[ruthlessratter]

Hello there,

I have started building my own CNC router and had a few questions about the drive for the long Y axis. I have decided on a rack and pinion system for driving the axis but wondered if the pinion needs to be spring loaded onto the pinion. I have looked at many homemade and kit built machines, these all seem to have the stepper motor on some sort of adjustable tensioner or spring to keep the pinion on the rack. Would it be better to have the motor fixed solid to the frame using a suitable shim to give a minimal amount of clearance between the rack and pinion? The only reason i can see for having a tensioner is if the rack itself is not completely straight and the tensioner would 'iron' out any of these disscrepensies. Is this correct or am i missing something obvious?

Thanks for your help peoples.

Dave

[Web Goblin]

Much better to have it spring loaded. This will keep the correct tension on the pinion. This will also allow for the pinion to move out away from the rack in case of anything getting jamed between the rack and pinion and will also allow it to follow the rack allowing for any deviation in it.

[ruthlessratter]

Obviously there would need to be two rack and pinion set ups, one for each side of the gantry. What is the best way to drive them - would it be a single motor situated in the gantry itself driving a shaft with a sprocket either end to drive each pinion on either side ....or .... twin motors indipendently driving singular pinions?

[JerryFlyGuy]

Ruthless I've got such a setup and while it works I would recommend a bit more attention to detail on the installation. My system uses pressure fed pinions using a gas cylinder to load the pinion into the rack. Over time this accelerates wear and I've had to replace both the pinions [a few times] and the rack [once]. Better would be to use your rail to temporarily mount your spindle and machine an aluminum surface to mount the rack to. Then using an adustable but LOCKABLE motor mount such that while the pinion is in contact w/ the rack and there is no lash, the pinion is not 'floating' or forced into the rack. This will give you the ability to control wear a bit better and give longer life to your system. I've got a machine that is set up this way and I've yet to replace the rack or pinion [however after 4+ yrs of cutting 8-15hrs /day its now getting time to replace the pinions].

The other option is a NEXEN rack system but thats a bit more money again..

Fwiw

JFG

[ruthlessratter]

hi,
Thanks for your thoughts. I agree with what you have said Jerry. I thinkl that as long as the rack is mounted correctly then the pinion should be fixed in a constant position with the option of adjusting it as and when is needed due to slight wear.
After further thought last night I have decided to run a 12mm shaft through the gantry which will drive a toothed belt down to matching pinions on each side.

[JerryFlyGuy]

Ruthless, this is certainly an option. How large is your gantry? I've seen routers that use this drive method and while it does work, there are instances where harmonic's from the belt are seen as defects in the part. It all depends on what quality of parts your concerned with etc [what is your largest area of concern.. ie; if it's an oxy fuel table having micro-meter accuracy in your drive system isn't a concern..]. Belt drive systems work but they are a compromise to a degree. If you do go the belt route but would like a leg up on quality of the drive system, spend a little extra and get the Eagle PD from Good Year. They have a herring bone pattern vs the straight tooth of a standard HTD belt.. the herring bone pattern will help [some] in the area of harmonic's and vibration due to belt speed and tooth cogging, this will help to reduce jutter and other defects in your parts.

Fwiw

JFG

[ruthlessratter]

The machine that I am building is a cnc router, the gantry is only 1000mm wide and will run up and down a rail 2000mm long. Do you think that it would be better to use two separate motors to drive the pinions, this would eradicate the need for any belts. My knowledge of mach 3 is fairly limited but i believe that you can run a second motor as a slave.

[JerryFlyGuy]

On a machine of that size I believe either system would work? You might also be able to get away w/ a single [center] drive ballscrew down the middle of the machine? I'm not sure what the 'rules of thumb' are for single point drives vs dual drive? [where the cross over point is?]

Mach 3 will Slave a second axis [usually the 'A'] to the X axis for dual driven gantry [this is how my personal router is set up].

The one thing that is nice about Dual Drive gantries is the ability to 'Auto-Square' the gantry during homing, where as a single point drive [either with a jack shaft with belts or a single center point] do not sqr themselves automatically.

As with EVERY design project, there is always a compromise w/ EVERY design decision. It all depends on what your project goals were when you set out down this road.. [and oh what a fun and exhilerating road it can be ]

Fwiw

JFG

[ruthlessratter]

Cheers for your advice, ill upload some photos and keep you posted on my progress. Thanks again.

