Leadscrew-ify Mendel/RepRap

I was considering re-designing the drive mechanics of a Mendel/Huxley to use lead screws instead of belts. Well aware of some of the changes that follow with that and all, but has anyone thought about this too? Any work been done/in progress?

Or is there a reason no one has thought about it? Costs might go up, but you get more precision. One would have to rework a fair amount of the values to get the machine printing within proper scale. Still, would be nice to have options, right?

---

- Andy
digifabindustries.blogspot.com

Hi Andy

The main reason people are not doing this is because of the limit on printing speed. Leadscrews are very slow.

I would imagine the only change you would need to apply to the firmware would be the steps per mm parameter for each axis.

Further, the gcode generating program would need to ensure the x and y axis acceleration and max speeds are reduced from that for the belts.

Cheers

... leadscrews need higher motor-torques and better (and so more expensive) drivers - i have NEMA23-motors with 2Nm torque and drivers with max clocking speeds of 200kHz in 1/10-microstepping with 10000 steps per rev. or 10MHz in 1/256 microstepping with 51200 steps per rev.

With my most common leadscrews with 5mm translation per rev. i'll receive 100mm/s (or 6m/min) with the 1/10-drivers or 976mm/s (58m/min!) with the 1/256-drivers, what's not so slow ... but i doubt i can use this speeds with my mechanical setup, though my mechanics are really rigid

So the comparison/decision between toothbelt and leadscrew is more depending on your intentions and monetary background ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Also if you are going to go the leadscrew method, use an acme thread (with anti backlash nuts) or ball screw. Typical hardware store threaded rod is not meant for use as a lead screw, its meant to hold things together and not constant use.

(can ya tell I'm a CNC service engineer?)

Leadscrews are not slow, if you pick the right ones. Multi-threaded screws get higher speed than single-thread screws and can go as fast as belts.

Leadscrews will let you achieve a higher-stiffness drive, but will typically have more friction and inertia than a belt. Whether or not precision will depend on how you design the system, and what leadscrew/nut you use...

High speed leadscrews don't turn the screw (it will start to whip) they turn the nut instead.

... atached is an image of my actually best linear stages - they're running with up to 100mm/s with rotating screw.

I haven't tried with the 976mm/s-drivers, but i think they should run until roughly 500mm/s fine too, if the screw's aren't to long and if they are perfect straight.

With longer screws you'll receive problems with high turning speeds, but this depends on the diameter and stiffnes of the screw and feeding/fixation methode ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

How would you constrain the nut in a Leadscrew mendel to slide perfectly (without rotating) ?
It would always have some sort of slop wouldn't it?
Would you maybe use a spring or bearings to hold it to a guide rail or aluminium extrusion?

Davebass, ballscrews are near zero backlash and for acme threads anti backlash nuts are available.

Now with the Z axis with standard threaded rod (not lead screw) the leadscrew without anti baclash is still ahead of the curve.
And a leadscrew nut isn't like a hex nut, its a threaded 'piece' which can have mounting holes/brackets.

As for keeping the 'screw' slop free, thats what thrust bearings are for.

(One model of the lasers I service, the 'screw' is held steady at a rather high tension and the 'nut' spins. It has a positioning speed of 120 m/min, yes meters a minute)

Dear all,

Hmm..... lead screw anti-backlash nuts are readily available either with a hex form or flanged with screw mounting holes (Igus, Ondrives etc), the load required to rotate a modern lead screw and anti backlash nut is very low (although most certainly more than the belt arrangement), but yes there trade off will be for accuracy over speed without a definite upgrade to the motor:

BUT BE WARE NEMA-23 is neither more powerful nor accurate nor of a higher resolution than NEMA-17 ... the NEMA bit is a mounting plate specification in this case and nothing to do with the step size, torque or accuracy ..... in selecting a motor make sure you select the required torque, speed, resolution and accuracy, you can get all these improvements in a NEMA 17 motor !

I'm personally going to be doing this too (in a couple of months) ....

(my background is mechanical and composite design engineer - specifically naval architecture just for reference) .....

Tim

I am currently in the process of building a mandelmax with 4 start acme leadscrew on the x and y axis and single start on the z, as well as lm8uu linear bearings on all axis. I'll publish it once it's done

chishgre Wrote:
-------------------------------------------------------
> I am currently in the process of building a
> mandelmax with 4 start acme leadscrew on the x and
> y axis and single start on the z, as well as lm8uu
> linear bearings on all axis. I'll publish it once
> it's done

oh, and thrust bearings too. The only thing I'm worried about (besides torque from the motor, where i'd just replace the motor with a more powerful one) is the processing speed of the gen6 electronics, does anyone know enough about this to tell me whether or not the mobo will step the motors fast enough so that the upgrades won't make the machine slow enough to be useless??

