Eliminating belts from the design?

CaptnBlynd · Oct 15, 2015

I am designing my latest FDM 3D printer. Since I am returning to school and will have use of high end machine tools, I am trying for a design on the light commercial scale. I do not like belts. I see them as a source of vibration and play. Everything I've done until now has been belt driven.
I am thinking about "rack and pinion". Secure the racks to the x and y axis, mount both motors on the carriage unit then have the steppers travel on the racks. Anybody tried this or used a machine that has it? Any other belt replacement ideas I can play around with?

I plan this to be a significant improvement over my usual home user toys, so if you have an interest in building 3D printers, this will be an interesting project. Multi-head, Bowden feed, high temp for exotics like nylon, all the fun stuff. If you are bored and want to see a couple of my earlier builds/designs, go to thingiverse.com and type "captnblynd" in the search field. 2 of my "Minion" series, complete with files you can print, are there.

hanermo · Oct 15, 2015

I am the opposite.
I think belts are great, very repeatable, very fast, and provide excellent dampening.

I would just use bigger belts.
They really should be tight !

The specs on a HTD5/15 are about 30 kg tension, iirc. 70 lbs.
Cant be done, directly, on stepper shafts.

Fixed belt, moving motor, or round belt with stepper driving ine end with *big* rigid coupler.
Either can be tight, and will provide excellent speed, and repeatability.

Just installing very rigid, accurate, belt drives on new brushless ac servos on some of my machines.
HTD8/30.

dsergison · Oct 15, 2015

hanermo said:
I am the opposite.
I think belts are great, very repeatable, very fast, and provide excellent dampening.

I would just use bigger belts.
They really should be tight !

The specs on a HTD5/15 are about 30 kg tension, iirc. 70 lbs.
Cant be done, directly, on stepper shafts.

Fixed belt, moving motor, or round belt with stepper driving ine end with *big* rigid coupler.
Either can be tight, and will provide excellent speed, and repeatability.

Just installing very rigid, accurate, belt drives on new brushless ac servos on some of my machines.
HTD8/30.

Agree. Rack and pinions are going to be looser due to gear lash. And now you have the motor inertia. And more cabling.

there is a damn good reason in every photocopier, scanner, laser engraver... you find a quality belt drive.

it you insist on dumping timing belts use metal belts or cable http://belttechnologieswebsitecdn.b...ntent/uploads/2011/08/Page-Images-Linear1.jpg

SRA2 · Oct 15, 2015

Servos seem to be the most precise without belts

dsergison said:
Agree. Rack and pinions are going to be looser due to gear lash. And now you have the motor inertia. And more cabling.

there is a damn good reason in every photocopier, scanner, laser engraver... you find a quality belt drive.

it you insist on dumping timing belts use metal belts or cable http://belttechnologieswebsitecdn.b...ntent/uploads/2011/08/Page-Images-Linear1.jpg

I recently toured a robotics lab with some really large 3D printers running off an inverted tripod design with servo controlled motion - they were crazy fast and highly accurate.

Rich

billzweig · Oct 15, 2015

Belts have a number of advantages, besides the simplicity and low cost. For precision work the timing belt is more accurate than rack as the belt is in contact with a larger number of the sprocket teeth and thus provides an averaging effect - the bigger the sprocket the more accurate. You do not have this in rack and pinion with the limited contact. One can use a leadscrew for the travel, this will be better than a rack, though much more expensive and complicated.
If you do not like the timing belts, there are other type of belt as the posi-drive etc.

CaptnBlynd · Oct 15, 2015

OK, that makes sense. By moving the stress to the frame and off the motor I can get belt tensions enough to minimize lash. Easy design even. Thanks. I've already started sketching that in.

awander · Oct 15, 2015

SRA2 said:
I recently toured a robotics lab with some really large 3D printers running off an inverted tripod design with servo controlled motion - they were crazy fast and highly accurate.

Rich

That is a called a "Rostock" design. They also use belts(at least the ones I have seen do).

SRA2 · Oct 15, 2015

Delta

awander said:
That is a called a "Rostock" design. They also use belts(at least the ones I have seen do).

I'm thinking of more of a delta robot design - I dont think there are any belts involved...

Take a look at this example:

Here's the root page: https://en.wikipedia.org/wiki/Delta_robot

awander · Oct 15, 2015

It just happens that I own a SeeMeCNC "delta robot" like the one in the photo. The belts are hidden inside the aluminum uprights.

SRA2 said:
I'm thinking of more of a delta robot design - I dont think there are any belts involved...

