Mendel Prusa X axis: slippage on the smooth rods, how are you dealing with it?

Hi again,

I am under the (theoretical) impression that, with the current Prusa Mendel design, for various reasons the holding force exerted by the motor and idler ends on the smooth X rods is insufficient, given the axial traction forces of the carriage.
This could cause slippage, precision issues and need frequent retensioning of the X belt.
Are you guys encountering any of those problems in practice? How are you dealing with them?

I adopted a preventive solution: shaft collars made from a drilled out and tapped M8 nut, pictured here [www.reprap.org]

But in the long term this would call for a redesign of the X ends holders. What options do you see?
I am not totally convinced by solutions calling for closing the rod channel ends, this would then need very precise cutting of the hardened rods, and limit adjustment possibilities.

I am thinking of opening the channel above the rods, replacing it with two separate roofpieces per end, and using four traversing bolts to secure the rods by pressure against the roofpiece. Precision placing could be ensured by transposing the V-channel from the top to the bottom of the rod channels. Maybe this is a bit "Old Mendel"?

Suggestions, experiences, inputs most welcome!

In my design, I have one x-end closed and one open.
This allows you to remove the rods if necessary (to remove an lm8uu carriage for instance).
For the open end, I have a 3mm hole from top to bottom that I can drop an M3 scre down. This prevents the possibility of the smooth rod coming out while printing.
On the closed end, I have a nut trap at the end so that I can screw an M3 screw into the channel so that it presses on the end of the smooth rod. This allows the spacing between the x ends to be held ridid and allows precise adjustment of that spacing without the need for precision cutting of the x smooth rods.

Hi Greg,

Very interesting design ! I'd love to take a look at it. is it in your github?

Side note: the cheap LM8UU bearings we are all using seem very prone to losing small steel balls each time we take them on and off the rods. So this is a manipulation I'd limit to a minimum... (hence the need to evolve the quick-change carriage)

Yes. In my git hub. It is only different in a minor way to the standard x ends.

Ok I have it. My clone of your repository was a bit old.

I like the design improvements and will definitely give 'em a try.
Also the smooth channel allowing for two LM8UU or a single LM8LUU is good.
What could (IMHO) be improved:
- The zip tie channels are single points of failure (I've had PrusaJr's design break clean just in that place while pushing the bearing in) and probably unnecessary. Faster, cleaner printing too.
- The "walls" for both the idler and the motor might be thickened by a mm or so. Rigid walls --> better parallelism --> belt wanders less on the motor's pulley.
- The "bridges"... racy and sportsy, yes... but thinking of all the newcomers with not-so-well-calibrated printers, oozing repStrap extruders, etc... Adrian took the choice of replacing bridges with walls. Which seems a sound choice. What about parametric walls one could turn on and off?

Connected thoughts:
There's one (slightly) variable point we do not fully control, that is the distance between the vertical rod axis and the vertical smooth bar axis. In my setup, between the X-end and the Z motor mount the difference in that distance is about 0.1 -0.2 mm, and the current Z motor mount doesn't not really allow for any adjustment.
This dos not impair the working of the axis, but is a significant cause of wobble at the lower end. I like to bolt my motors ;-)
Attached is an (untested) small mod of the z motor mount that should allow for variable positioning of the motor axis...

About Greg's new x-ends:

Here is - more or less- what I meant in the previous post, at least for the new idler end (no bridges and no zip tie holes). Still need to add control of those features by configuration variables.

The motor end is putting to test my rudimentary openscad skills.
I did away with the hourglass design of the motor mount, which, apart from aesthetic criteria, is mostly weakening the mount.
But this mount also, as Nophead observed in Thingiverse, anchors the X motor quite far from the center of gravity of the axis. I'd prefer a more centered motor mount (like the previous), but one with much stronger walls.

I agree that there is a problem with the lack of adjustment of the Z mount and rod clamp. I have used paper shims between the rod and the z motor mount to fix alignment problems. It would be nice to have an adjustment for this, but im not sure sliding motor mounts are the best way to do it. I prefer screw driven adjustments (like my belt tensioner), but that would require more vitamins.

As for the bridges, you should calibrate properly before printing and then they are not a problem. Your design will use much more plastic. You could have a skinny wall broadening out to the support to save on plastic but still not require proper calibration if you really want to avoid bridges.

Greg,

I agree that the crude adjustment provided by sliding motor mounts is not the most refined. Could do the job, though.
The "canonical" way to have fully adjustable motor centering would use three bolts in the circumference spaced 120 degrees apart. Presumably three M3x30 socket caps, plus three captive nuts, that isn't ready a lot of vitamins, and we are using the precision given by metal rather often.
Not unlike certain of your shaft mounts. I may try a stab at it when I have some free time (delay: months alas).

(rant mode)
The background problem is that we have a species well optimized for fast breeding (low in plastic and parts count), and in that it is rather successful (as compared with its ancestors). But this does not mean optimization for other important functions like precision, dampening of vibrations, etc. We now have a "generalist" deemed to spread and occupy the ecosystem (a rodent sort of), with limited ability for specialized tasks. Like everything else, that is a fluid and moving compromise. The agonizing question behind reprap and life in general is: is it all only about fast differential breeding?

Low plastic, thin walls: the compromise is visible, I get the X motor wall showing flexum as soon as I want to set adequate tension on the belt. Not a lot, it is still totally workable, but because of the lack of parallelism the belt wanders on the pulley. That's why I'd like to set a thicker X motor mount wall for my machine (or adopt a different design, Emmanuel is showing us very enticing X bars in vertical configuration).

Bridges finally: got good calibration and results, even with my monstruous franken-extruder, I'm obsessional-compulsive enough to spend the necessary time for that, but we must stay aware that lots of people and especially newcomers aren't mentally prepared to do all the stages of calibration at first. With all those marginally calibrated machines, I'd go the safe way, dead easy to print blocky pieces, leaving the beautiful design features for the arts & performance contests. Because user frustration isn't good for breeding (TM).
(/rant mode)

Edited 1 time(s). Last edit at 08/20/2011 02:56AM by Lanthan.

Lanthan Wrote:
-------------------------------------------------------
> (rant mode)
> The background problem is that we have a species
> well optimized for fast breeding (low in plastic
> and parts count), and in that it is rather
> successful (as compared with its ancestors). But
> this does not mean optimization for other
> important functions like precision, dampening of
> vibrations, etc. We now have a "generalist" deemed
> to spread and occupy the ecosystem (a rodent sort
> of), with limited ability for specialized tasks.
> Like everything else, that is a fluid and moving
> compromise. The agonizing question behind reprap
> and life in general is: is it all only about fast
> differential breeding?

Right now that is what it's all about. As these printers breed and saturate the market, building more complex and precise machines becomes easier.