Aside from the eye-watering cost of high helix lead screws (25mm pitch or higher) and corresponding anti-backlash nuts, is there any reason not go with them? I’ve not seen many leadscrew builds. IIRC I have only seen one, but that used standard ball screws, roughly 10mm pitch.
Do you think the lower torque for the high-pitch versions would be a factor? Does precision suffer as a result of the higher pitch?
I was thinking of using igus high helix lead screws with their plastic nuts. Very quiet (I have a sample).
Not sure about the torque, though. I did consider it, and figured with going with two motors for the y-axis driving a cross beam with one motor, a nema 17 with something in the regions of 60Ncm driven by a smoothieboard.
The discussion brings to mind another question. What advantage would a high pitch screw provide to the printer? It seems that the disadvantages are foremost.
My thinking was that it would reduce or eliminate ringing, and generally result in higher precision than belt-driven solutions.
With regards to disadvantages, I’m trying to discover what they are. Can you shed some light on that?
I thought Mark Rehorst’s comments were directed at the disadvantage side of things. A more powerful motor and driver board and higher power supply seems to add cost. More vibration would appear to be a disadvantage too.
Why not a low-pitch screw drive?
I’m mostly ignorant of one aspect here. What part of a design contributes to ringing? I interpret the term ringing to mean a deviation from desired form caused by a mechanical characteristic of the construction. If ringing is caused by belt drives, what is the characteristic of those drives that creates it?
By ringing, I refer to this: http://support.3dverkstan.se/article/23-a-visual-ultimaker-troubleshooting-guide#ringing
It’s a speed-related issue, and as far as I can tell this is due ultimately to the belts having a very small amount of stretch. Do please correct me if I am mistaken.
That’s a fairly comprehensive guide and very interesting reading! The portion about ringing referred to acceleration and by extrapolation, one could assign belt stretch to that characteristic. I’ve not done any direct research about belt stretch, but have found that the entire assembly has to be assessed when tensioning belts. I had some tower supports begin to tip inward when tightening a corexy design, limiting the tension I could apply. It’s assembled now and I have no spare belts to test, but one would expect that eventually the stretch is “pulled out” or the structure fails.
Is a secondary benefit of moving to a high pitch screw expected to be a greater tolerance for high acceleration?
Why not a low pitch screw drive? I can see it would require faster motor rpm for the same speed, but lower current requirements from the motor and driver board.
@Mark_Rehorst great stuff. From the build I recall the chap who made claims he no longer suffers from ringing, which is what sent me thinking about using lead screws.
Now I wonder if there is a way of testing whether it is the motors or the belts responsible for ringing, or which contributes more, using a printer rather than building something specific?
@Mark_Rehorst @Fred_U Ringing is caused by the belts, not by the motors. If u “strech out” a belt by tensioning it, the belt gets damaged and you can throw it away. With my build of the DICE (micro-coreXY) I can easily verify that ringing is directly dependend to the length of the belt used. Otherwise, there would always be the same size of the ringing artifact, cause we us all 200 steps NEMA17 all the time 
Ringing is due to many factors, including the mass of the moving parts and the flex and play in the assembly. Switching to a screw might reduce some minor flex but it also adds a large rotational mass. You will have to lower the acceleration and jerk values to account for the mass and just lowering those would address the issue when using belts.
@Wylie_Hilliard For that I would imagine that the more powerful motors would cope with the greater rotational mass.
Where the ringing comes from depends on the printer design.
From the math I’ve done, rod flex and stepper torque springiness contribute MOST of the ringing in MOST printers. (The rod flex is in whichever Cartesian stages are dynamically side-loaded. Z stage in i3 type printers, X and Y in XY gantry printers.) Belts will contribute proportionately much more in long-belt, low-mass, rigid-rail printers like lasercut-sheet CoreXY (such as @Rene_Jurack 's DICE).
A screw will (essentially) eliminate belt stretch, and by virtue of the much smaller travel per motor step, decrease the visible impact of stepper rotor oscillation upon direction changes. It won’t do anything for rod flex. But I don’t think anyone is dumb enough to pair a giant screw with flimsy rods…
The obvious downside to a screw is motor RPM required. If you just want to print slow, and use a “fast” ballscrew, then you won’t have any issues. But it’s EXTREMELY difficult to configure a drivetrain to hit high linear speeds with a screw. You need very, very high drive voltages to overcome the motor back-emf and maintain current and torque.