[ahren]

Having gone through the development of our rack and pinion drives about 4 years ago, I can say that unless you can more or less perfectly align your gear rack and your linear motion system, you'll constantly be fighting a statically mounted pinion drive. We've had great success with our pivot mounted and spring tensioned drives, and it's really nice to be able to drop the drives off the rack easily to work on things, and then hook them back up again without a lot of hassle.

As to belts, we use GT series belts that are low-backlash and offer excellent performance. We dual-drive our systems, both to avoid mechanical complications and to get the extra power from a second motor pushing the gantry, which allows for higher speeds.

Good luck with your build!

Best regards,

Ahren
CNCRouterParts

[JerryFlyGuy]

Having gone through the development of our rack and pinion drives about 4 years ago, I can say that unless you can more or less perfectly align your gear rack and your linear motion system, you'll constantly be fighting a statically mounted pinion drive. We've had great success with our pivot mounted and spring tensioned drives, and it's really nice to be able to drop the drives off the rack easily to work on things, and then hook them back up again without a lot of hassle.

As to belts, we use GT series belts that are low-backlash and offer excellent performance. We dual-drive our systems, both to avoid mechanical complications and to get the extra power from a second motor pushing the gantry, which allows for higher speeds.

Good luck with your build!

Best regards,

Ahren

Ahren, I agree that if you don't want to go to the effort of properly aligning the rack to the rail your only option is to use a floating pinion. Care must be taken however, to make sure sufficient pressure is supplied to the pinion to maintain it's contact w/ the rack under maximum acceleration. This then makes the pressure system dependant on motor size. On my machine [which uses this style system but w/ gas shocks to supply 'counter pressure'] the force encountered [trying to seperate the pinion from the rack] was sufficient enough that it required a fairly good amount of 'counter pressure'. Of course to be safe, this 'counter pressure' needs to be 2-3x [or more] the max seperation pressure otherwise there would be positional error during high acceleration moves as the pinion would partially climb out of the rack. The down side of this senario is that the counter pressure causes pre-mature pinion [and rack] wear. In my case I'm operating at 90% of the rated pinion loades so the pressure is at the upper end of the range. I am seeing ~500-800hrs operation per pinion set. The pinions are changed when imprefections are seen on the finished parts. [more often it's 'judder' marks on the part and sometimes [less often] a dimensional failing due to reduced pinion size]

I own a machine that has the pressure mounted pinion and I operate and maintain a machine that has the fixed pinion setup [at work].. I prefer the fixed pinion setup for it's ease of maint. and fewer head aches.

I am in the process of planning several upgrades to my machine and one of them is to fix the relationship of the rack to the rail and thereby elliminate the floating nature of the pinion to the rack.

It's actually not really that difficult a process to complete. If your rails are straight and installed, it requires a simple [temporary] mount to be made for a router body that directly mounts to carriage on the rail. This carriage is then hand fed down the rail and the router machines a sqr true edge into an aluminum bed [pre-mounted Alum FB material located where the rack will mount. The rack 'bed' is now sqr and parallel [in relation to the rail] for the rack to be mounted to. [if one is opposed to 'hand feeding' a router to do this process a simple cable drive system could be arranged w/ very little extra effort] Via this process you've got a matched rail to rack mounting bed and the need for a floating pinion is elliminated. Step the rack/pinion up to a Nexen system and you've got the equivilant of a Ballscrew for smoothness of motion [and accuracy at +/- 0.001" over the entire length of an installed system]

As with every decision, there is a level of cost [and or time] associated with it. It has been my experiance that the added need to purchase new parts, have them machined and installed plus [possibly] re-doing parts and the unwanted down time are offset in an economically viable way by completeing this process. On a machine that was intended for a less strigent 'up time' and the willingness to tinker now and again the simpler method is probably satisfactory.

I think the one thing about belts is that no two designs are exactly the same and no two installs react the same. One install will have harmonic issues at one speed while the other will have the same but at a different speed. Again belts come down to economics. You'll not see them on many high performance [commerical] systems due to these reasons. On a hobby level system they can be more than adequate.


To each their own, to be sure

JGF

[dmalicky]

I imagine this has been tried, but I've no idea how well it works, or not:
1. Mount the rack to the side of the frame/gantry, with teeth pointing down. Drill holes through the side of the rack and bolt to frame, so the "back surface" of the rack (opposite the teeth) is clear.
2. Use a floating pinion/motor/plate assembly like CNCRP, but instead of spring loading, locate the pinion by running an adjustable-position skate bearing on the rack's back surface. That is, the skate bearing follows that back surface, which sets the pinion position relative to the rack.
Then the rack doesn't have to be mounted as straight, but the pinion stays at the same distance from the rack.

