asking for rail binding opinions

i have read many comments about rail binding. i am about to mount my axes on my first machine and am concerned that if my rails arent exactly straight, they may bind. i am using brass bushings in printed holders.

is there anything that i can do to reduce the risk?

has anyone tried a silicone adhesive in stead of epoxy to bond the bushing holders to the axes? to allow some slight auto-adjusting?

zipties (tie-wraps) probably add more flexibility than an epoxy connection too.

thoughts and comments welcome.

Edited 2 time(s). Last edit at 09/26/2012 12:20PM by rrr7.

Measure between the center of your X bars, are they the same?

The more ridid you can make your setup the better... BUT the more accurate your measurments have to be...

Get out a piece of string, and start half-ass measuring things, if you can get a caliper in there even better, if not measure with string and have someone measure it with calipers, or even a tape measure for now, get EVERTHING the same, make sure your axis is at the extreme that it will ever be before you measure anything, as the asix will change whatever it is (rails, lead screw, belt, etc.) to what it WANTS, that is the measurement you need. Don't put your X all the way to the left, then measure the right, that makes no sense... see what I mean? take things apart, file them, cut them, etc to get the "perfect" and then you can have better prints.

I haven't researched brass bearings/busings on these things, but I assume it's Oilite brass? If not your wasteing your time... I would ream out your bushings to atleast .2mm more, do you have a reamer (not a drill bit)? OR, if you have some left over drill rod put some FINE rubbing compound on it and run your bushings up and down it about 50 times or so, then clean them all really good and re-install on your machine, actually that's not a bad idea to do from the start anyways... SMOOTH is the key in everything, don't be afraid of rubbing compound or 1200 grit sand papper..

There shouldn't be any need to ream a bushing out unless it doesn't slide smoothly on the rod. Scratches and imperfections should be sanded out of the rods though. A knife sharpening stone works nicely.

Make sure the bushings slide nicely on the rods, then assemble everything, best guess. I'm guessing you're building a Mendel of some kind, so once the frame is assembled and tightened, try the axes and make adjustments until everything runs freely. This kind of adjustability is built into the design. That's part of why everything is threaded rods.

Once thing I did find necessary when installing bushings into the carriages (MakerGear Prusa) was to use the smooth rods to align the bushings to each other. Assemble the carriage and slide a rod though. The rod probably drags. Hold the rod on either side of the carriage and press against the carriage with your thumbs. One over each bushing. Press with enough force to move the bushing in its bore, but not enough to bend the rod, of course. MakerGear's Prusa came with self-aligning bushings, and this process aligned them. Bushings that are not self-aligning may benefit from the same operation. LM8UU bearings shouldn't need any of this.

I've found that even with LM8UU bearings, a high degree of accuracy in the linear rods is necessary - they'll move if your rod isn't well aligned, but they'll wear out quickly and/or cut grooves in the rods.

On my latest build, I've used a hybrid approach - LM8UU (or printed equivalents) on one rod, and a pair of roller bearings on the other rod. With this arrangement, precise alignment of the rods is not necessary, because the roller bearings can drift side-to-side from one end of the rod to the other.

rrr7 Wrote:
-------------------------------------------------------
> i have read many comments about rail binding. i
> am about to mount my axes on my first machine and
> am concerned that if my rails arent exactly
> straight, they may bind. i am using brass
> bushings in printed holders.

As other people have said, the more rigid the better - on the other hand, you need good accuracy for a rigid setup or it will bind. Raldrich's method works well but would require you to make changes that would be easier done when you've got the machine working.

Do not do anything to the bearings if you can avoid it - as Dale says, make sure the rails are smooth. If you have a spare bearing you can run it up and down the rails with some polishing compound on it but do not use the bearing afterwards and make sure the rails are well cleaned after.

The basic principle is to constrain the parts as little as possible, any misalignment will then show itself as movement of the unconstrained elements rather than bending of the parts.

The side with two bearings on it is the master and the main thing to watch with this is that the bearings are in line. I would be tempted to epoxy in the bearings for this side with just the rail in place to make sure they are in line.

You can then mount the rails and the carriage - measure the positions and tighten the master side rail down in the measured position.