Take a look at this example: View attachment 151751

Here's the root page: https://en.wikipedia.org/wiki/Delta_robot

Pete F · Oct 15, 2015

A belt design introduces neither vibration, nor more backlash over alternative linear motion drives. While any engineering is a compromise, I feel you are barking up the wrong tree here. What is far more important is to reduce the flying weight and system inertia in order to increase speed. You may want to look at Hbot/CoreXY printers to achieve that. There are other drive configurations available that don't require belts.

The alternative suggested is not called a Rostock, that is just one version, it is a Delta printer, and yes they can be built without belts, with the stepper motors directly coupled to screws.

garyhlucas · Oct 15, 2015

The main reasons belts introduce vibrations is they are undersized, the pulleys have too few teeth, or the pulleys are poor quality with an inaccurate profile. That delta printer shown is a rostock variant, and a poor one at that. Goggle Berrybot to see what an awesome rostock can do. I started building a large one of my own. I have all the materials, and the design done in SolidWorks but somehow life intrudes.

motion guru · Oct 15, 2015

The Bell Everman servo belt actuators give extremely accurate and dynamic response for reasonable cost for large systems.

We have built a number of larger high speed linear motor systems with very high accuracy and stiffness. . . 4m/sec rapids and 1/2 micron resolution and 5G accelerations, but your pocket book needs to be pretty stout.

CaptnBlynd · Oct 15, 2015

"The main reasons belts introduce vibrations is they are undersized, the pulleys have too few teeth, or the pulleys are poor quality with an inaccurate profile. That delta printer shown is a rostock variant, and a poor one at that. Goggle Berrybot to see what an awesome rostock can do. I started building a large one of my own. I have all the materials, and the design done in SolidWorks but somehow life intrudes."

Want some inspiration? Look what these people did with a rostock. Printed carbon steel on a home machine. https://www.academia.edu/5327317/A_Low-Cost_Open-Source_Metal_3-D_Printer

hanermo · Oct 16, 2015

Exactly.
On CNC systems, with the mentioned medium HTD I could get 1 micron resolution *mechanical movement* at motion-end, in 2004.
As measured via DTI.
Via belts, driving a fairly rigid system.

You just need a good belt system, and enough overhead.
Ie servos.

motion guru said:
The Bell Everman servo belt actuators give extremely accurate and dynamic response for reasonable cost for large systems.

We have built a number of larger high speed linear motor systems with very high accuracy and stiffness. . . 4m/sec rapids and 1/2 micron resolution and 5G accelerations, but your pocket book needs to be pretty stout.

automater · Oct 29, 2015

CaptnBlynd said:
I am designing my latest FDM 3D printer. Since I am returning to school and will have use of high end machine tools, I am trying for a design on the light commercial scale. I do not like belts. I see them as a source of vibration and play. Everything I've done until now has been belt driven.
I am thinking about "rack and pinion". Secure the racks to the x and y axis, mount both motors on the carriage unit then have the steppers travel on the racks. Anybody tried this or used a machine that has it? Any other belt replacement ideas I can play around with?

I plan this to be a significant improvement over my usual home user toys, so if you have an interest in building 3D printers, this will be an interesting project. Multi-head, Bowden feed, high temp for exotics like nylon, all the fun stuff. If you are bored and want to see a couple of my earlier builds/designs, go to thingiverse.com and type "captnblynd" in the search field. 2 of my "Minion" series, complete with files you can print, are there.

If possible go with direct drive or 90 degree gear boxes this will make the overall footprint larger but it will be well worth it and yes there will be some that will play the torque card etc. but look into it yourself.
Bob

Skeeter · Oct 30, 2015

I have a friend that has built him a cnc router.(5x11 ft.) He used a Kevlar timing belt.This was to eliminate backlash.He also used one ball screw arrangement on the up and down movement on the router. The ball screw set is quite pricey.

These units do work very well.

CaptnBlynd · Nov 2, 2015

Why not ball screws? This seems to be the system of choice for 3 axis CNC.

Oldwrench · Nov 17, 2015

In low-inertia conditions belts are inherently faster than screws, they are always smoother than gearing, and they don't exhibit localized wear. We have a laser engraver with belts and its stability is superb--it will mark tiny 4-point type over an 18 x 24 area with perfect repeatability. I don't know how heavy the deposit nozzle is on a printer but it can't possibly be enough to need a rack and pinion gear or a ballscrew except for maybe elevating the table. JMO, I'm not planning to build one.

Eliminating belts from the design?

Plastic

Titanium

Diamond

Plastic

Stainless

Plastic

Stainless

Plastic

Stainless

Titanium

Stainless

Diamond

Plastic

Titanium

Plastic

Aluminum

Plastic

Titanium

Similar threads