Best Belt-Less RepRap and Printer Designs (Rack & Pinion Rostock )

@crispy1: nah, I knew the Trinitylabs leadscrew already, but was not too keen on ordering stuff that does not have publicly available (as in, you don't have to contact the seller) datasheets and / or performance information. To add to this, I am located in Europe, so importing them would be bothersome.
As both Igus and Reliance Precision Mechatronics provide a wealth of information on their sites, it was only natural to quote their data.

@uGen

Fair enough, I can respect those reasons.

I saw Trinity Labs had them

Not sureI am ready to drop $100 on them.

I did however pick up one of their little stepper motors with attached delrin gearboxes though.

I think 100$ is a fair price for these things. RPM charges around 100 pounds for their better (more suitable) versions. Igus is the cheapest, but I guess theirs is better suited for hand control mechanics (one example application was a lever for a locomotive that utilized the backdrive of the high lead in order to drive the screw by running the nut along it).

Sorry if you had trouble finding specs on the leadscrews I sell at trinitylabs. Here are some graphs that will show you performance as far as torque and speed for varisou sized nema17 and nema23 steppers in 1,2 or 3 stack running these 10 start 25mm pitch screwws with the constant force antibacklash nuts>>

These graphs were done in the lab at PBC while we were codeveloping the specific screws we use in our Aluminatus printers and the ones we sell in our store. This is made form true recorded data I witnessed firsthand >>

[brainspl.at]

Here is the vendor's information PDF about the technology in these leadscrews and constant force antibacklash nuts

[www.pbclinear.com]

You can get leadscrews and nuts direct from PBC or one of their distributors but you will pay more and the customized version I have had made specifically for the motion profile of 3d printing are not a stock available item from them.

Not trying to sell anyone here by any means and won't link to my shop but wantd to provide you with the spec sheets you were looking for and information on alternate sources then myself.

FWIW we use triple stack nema17's at 112oz-in with 10 start 25mm pitch PTFE coated screws and can attain a top speed of 390mm/sec with 40,000 acceleration setting for the XY axis in marlin firmware. But be ware that screws this fast do max out the 8 bot corens in arduino based electronics and work even way better on smoothieboard or 4pi or a 32bit arm based motion control platform.

Cheers-
-Ezra

Ezra, thank you for clarifying things and providing the .PDF's!
On paper, these leadscrews really look impressive and I am sure that they will also perform well. However, when I looked into this whole matter of alternative drive mechanisms, it became quite clear that wear life of these fast leadscrews would one of the more limiting factors (along with complexity of implementation due to the bearing blocks). Speed absolutely is not the problem; in fact, my configuration with a 20 tooth GT2 pulley and a pulley system similar to the Printxel would be slower than your leadscrews.
So, do you know the service life of the leadscrews? PBC really does not seem to like publishing this data as openly as others although their products seem to stack up quite well against the competition.

uGen-

We don't actually know the limits oif lifetime on these leadscrews with the constant force nuts and the SIMO stages I codeveloped with PBC. I recently visited their headquarters in chicago a week or two ago and got to see the whole room deveoted to the SIMO stages for TrinityLabs and the Aluminatus.

And they still have the very first production SIMO stage like on the Alu8minatus XY axis in a test harness and it has been running back and forth with a 4 kilo brick on the carriage at 300mm/sec since August 21st 24/7 non stop and still is not showing *any* signs of wear on the slider or the leadscrew,

The constant force antilash nuts are truly something special. rather then the traditional preloaded antilash nuts that push apart the twoi ends of the leadscrew nut with springs and compensate for wear by letting the plastic wear away more and more until it doesn't really compensAte for lash anymore the constant force nuts have round metal bands kind of rolled up like the element in a ciggarette lighter in a car wrapped around 3 places on the leadscrew nut where there are 3 or 4 "fingers" that have threads on the inside. So instead of getting contact with the screw and the threads only on one small ti[p pof the leadscrew nut the constant force nuts get cponstant force contact for like 20mm of thread and therefor do not wear nearly as fast.

I am not personally familiar with the supposed lomnger lasting screws anbd nuts linked above in this thread but the first production SIMO stage configured identically to the ones I am shipping on the Aluminatus has been running back and foreth across 320mm of travel at 300mm/sec speed and 10,000 acceleration setting since August 21st which is over 5 months of 24/7 300mm/sec back and forth travel without showing any signs of wear yet. If that won't last longer then most of the rest of yoru printer I do not know what will.

So while I cannot give you s[pecs or stats from the manufacturer, I can give you hard seen with my own eyes info about one of these sytages running almost at full speed back and forth with 4 kilo brick on the car for close to 6 months now without signs of giving up. This is a lot more then "a few prints" that folks are saying igus screws would be capable of doing. I think those igus screws would last a lot longer then a few prints honestly but i know for fact that I have no worries about the PBC linear constant force nuts and multistart leadscrews wearing out prematurely.

In fact I warantee the X and Y stages for a full year after purchase with free replacement no questions asked if they fail in the first year. Of coursze I'm not gonna cover your stepper motor because people can do dumb things and ruin those easily but the linear stage anbd leadscrew + nut is going to last you a full year of as much printing as you can do or I will replace it.

I have nbo doubts that Aluminatus printers will still be printing 5 yearsa from now based on the ones going out the door today.