That sounds...expensive.

---

- akhlut

Just remember - Iterate, Iterate, Iterate!

[myhomelessmind.blogspot.com]

IMHO, leadscrew-driven X and Y axes aren't worth it unless every other component of your bot is already better and it's the X-Y motion that's holding you back. The weakest link in the chain is usually elsewhere. If you upgrade to screw-driven X and Y axes with high-quality 4-start ballscrews, and accordingly use to huge beefy NEMA 23 motors to drive them, you will indeed have fast, accurate X-Y motion… but it's wasted if you still have a not totally flat print bed, a wobbly hot end, a low-quality hobbed bolt, lousy bushings, bearings, or rods, imperfectly-printed X-ends, and so on and so forth. There's usually much lower-hanging fruit if you're looking to upgrade the accuracy of your X-Y positioning.

Edited 1 time(s). Last edit at 01/21/2013 01:59PM by Pointedstick.

if you use mach 3 hooked up to a parallel controller board, or a specialized controller it should work. without using belts, the timing inconsistencies in the stepping of the arduino, will limit acceleration. there are micro stutters and pauses even in processing data to gcode from information that is buffered properly. this has been debated back and forth before, but few people have both reprap electronics and mach 3 electronics to compare results with. And anyone I know that has a reprap and has threaded acme, does not use standard reprap electronics.

if using reprap electronics stay with belts. since you seem to have the resources for acme rods, and backlash nuts, i would hope you would invest in some sort of specialized controller to drive it.

surprisingly, these upgrades are only adding a few hundred bucks to the overall cost. the threaded rod can be purchased for around $30 a length and each nut also costs about $30 (for bronze machinable ones.) This will be my 3rd reprap, so i have some experience building them properly. @jamesdanielv thanks for the tips on the controller! It's FIRST robotics build season, so the project is on hold for about a month, but i'll let you know how it goes.

Cheers!

Quote

without using belts, the timing inconsistencies in the stepping of the arduino, will limit acceleration. there are micro stutters and pauses even in processing data to gcode from information that is buffered properly.

I'm pretty sure Teacup firmware doesn't do such stuttering. Looking at Repetier Firmware sources I find a proper serial line interrupt, too.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

personally i think there are benefits to be had to doing this, not nessarily for 3d printing but also for other applications,

lapping cheaper M8x1 or M8x1.25 threaded rod isn't difficult and could provide accurate positioning
antibacklash measures are a dime a dozen all with their place, the easiest and simplest way to go about that would be a springloaded half nut style printed piece

the most interesting change that would be able to take place which i think is the most important is the electronics

theoretically if you were running a leadscrew with a 1mm pitch then this would allow you to run full stepping or 1/2 stepping and still get 0.0025mm per step at 1/2 stepping of course the error in the leadscrew would mitigate this a little bit but you'd still be able to achieve less than 0.1mm without too much hassle,

where the electronics side of things could change is that much cheaper stepper drivers which are only capable of full or half stepping to be used resulting in cheaper and cheaper electronics and therefore cheaper and cheaper printers and a wider user group,

I'm tempted to model the parts needed just for giggles,

---

-=( blog )=- -=( thingiverse )=- -=( 3Dindustries )=- -=( Aluhotend - mostly metal hotend)=--=( Facebook )=-

Where you able to get the firmwear to work?
I want to use a lead screw for my bed, the same kind which im currently using for my z axis. They work great on my z axis but wont with bed, is there more i need to change in the firmwear besides steps per unit to get it to work? im using sprinter currently and my z axis is set to 1700steps, but if i go hight then 300 steps on my x axis the motor wont work and starts beebing. I already tried to mess with current and didnt work.
I dont get it im using the same hardware as my z axis so if anyone got any ideas to get it working, let me know .Tks

Maybe you need to turn down the max feedrate of the axis. With a 1mm pitch screw the motor spins 10 times to move the axis 10mm, compare this to less than 1 spin of the motor to move a GT2 gear 10mm. If the screw is trying to move 200 mm/second then you are asking the motor to spin 200 times per second compared to les than 20 spins of the GT2 gear per second. The motor cannot spin this fast so it just grinds.