[Johnsnd]

After much searching on the web and CnCZone I was unable to find the Pros and Cons of rigid/flex coupling.

Wikipedia says that "Flexible couplings are designed to transmit torque while permitting some radial and axial and angular misalignment". But what is meant by "misalignment". I would think that the rigid coupling would provide more accuracy. Is the error introduced by a flex coupling negligible or is it just accounted for in software.

The mechanics of CNC always get me stumped, seeing how I am a computer engineer, microelectronics are more my area.
Thanks in advance …

[Johnsnd]

Not sure what the protocol is on answering your own posts but here is the information i found on alignment of couplings. This diagram explains the need for flex couplings but I am still not sure what kind of error would be introduced.

[MrWild]

You do not want to use spiral cut couplers. They can wind/unwind a bit which makes for poor accuracy

[Al_The_Man]

Rigid couplings usually imply that alignment is perfect or ideal, which it often is not.
There are many types of couplings that are designed especially with CNC or accurate positioning in mind, some of the higher end ones are often out of practical reach for the
DIY'er.
One simple economic one is the Oldham coupling.
One I have used with success is the Lovejoy Uniflex, this is essentially two opposite wound springs, one inside the other.
http://www.lovejoy-inc.com/content.aspx?id=352
Al.

[Johnsnd]

Thanks Al_The_Man for your input. I was looking into the uniflex spring couplings (figure 1) and they seem quite expensive even for a DIY'er. The cheapest I found one was around $100 and that was the smallest one.

The Oldham couplings (figure 2) seem to be quite similar to the spider couplings (figure 3). I assume that they would be about the same in cost vs. accuracy. One advantage of these is that you can buy them in separate pieces.

My dilemma is that my lead screw is 1” diameter and my stepper motor is 6.35mm in diameter. So finding a coupling that will accurately attach to both is unlikely? I guess I need to find someone to turn down a 6’ft lead screw?

[Al_The_Man]

I am assuming you got a deal on the Ballscrew to be that large for what that size of motor must be?
If by 6ft leadscrew you mean acme etc, then it could be worse, when you consider the co-efficient of friction and inertia of a 1" dia screw especially if high lead.
Have you done any sizing for the stepper?
The Uniflex used to be a fairly decent price, but I noticed they have gone up drastically in the last little while.
Al.

[Johnsnd]

First and foremost for all who find this thread looking for the answer of rigid vs. flex coupling … choose FLEX couplings. The reason is for most DIY applications you are not able to perfectly align the stepper motor with the lead screw. This misalignment causes noise and reduces the life of the coupling. I think it might also cause whipping but I am not positive on that one. There is a negligible amount of error introduced in most flex couplings and some of the flex couplings are designed for cnc with zero-backlash in mind.

Actually al_the_man I didn’t get a deal on ball screws. I was thinking I could use a 400 oz/in stepper (Seen Here) to drive the x-axis of my 5'x3' cnc machine. Specifically using 1" acme threaded rod . But your subtle comments about friction got me thinking and it seems you are right, I need a different route.

So I have decided to go with timing belt system instead. So far I am still doing research on what exactly I need for that but it looks promising. I know that most threads have spoken of a “gear reduction” with a stepper motor having a small cog and a belt to the larger drive cog. The cost is about the same. I guess before I continue with this topic, should I start a new thread on belt systems?

Oh and PS, thanks a lot Al for taking the time to give me your input, you are “the man”

[Al_The_Man]

Thanks,
And another reason is snapped shafts.
http://www.cnczone.com/forums/showthread.php?t=81989
Al.

[bad bearings]

Unless you can bore and face your motor mount concentric to less than 0.001" TIR you ought to be using a flex coupling. Similar problems occur with ball rails and screws. I believe that most positioning applications bellows couplings work the best in terms of torsional rigidity and angular and offset misalignment. You should really analyze your application and figure out how much slop is acceptable. If you don't have the equipment to bore your motor mount concentric then I suspect the rest of your machine does not require a lot of accuracy. Try to use components that allow the most play possible--bushings for at least one set of rails. Self-aligning linear bearing pillow blocks etc...

There are many designs on this site that ignore the need for precision alignment of rigid rolling element components. When you're dealing with steel ball bearings of any kind you don't have much deflection to work with so you will have to make a large capital investment in precise machining to get things lined up correctly.

If you can't align components to a few tenths then you should increase your tolerances, use self aligning bearings, acme leadscrews or timing belts or better linear bushings. Some cheap lovejoy spider style couplings should do fine.

Good luck.

[RICHARD ZASTROW]

When using couplings I prefer Metal belows couplings (rated for torque max).

They allow for some misalignment and axial movement while keeping rotational accuracy, no lost rotary motion.

But, you're right, they ain't cheap.

Dick Z

[Hirudin]

When using couplings I prefer Metal belows couplings (rated for torque max).

They allow for some misalignment and axial movement while keeping rotational accuracy, no lost rotary motion.

But, you're right, they ain't cheap.

Dick Z

I've never heard of those before (not a surprise). I like the idea though! Where can we buy some?