In fact I bet you would wear out your GT2 belts and pulleys ,multiple times befotre you come close to wearing out these screws.

-Ezra

Where are pics showing higher beltless quality?
Pics on site do not show it.
I see repeative structures on vertical walls
and wavy traces after head direction changes.

I would be really interested in seeing the video of the SIMO stages moving back and forth @ 300mm/s @10k accel with a 4 kilo brick on them. I searched and couldn't find any of your vids with any speeds higher than 180mm/s and certainly not with a 4 kilo brick on them.

Thank you for the insight; as you have described the constant force nuts, they do make sense. Nice to see you trying out a different approach, too, because somehow generally, printer development seems to be a little bit stagnant at this time.
Maybe I will implement one of these leadscrews in my printer, too, but I have purchased my drive system already.

I am not too sure about your statement regarding wearing out the GT2 parts, though. In vehicles and industrial equipment, timing belts have to endure much more stress than on our printers, still, they hold up quite well because they are adequately specified and correctly implemented unlike in some of our printers. I chose the original Gates GT3 belt that is rated for higher tor higher torque than the previous GT2 generation precisely because I wanted a long-lasting system (I just don't know where the GT2 belts in various web shops come from; GT3 is downwards compatible). In addition to that, the idlers over which the back side of the belt runs are larger than the pulleys, just according to specs.
So, in theory, it should last me a long time since I constructed the belt path according to Gates' specifications, but I have yet to prove this assumption.

Seriously guys, spend five seconds researching your claims before making them.


Modern belts have essentially zero stretch. Your belts are not stretching, your machine has just gotten loose and all the parts are jiggling around.

Belts are used literally everywhere in ultra high precision machines. Cnc mills costing more than most houses use belts to drive their ballscrews. Such mills maintain accuracy in the micron range over decades. Gantry mills and routers use belts directly for linear motion and regularly move hundreds of pounds with 0.001" accuracy.


Leadscrews are acceptable but in most cases not beneficial for 3d printers. Belts are already accurate and rigid enough for a 3d printer, so a screw provides no advantages but increases the inertia of your load and limits top speeds.

PS. DumpsterCNC came up with your "constant force nuts" first: [www.dumpstercnc.com].


Regardless of how many dollars you spend on screws, belts, or other form of linear motion, the results will always be the same. If you can twist your printer into a pretzel by hand, the frame is going to move far, far, more than the belts or screws ever will.


If you are genuinely interested in making your machine perform better, people have been building machines for centuries that can hold tolerances three or four orders of magnitude tighter than printers. You don't need to reinvent the wheel.

691175002 Wrote:
-------------------------------------------------------
> Seriously guys, spend five seconds researching
> your claims before making them.

In some sense you are correct in that the fibres in the belt should not stretch ( but still do and that is why you can buy really expensive belts with Kevlar or steel in them). But the rubber on the outside can deflect a little as the pulley changes direction. On a CNC milling machine with belts driving a rack or lead screw this is not an issue because they change direction very slowly. But on a 3D printer it changes direction really fast. For the best prints we want really high acceleration and belts start to show backlash at that point. The above example from Trinity labs is changing direction and accelerating at over 1g (goes from 0 to 300mm/s at 10,000mm/s^2).

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

Quote

n some sense you are correct in that the fibres in the belt should not stretch ( but still do and that is why you can buy really expensive belts with Kevlar or steel in them).

I think all timing belts have Kevlar or steel in them and they are not massively expensive. Rubber belts without reinforcement are not timing belts and would be useless in this application.

That said they are not infinitely stiff, especially when they hang in a catenary. I wonder if turning them 90 degrees like the original Mendel X axis is significantly better as they would sag less. When used to drive a lead screw they are vertical, so would not sag with gravity.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> n some sense you are correct in that the fibres in
> the belt should not stretch ( but still do and
> that is why you can buy really expensive belts
> with Kevlar or steel in them).
>
> I think all timing belts have Kevlar or steel in
> them and they are not massively expensive. Rubber
> belts without reinforcement are not timing belts
> and would be useless in this application.
>
> That said they are not infinitely stiff,
> especially when they hang in a catenary. I wonder
> if turning them 90 degrees like the original
> Mendel X axis is significantly better as they
> would sag less. When used to drive a lead screw
> they are vertical, so would not sag with gravity.

All general purpose timing belts I have seen start with fibreglass reinforcements and then Nylon and the Kevlar or steel, the latter two being the best and most expensive.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

Just remember that if the belts have steel strength members you will need to be mindful of minimum bend radius requirements. Otherwise the fibers will fatigue and break over time, even with low tensile loads.

Wired1 Wrote:
-------------------------------------------------------
> Steel wire has to be the way to go. It doesn't
> stretch (normally) and is flexible enough to do a
> couple of turns around the drive motor.

Absolutely !
And I think wire-driven axis will be the standard system in repraps. It's easy to find (stainless steel fishing wire) and it doesn't need precision machined pulleys, simple printed cylinder will work just fine.
The important thing to remember is that it needs to be flexible. 7 strands wire doesn't work, I tried and it didn't pass the stress test, 49 strands wire works great.
You can see my system on the Heavy Mendel thread : [forums.reprap.org]

Things I printed with it for the print something really tiny competition :
[forums.reprap.org]
[youtu.be]