A less obvious issue with screws is that they tend to push wobble into the linear stage. Screws are not as straight as linear rods. But they are similarly stiff – an 8mm screw is nearly as rigid as an 8mm rod. So the screw’s poor straightness will push on the rods unless you’re quite careful with the linear stage design or have very, very stiff rails.
Fast screws apply more torque to the linear stage through the nuts and are less efficient at converting motor torque to linear force. Usually not a big deal though unless you get stupid with the lead.
@Dani_Epstein Larger motor has more mass, requiring even more power.
The most likely reason that some have seen an improvement in ringing is the fact that lead screws require a lower speed and acceleration.
@Wylie_Hilliard Of course, but the mass/power ratio and hence the inertia and the power to overcome it changes with the motor dimensions, from what I understand. That is to say, the long bodied Nema 17 stepper motors will have a better mass to torque ratio than the short bodied ones. So of course it means bolting bigger motors on, and more mass, but the power increase does not appear to be linear. Then again, I could be wrong! It’s not my field.
@Fred_U more vibration in there motor and frame but compared to belts surely less vibration of the hot end that now has a stiff connection to the frame.
@Ryan_Carlyle Not really sure how I missed your comment, and its full of useful information!
From what I understood from you, there are four components that contribute to ringing: the stiffness of the build, belt stretch, guide (or lead screw) induced wobble and torque springiness from the motors.
I was debating between coreXY or lead/ball screws, but the belt length of the coreXY worries me a bit, since I want to make something with a bed in the regions of 400mm x 300mm.
So my thoughts were to deal with the various issues as follows. The build stiffness: prestressed concrete and some spare parts from an old battleship I have lying in the garden. Failing that, some beefy aluminium extrusions and some creative thinking.
The belt stretch issue appears to be obvious, use ball screws or high-helix lead screws with anti-backlash nuts.
Reduce wobble insofar as possible by using something like MGN12 or MGN15 linear guides. If possible, create a floating mount for the ball/lead screw nuts so they only “push” forwards and backwards, i.e. along their main axis, and any sideways movements will be absorbed by the floating mounting, hence reducing the effect of wobble.
Now, I had not considered the motors factoring at all, so this is new to me. Of course on really short movements such as a narrow infill pass, where is almost no room for acceleration and deceleration, perhaps the motor springiness is going to be a much bigger issue as well. This would affect something with a textured surface
Also, I have no idea how fast a stepper motor can really go. Although I can get my UM2 to zip the print head pretty quickly, I still don’t know how fast those steppers are actually going. With 12mm ball screws the lead is around 4mm, and getting a 10mm lead will need something in the regions of 20mm diameter. So speed of the motor is going to be pretty critical in order to get 50mm/sec printing or higher.
So I figured maybe go for igus 10mm diameter 25mm pitch lead screws with plastic anti-backlash nuts, although I confess I do not know what precision one could expect from them. I have a sample of a regular plastic igus nut on one of these and its difficult manually gauge if there is any backlash at all.
All of this still leaves me wondering…
@Dani_Epstein sounds like you’re building a CNC router, not a 3D printer! That’s all overkill for the kinds of forces we’re dealing with. Check out the Fusion 3 F306 CoreXY design. That’s spectra instead of belts, but in any case it produces exceptional print quality with 300mm cube build volume and fairly simple construction. (Lots of clever design details but nothing you couldn’t mimic.) If you’re worried about belt stiffness, switching to 9mm wide 2mm pitch GT3 belts will greatly reduce belt stretch.
Really great thing about CoreXY for ringing is that the contribution from belts and motors is rotated 45 degrees out of alignment from the contribution from rod flex, so they don’t add constructively as much as a simple Cartesian printer would. All the drivetrain forces are split between multiple load paths.
Thanks Ryan, I know it sounds more like a router than a printer. I have some pretty large objects that I regularly print, and need them printed reasonably accurately and of course as quick as possible. So it looks like I will have to give up on the whole lead screw thing until I can afford many monies for servos. CoreXY it’s going to have to be.
0.9 degree steppers will decrease rotor ringing magnitude. (Just make sure you run 24v PSU so you’re not RPM-limited.) Wider GT3 belts will reduce belt stretch. (You can still do 2mm pitch and use GT2 pulleys with GT3 belt of you want.) Rails mounted on extrusions will maximize your stiffness to weight ratio.
300mm is kind of the inflection point where design and parts selection starts to really matter.