Design Idea for tractor style belt on belt.

I have been gathering ideas for building my own ~24"x36" CNC Router. I really like the belt on belt design for linear motion. I have an idea for a variation on the belt on belt that I haven't seen before and I would like input as to positives or negatives.

It seems the existing belt on belt designs engage the stationary belt with 2 sections of belt.(See right side of pic below) This then requires adjustment to tension the belt while ensuring proper tooth spacing between the different sections of belt. Lots of layouts could have this facing the inside of router and visibility of the 2 belt adjustments maybe hampered. 2 reasons I could see for this is backlash prevention when there is space where the belts engage or it uses a more readily available single side belt.

My idea is to use a double sided belt (See the middle pic on the left side). This would reduce the adjustment to just belt tension as tooth spacing would be fixed. It also seems to be a simpler design with fewer moving parts. From a backlash perspective, I think finding a belt on belt profiles that press into each other instead of having a bit of gap between the teeth would reduce backlash significantly. So I did a comparison of various 5mm belt profiles meshing shown below. It looks like the T5 on AT5 belt has no space between the teeth while maintaining good engagement.

Let me know your thoughts.

Thank you,

Jerry

Hi Jerry - Why on such a small machine? The reason to use belts is to reduce cost. But two belts cost twice as much. A 24"x36" machine is well served by a single U drive or a continuous belt. A well tensioned belt has zero backlash anyway as it is snugged into the pulley as its designed to do. If you are looking for extra precision over an AT belt get an ATL belt about 25% more cost but its a zero gap belt for precision applications. The issues with the everman and tractor style systems are: 1) if you glue the static belt down the drive belt will stretch slightly as it beds in then it will have a slightly different pitch to the glued one hence it becomes a mismatch. If you don't glue the static down you still have the same problem and I expect it will be painful to tune the two together. If you are worried about stretch on a 36" machine its really not a problem if you size the belt correctly. I find they bed down and you reset them a few times and within a couple of weeks they are good forever. For 36" I'd use a 16mm or 25mm belt system. I prefer continuous drives so the motors are not on the gantry. It will be a close call in cost to a ballscrew system as you have so many idlers and eccentrics to deal with. My guess is that it will cost more and take up more space. Space is critical on small machines...

The tractor style works and its used on large plotters but a router implies you want to cut something? (plastic timber? metal? ) again a small machine is fine with a conventional belt (use a big one) or a ballscrew... using the double sided belt solves a couple of issues but its a small gain compared to using a ballscrew. If its test for a 20m bench then that's another story... use an ATL belt for the static belt and a std spec loop for the tractor.

Re: mixing belt profiles is a bad idea, the contact loads will wear them quick sticks, the load has to be well distributed and they are designed to do this around a pulley not to each other...

If you pursue the idea don't use the middle idler its not doing much. The work is being done at the ends of the traction section. I assume one will be an eccentric so the traction bit can be stretched slightly... It the first two or three teeth transferring the load...Peter

Pete,

Thank you for the valuable feedback. Can you tell me what a U drive is?

I'm looking at carving wood, wood on a 4th axis and 1/4" to 3/8" aluminum. The aluminum work would be making things like plates for adding CNC capabilities to my manual engine lathe & mill along with other maker projects. I'm looking at a 2-3kw spindle. When you mentioned "normal" belt drive it seems a lot of people use that system for wood, but not as much with aluminum. I'm looking a machine that is rigid enough to handle aluminum with moderate cutting speeds. Am I correct in thinking ball screws would be the best for cutting aluminum on a machine this size? Does screw whip become much of an issue with a ~38" long ball screw?

Thank you,
Jerry

Hi Jerry - for aluminium cutting you need a stiff system. Although belts can be sized to do this, I'd go ballscrew. They will be significantly stiffer and cost the same as you are planning. Whip is not a problem at that size machine. Hiwin publish formula to check this.

https://motioncontrolsystems.hiwin.c...n-configurator

U drive see image attached. A U drive is a moving motor solution. A continuous belt is a stationary motor solution. The U drive uses a shorter belt but it bends the belt in two directions which fatigues it faster.... I prefer the continuous belt. "tools" is one of my small machines. 10mm AT belt. Cuts plastic, timber etc very well. Struggles for various reasons on aluminium. Was intended as a plotter but once I build something I tend to torture test them...

Cheers Peter

I will brush up on ball-screws. I see it as being simpler than the belt on belt and more rigid than normal belt drive.

A couple more questions (if I'm not getting annoying :) )

I'm trying to squeeze the router cross ways into a galley shaped workshop so over length is a factor. If I use hybrid steppers, those get pretty long and would stick out 4-6"/ Do you see using a belt drive to be able to put the motors under/in the frame (See pic) as an OK idea or will this be a sacrifice over a motor directly coupled to the ball screw? I have a ways to go to figure out if there is an advantage to be had by a motor to screw speed ratio available when using a belt.

Also, in the pic, the ball screws are above the slides and protected by a channel over the top of them. But I was thinking having the rails above the screws would add a little more ridigity to the moving gantry. I'm looking at using 20mm linear ball bearing rails.

Hi Wire - Why use hybrids? They cost more and I'm not sure their cost/benefit ratio is there. Make the gantry (X axis) wider then the depth (Y axis) to improve its fit? Reflexing the motors is good to save space. Its done a lot I don't see a big negative. Really comes down to the target accuracy. If you use a direct coupling you get the best accuracy if you use a belt you have some compliance. In timber work no big deal but in the aluminium this maybe an issue. What tolerance or accuracy are you aiming at... but a short wide (20mm or 25mm belt is very stiff) 20mm rails is more then adequate for what you are doing get medium or high preload cars though makes a difference when cutting metal and they wear slower... Peter

Swapping the X & Y lengths has got me thinking. I see several advantages to that and a 40" vs 28" gantry is not that big of an increase as far as ridgity is concerned. Time to start skectching some more. Thank you for all your help and have a good night!

Hi Wire - Lunch time here a while to go to bedtime. :) Peter