You now need to align the other side to be parallel to the master (measuring is not accurate enough for this) - leave one end fairly loose (enough for a millimetre of movement) and tighten the other down fairly firmly. Slide the carriage up and down - you will see the end of the rail on the loose side move side to side - adjust the tight end position until the loose end of the rail does not move side to side as you move the carriage (you may also need to move the loose end side to side to make sure it is in the centre of the movement of the rail). You can now put the carriage in the centre of it's travel and epoxy the bearing in place.

When it has set, recheck the adjustment procedure and then tighten down the loose end - as you do so, slide the carriage up and down and adjust its position so it does not bind.

Happy aligning :-)
Mark

I ended up doing a quick test on silicone and came up with a bond line .030" thick and a bond area of .125 sq in, it holds up to 5.8 lb shear max until it let go of the wood. it deflects .030" at 2lb shear, and will deflect up to 10deg at very small torques (less than .5 inlb). the silicone easily peeled off the roughened ABS after the wood released. obviousely poor adhesion to both surfaces and way too flexible, even if the bond area is increased.
Neil in youtube repraplogphase made a comment about more flexible adhesive than epoxy for bushing bonding, so I think that I will go with a polyurethane (mildly flexible) gorilla glue. The polyurethane should give a strong bond and allow some small flexation to prevent binding on the rails (at least that's my story until proven otherwise).

Edited 1 time(s). Last edit at 10/08/2012 05:11PM by rrr7.

Well, I did not like the warnings about sensitization on the gorilla glue label, so I went with good old JBWeld (epoxy) in stead and I am glad that I did. I don't think that my concern about binding was as critical as I had imagined with the variations in straightness given the small diameter of rod (5/16) and flexibility of both rod and the prusa frame. Keep in mind I am trained in machining, so I am critical of tolerances and straightness in mechanisms. In my experience having done it now, the mendel prusa design seems pretty tolerant of small build variations. Here are some details of what I did and found:
I got the rails inline with what looked like less than .005 inch deviations along the lengths checked to a flat surface (build plate) on top of the rods watching for light leaking through between the rods and the plate. Rotating the plate and rechecking to verify it was the rods which were deviant and not the plate. I rotated the rods to get the best straightness towards the top since that will affect the running flatness of the print bed. If there are larger variations in distance between the bars, the lower platform stiffness keeps the bushings fixed and forces the rails to deflect to stay in-line. It does not require adhesive flexibility to accomodate this. I used 1/16 inch dia welding rod as my rail jig (as well as the other jigs to check frame distances) and used this along the length between the rods. I did end up shimming one rod to the clamp using 6 thicknesses of printer paper inside the bottom of the clamp to raise that rod end since the plate showed that end low consistently. In the end, the JBWeld holds sanded ABS well and the platform slides with marginally little force. I can hear and feel slight variations (9 high-spots) in the sliding force (roughly measured with a spring kitchen scale to 100-300g) along the length with the newly installed unlubricated (Mcmaster p/n: 9368T115 Graphite SAE 841 Bronze Sleeve Bearing) against the hardware store sourced rods (not tightly toleranced rod) but the variation in force along the length is periodic in stead of in one place which means that the rod adjustment is as good as it can be. The force is higher and more variable than I would like, but the low cost and quickly available rod does work. In the end, I have no doubt that the machine will print but I will have to wait until I get to high speed printing to see if the cheap rods end up making any difference at all against the NEAM17 steppers.

Edited 1 time(s). Last edit at 10/14/2012 12:18AM by rrr7.

3 point suspension can help.

All measurements work from the Zero set for the X, Y and Z axis. It is on the side of the traveler
nearest the zero for the respective axis, to have the two bearings; and on the farthest side to allow
the single bearing room to float, preferably in the central position for the axis.

For the z-axis, I have let the rails on that farthest side to float on the x-plane.

I figured out why my y-axis has so much friction.
It is not due to alignment, it is due to tolerance buildup in the rods and bushings.
I tested all the bushings to a rod before beginning assembly, but failed to test all bushings to all rods. Some of the rods ae too tight to some of the bushings. before assembly that can be fixed by matching bushings to rods or by polishing either or both to get a good low friction slide. My mistake for assuming they were consistent. now that the Y-axis is assembled, I will just polish the rods in place.