[JerryFlyGuy]

Unfortunatly that doesn't keep the rack sqr to the pinion. It would work but you'll find it creates un-even wear and could add to the potential for inconsistance results [movement].

JFG

[louieatienza]

I imagine this has been tried, but I've no idea how well it works, or not:
1. Mount the rack to the side of the frame/gantry, with teeth pointing down. Drill holes through the side of the rack and bolt to frame, so the "back surface" of the rack (opposite the teeth) is clear.
2. Use a floating pinion/motor/plate assembly like CNCRP, but instead of spring loading, locate the pinion by running an adjustable-position skate bearing on the rack's back surface. That is, the skate bearing follows that back surface, which sets the pinion position relative to the rack.
Then the rack doesn't have to be mounted as straight, but the pinion stays at the same distance from the rack.

If you have the motor mount on a pivot, then you'd have varying clearances when the motor mount moves. The motor mount would have to be made in such a way that it moves perpendicular to the rack. The other thing with a pivoting motor mount that I see is, if the angle of the pinion path-pinion-pivot is around the 45deg. or 90deg. position then any movement of the mount would translate to a very small position error. I'm sure it's very small but it is there. With the pivot coplanar to the pinion path any movement of the mount would yield a negligible position error.

Also not mentioned as far as wear on r&p systems is lubrication. All commercial r&p's have them, but I don't remember seeing any DIY r&p builds that have an auto-lube system of some sort.

Setting a slave axis in Mach is easy, though with r&p you may have to reverse the motor's direction.

I'm not sure getting a 2000+mm center driven ballscrew is the best option, duw to the need for a large diameter screw to prevent whip and sag, and a larger stepper to counter the increased inertia of the larger screw. Unless however, you use a spinning nut design.

[JerryFlyGuy]

louie I agree with you on the principle of the R&P system. Mine is designed such that the center of the pinion is in line w/ the center of the piviot point of the tensioner, with this setup the +/- error imparted by the pinion moving in and out is fairly small [well inside all my other tolerances] so it can be done.

Lubrication in a system w/out bellows would be a magnet for dust and grit etc. Mine aren't lubed for this reason, not that I couldn't have devised a system that would have. This does accelerate wear but I am using hardned R&P which helps in this regard.

I've got a 2500mm ballscrew on my Z axis [5/8" od IIRC] and I spin it plenty fast w/ no whip to speak of [I've never seen it]. The bigger concern is if the gantry would stay sqr on a setup like this.. anything can be made to work but it may require to large a spacing on the X axis rail carriages, such that it negates the advantage of the single drive point solution.

To everything there is a compromise

JFG

[dmalicky]

Jerry and Louie, Thanks for the replies and info. Still, I'm not sure why the idea would be any worse for alignment/squareness than spring loading like CNCRP? (I'm assuming a well-executed pivot for the motor/rack plate, square to the rack.) Hmm, actually it should be better, as the "follower bearing" on the back surface of the rack would apply its force directly opposite the pinion, so the plate and pivot experience ~no twisting torque. (With spring loading, the spring force is a few inches out of plane from the rack and pinion.) Or maybe the concept isn't clear and I need to post a sketch?

Yes, I agree the motor plate pivot should be at the same "height" as the rack teeth, so the pinion moves perpendicular to the rack. I'm surprised that more machines don't do that. With the more typical pivot location, I would think the pinion slips at different accelerations for + and - axis movements (or needs excess spring force to prevent slip).

[louieatienza]

Jerry and Louie, Thanks for the replies and info. Still, I'm not sure why the idea would be any worse for alignment/squareness than spring loading like CNCRP? (I'm assuming a well-executed pivot for the motor/rack plate, square to the rack.) Hmm, actually it should be better, as the "follower bearing" on the back surface of the rack would apply its force directly opposite the pinion, so the plate and pivot experience ~no twisting torque. (With spring loading, the spring force is a few inches out of plane from the rack and pinion.) Or maybe the concept isn't clear and I need to post a sketch?

Yes, I agree the motor plate pivot should be at the same "height" as the rack teeth, so the pinion moves perpendicular to the rack. I'm surprised that more machines don't do that. With the more typical pivot location, I would think the pinion slips at different accelerations for + and - axis movements (or needs excess spring force to prevent slip).

