a long screw you can fix it and rotate the nut instead.

For quick movement with longevity you need multi-start high lead ball screws and that is not compatible with low cost. They are at least an order of magnitude more than belts and pulleys and not as fast.

You tend to need anti-resonant drives to accelerate through the mid band resonance, which are again more expensive. On the plus side you can mill as well as 3D print if the rest of the machine is stiff enough.

---

[www.hydraraptor.blogspot.com]

I went with leadscrews because they're more accurate and i figured id be able to make a more rigid system which means i could run it faster. backlash is drastically reduced as well. i'm pretty sure i had a reason for going imperial on the leadscrews, i think it was simply availability, but im not 100% sure. it was a while ago.

here's a render of what's inside that v-shaped motor bracket:
Render2

from the motor towards the centre of the frame, the parts are arranged in ths order:

-motor
-motor bracket
-shaft coupler (i 3d printed a bunch of these)
-leadscrew
-shaft coupler
-die spring
-thrust bearing
-flanged bearing
-y axis face plate

edit: prettier picture:
render 3


also, i should mention that i used 4 star leadscrew for x and y axis and single start for z axis

Edited 4 time(s). Last edit at 01/10/2016 01:11PM by albatroopa.

[www.thingiverse.com] This usable? Maybe with a "threadless ball drive"

Edited 1 time(s). Last edit at 01/10/2016 01:08PM by Tunames.

I have 1/2" lead acme screws in the Z axis of my printer. See if you can spot any problems:



As I only sort of understand it, Imperial threads are only a problem when the pitch is small.

---

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

Quote
the_digital_dentist
I have 1/2" lead acme screws in the Z axis of my printer. See if you can spot any problems:


As I only sort of understand it, Imperial threads are only a problem when the pitch is small.

i'm assuming you mean 1/2 in acme leadscrews. a 1/2 in lead means that in 1 revolution, the nut would travel 1/2 in (pitch*number of starts=lead), whereas a 1/2 leadscrew alludes to the diameter. approximately how many hours has that leadscrew run for? it looks absolutely mangled...

The problem with the "inch" lead screws may be due to rounding done by the software that at some point has to compensate for accumulated rounding errors not unlike calendar leap year. The absolute rule is to have all in metric, lead-screws, encoders, internal representation in the programs etc... and for people still using it to perform the "inch" unit conversion only once, at the human input/output interface.
I wouldn't buy a non ISO compliant machine.

Pitch is the spacing between the threads. Lead is the distance moved in a single revolution. I have 1/2" lead (not diameter) screws- 1/8" pitch, 4 starts. I'm not sure what the diameter of the screws is.

The picture is a micro photo of the corner of a printed cube, 0.25 mm layers, not an acme screw. I posted it to illustrate the lack of z artifacts even though the screws that lift the X axis use Imperial threads.

---

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

Ahh, lol that puts it into a whole new perspective. I see it now.

Quote
MKSA
I wouldn't buy a non ISO compliant machine.

What is an ISO compliant machine? ISO 8217? ISO 26000? Which ISO spec pertains to 3D printers?

---

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

Quote
the_digital_dentist

Quote
MKSA
I wouldn't buy a non ISO compliant machine.

What is an ISO compliant machine? ISO 8217? ISO 26000? Which ISO spec pertains to 3D printers?

By ISO I mean all parts compliant to ISO standards: fasteners, bearings, pulleys, gear, lead screws, belt, rods, tolerance classes etc... I am OK if their production is not ISO 9001, neither their documentation usually written in chinglish
SO no hassle to find parts to replace, make them, adjust them ... Plenty of parts can be reused from old printers, PCs, bike, cars etc... Example an old printer contain valuable small screws, bearings, rods ...

I have never seen a 3D printer marketed with any sort of reference to any ISO standard. When you go shopping for a printer, how will you know that each part in a machine is compliant with whichever ISO standards are relevant? How will you know which standards are relevant for each part?

Are you talking about metric vs imperial dimensioned parts?

---

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

How bad is the lead screw longevity issue?

I still wonder if we could get good enough (2500 hrs?) lead screw lives by minimizing mass and acceleration and doing some research on lubrication, number of starts, lead pitch, diameter, nut length, nut/screw material combinations. Equally well would be to find a source of cheap ball screws+nuts or a feasible way to produce them ourselves.

The possibility to CNC would make an interesting RepRap machine.

I tried hard to spot periodic errors caused by 1/2 in acme leadscrews. Didn't find any.

---

torbjornludvigsen.com

The dip you're seeing in the center of the part is caused by pincushion distortion in the camera lens- I used my Droid Turbo phone to take the picture through a microscope and did not correct the image for the camera's distortion. You can see the distortion more clearly in this video, made with the same Droid Turbo/microscope combo: [vimeo.com]

I should have taken that photo with the part pressed against the edge of a steel ruler.

How did you derive that graph?

Edited 1 time(s). Last edit at 01/13/2016 11:02AM by the_digital_dentist.

---

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

I expected it to be optics, but it could have been just a curved figure. Cool microscope solution.

I just zoomed in the picture and drew the graph by hand in Inkscape, nothing rigorous.

---

torbjornludvigsen.com

Consider High speed leadscrew or multi start leadscrew

[buildyourcnc.com]

The Aluminatus A1 printer uses a PCBLinear guide rail with a High speed leadscrew.

[www.pbclinear.com]

Probably too expensive for your application

Thanks for the links. The pcblinear one led me to read about the Aluminatus.

Do you know anyone who have used one of these printers for a while? It would give us more exact knowledge about the wear problems.

Emailed Vermont Rapid Prototyping about it.

Quote
cozmicray
Probably too expensive for your application

Do you know how expensive? A bit cheaper than ball screw would be nice.

---

torbjornludvigsen.com