If you're using a "follower" bearing of some sort opposite the pinion with a pivoting system, then whenever the pinion/bearing moves relative to the pivot, the clearance between the pinion and bearing will change. You'd either have to spring-load the follower bearing or have the whole assembly move perpendicularly to assure the "gap" doesn't change relative to the rack.

[JerryFlyGuy]

Dma I think it could be made to work however, the effort to do so would be suprising. To make it work well, it'd take alot more than just the simple system your describing. As louie has pointed out, the two [pinion and bearing] would have to move independantly to each other but relative to each other so now you've got two pivot points [they could be co-located but it's still two pivoting arms]. If you go w/ the perpendicular slide then you'll run into flexing issues. It'd be interesting to see a drive done this way but I don't think it'd perform nearly as well as properly exacuted R&P system on a fixed bed w/ a fixed pinion. Any time a flexing or moving bracket is added to a system it has the potential to induce flexing, mulitiple moving brackets stacked on each other and your building a big ol' tuning fork.

I'd hazard a guess that to make the system your describing, would take more effort to accomplish than a simple fixed pinion system and by a fairly good margin.

I'm not sure why people shy away from building their system with fixed rack and fixed pinion. It's actually very easy to do, the most difficult part being building the temporary mounting for your router head to mount it directly to a rail carriage [or two]. Once thats built it becomes a very very simple proceedure to exacute.

[Actually I do know why.. when I did mine w/ the floating pinion I thought it'd be much easier to do using the floating pinion. No rack alignement worries, no temporary setups etc etc.. in reality I spent more time trying to 'beat' this than I would have spent if I'd just have stopped and done it right the first time.

Fwiw

JFG

[dmalicky]

Thanks for the replies. Yes, if the assembly had to rotate a large amount, that clearance variation would be deal-breaker. But it only needs to rotate a tiny bit, and I think the clearance variation is some function like (1- cos(small angle)), which is very stable. It's easier to analyze with CAD than geometry... let's say the rotating plate is 6" from pivot to pinion, and the clearance between the pinion_pitch_circle and follower_bearing (rack "thickness") is 0.500". Now, how bad should the rack be mounted? I'd guess < 0.030" parallel to the rail is typical... I'll use 0.060" to be safe. The first attachment shows a clearance variation (minimum vertical distance between the roller and pinion) of 0.00006" for either cw or ccw rotation.

Starting with a 30 deg rotated assembly gives the same average clearance variation, but it's different cw vs ccw (second attachment). It's still within 0.0001", so the follower bearing should work for that config, too.


Unrelated to the follower bearing: given the tiny rotation angles, the system could be further simplified by dropping the pivot pin/bearing and making the plate/arm mounting area flexible enough to accommodate the 1/2 degree rotation. I.e, change the pivot pin to a flexure / living hinge. The flexure would need to be stiff in longitudinal torsion, though, to keep the pinion square to the rack.

[louieatienza]

Thanks for the replies. Yes, if the assembly had to rotate a large amount, that clearance variation would be deal-breaker. But it only needs to rotate a tiny bit, and I think the clearance variation is some function like (1- cos(small angle)), which is very stable. It's easier to analyze with CAD than geometry... let's say the rotating plate is 6" from pivot to pinion, and the clearance between the pinion_pitch_circle and follower_bearing (rack "thickness") is 0.500". Now, how bad should the rack be mounted? I'd guess < 0.030" parallel to the rail is typical... I'll use 0.060" to be safe. The first attachment shows a clearance variation (minimum vertical distance between the roller and pinion) of 0.00006" for either cw or ccw rotation.

Starting with a 30 deg rotated assembly gives the same average clearance variation, but it's different cw vs ccw (second attachment). It's still within 0.0001", so the follower bearing should work for that config, too.

Unrelated to the follower bearing: given the tiny rotation angles, the system could be further simplified by dropping the pivot pin/bearing and making the plate/arm mounting area flexible enough to accommodate the 1/2 degree rotation. I.e, change the pivot pin to a flexure / living hinge. The flexure would need to be stiff in longitudinal torsion, though, to keep the pinion square to the rack.

This is a great illustration, but it in fact proves JerryFlyGuy's point. If you're using even supported round rail, the bearings probably have about .0001"-.0005" clearance. Why would you even have to bother then with the follower and pivot? It becomes redundant. All you'd have to do is datum your rack to the linear rail. With a few basic tools and some home brew jigs, it shouldn't be the too much fuss to install the rack accurately enough. Way more accurate than 1/32" (.032") for sure!