[reallyhuman]

I realize this is an aging thread, but felt compelled to drop a bit of info here that I found interesting.

An essential component of jaw couplings, flexible elastomeric spiders transmit torque and accept misalignment | Mechanical Drives content from Machine Design

In the last paragraph of the article, a 'product manager' from Lovejoy, Inc. states:

Jaw couplings are not suitable for most engine-driven or frequent start-stop-reversing applications because of backlash (hub rotation allowed by the spacing between jaws and spider legs). Also, this backlash makes most jaw couplings unsuitable for positive-displacement (pump) and precision motion control applications.

I find this intriguing, due to the fact that hub/spider jaw couplings (Oldham, Lovejoy etc..) are generally accepted as a suitable coupling solution in (hobby) CNC applications. What I take from this statement is the manufacturers wouldn't recommend their use in that particular application?

With that said, I am in the same boat as diecutter. I can live with the maybe .0015" - .003" backlash I may encounter as the spider wears. I get replacements from mcmaster for like $2.50.

Just sayin'.

[wizard]

Yeah this is an old thread!

One problem I found after reading a good portion of it is a bit of confusion with respect to product names. For one an Oldham coupling design is not the same thing at all as is a LoveJoy spidered coupling. A minutes search should highlight the difference to everyone.

I realize this is an aging thread, but felt compelled to drop a bit of info here that I found interesting.

An essential component of jaw couplings, flexible elastomeric spiders transmit torque and accept misalignment | Mechanical Drives content from Machine Design

In the last paragraph of the article, a 'product manager' from Lovejoy, Inc. states:


I find this intriguing, due to the fact that hub/spider jaw couplings (Oldham, Lovejoy etc..) are generally accepted as a suitable coupling solution in (hobby) CNC applications. What I take from this statement is the manufacturers wouldn't recommend their use in that particular application?

There is a considerable difference between an Oldham and a LoveJoy in CNC or motion control work! A LoveJoy coupling, in the most common form is not suitable for motion control application where accuracy is needed. That being said there are couplings marketed that look similar to LoveJoy style couplings that are marketed for motion control applications. Generally these couplings have a different design to the spider and jaws that contain. Oldham couplings can be great for motion control when a coupling is found that can handle the tongue involved. To be honest though I often prefer the Helical style couplings.

With that said, I am in the same boat as diecutter. I can live with the maybe .0015" - .003" backlash I may encounter as the spider wears. I get replacements from mcmaster for like $2.50.

Just sayin'.

I'm going to suggest that most can't live with that backlash even if they think they can. Beyond that I've generally found that LoveJoy style couplings have more freedom than that. Plus this spiders can wear real fast in some applications. Of course there are all sorts of "spinner" materials available but in order for the coupling to take up mis alignment you need to have some clearances in the spinner material of you need to resort to precise alignment of the coupling.

[pippin88]

Disc couplings are quite good. Harder to find but from my reading offer the best performance. Get double disc ones - they allow angular and side ways misalignment. Should have the lowest backlash (if any).

[louieatienza]

Lovejoy makes many different types of spider couplings, with different tolerances, and the spiders come in different hardnesses depending on the application. The L style most commonly seen and used have some noticeable play and are not designed for positioning, rather just for transmission of rotary movement. A spider coupling can be beneficial in axes under stepper control, as the spider acts as a decoupler between screw and motor, which can help in a system that's prone to resonance.

[Devastator]

I use belts.

[Eldon_Joh]

I was a bit supprised to read that the helical couplings are known to fail (recently in another thread), but upon finding out they were made from aluminum, it was of course no supprise.

you guys know that aluminum does not have a fatigue limit.. right?

[wizard]

I was a bit supprised to read that the helical couplings are known to fail (recently in another thread), but upon finding out they were made from aluminum, it was of course no supprise.

They can and are sold and made from a variety of materials. Stainless is a common material.

you guys know that aluminum does not have a fatigue limit.. right?

As for failure I've seen all sorts of couplings fail over the years. Sometimes from poor installation, sometimes from things going horribly wrong (hydraulic pump disintegration) and some times from old age. There is still a real need to properly size what ever coupling you decide to use. That is if you use a coupling at all. As someone above mentioned, sometimes a belt drive is a smart choice.

[Grunblau]

I know this is an old thread, but while it is resurrected, I thought I'd throw in my $.02 since I have experience with all three on various builds.

First is the LoveJoy coupler. I used these on the MDF machine build and had no issues. Everything was tight and I doubt I would have been able to notice any backlash issues. Good value for a good coupler with no noticeable backlash.

Next was the helical coupler. In my build thread for my machine, you will notice that I replaced these immediately. This was because I had observable backlash when the machine was changing direction. May be not a big issue on most hobby cnc's where the acceleration is low. On the Platform I could watch them spool up like a spring when switching directions especially on the Z-Axis. At that time, I switched to high quality Lovejoy style couplers with hard plastic spiders for the Prototype machine. These were expensive from McMaster and remain my favorite connector as they allow for little slop for alignment as I don't need or want any.

The current machines now uses the Oldham style couplers from CNC Routerparts. These allow for misalignment but work as well for power transfer as the Lovejoys. Just makes it a bit more difficult to align the axis when the coupler allows for misalignment, where a spider coupler locks everything inline.

Personal opinion,

Helical- Never again for me since I could physically turn the axis lead screw 5-10 degrees by hand with the steppers locked.

Lovejoys- Good cheap option for hobby CNC... really like the version I got from McMaster, but they were like $40 a piece I think.

Oldham- CNC Routerparts has a very nice and inexpensive version of this and they work great! They allow for misalignment, but I would rather they didn't.

Hope this helps!

Brian

[UA_Iron]

I was a bit supprised to read that the helical couplings are known to fail (recently in another thread), but upon finding out they were made from aluminum, it was of course no supprise.

you guys know that aluminum does not have a fatigue limit.. right?

Aluminum does have a fatigue limit, every material represented on an S-N curve has a fatigue limit - it's how the line gets drawn that represents that limit.

I know what you're trying to say - steel has an asymptotic fatigue limit at a given stress where no number of cycles will cause it to fail, while aluminum will continue to decrease in strength as the number of cycles increases until it fails. Essentially we're talking about the 'toughness' of a material. 7000 series aluminums are probably your best choice for fatigue resistance, 7075 and 7050 being very good choices. Polymers and composites have fairly poor fatigue properties as well - this is why steel dominates in the structural engineering field.



Back to the issue of couplers -

Oldham couplers: I will only use these under certain circumstances; those circumstances being when extremely small packaging/clearances are needed, and when I cannot guarantee shaft concentrities due to some overconstraint. These actually do a great job at dealing with parallel offsets, angular offsets and axial misalignments if the disc portion is of sufficient lubricity (ie Delrin/Acetal) - but they are awful if you are trying to line up two shafts concentric before bolting something else down (like an encoder). In fact, aligning shafts with an oldham coupler is painful to do; I have used an indicator to get the oldham couplings concentric before bolting down an assembly.

Attachment 267994

Helical couplings: Really should be used forangular and some parallel misalignments only. Introducing axial misalignments will stress that right angle where the slit is relieved into the shaft coupling ends, right angles are bad for fatigue.
Lovejoy couplings: Can accept all 3 misalignments to a degree, and maybe the best for axial misalignment (speculation). The typical soft polymer spider piece allows for some backlash, which is probably fine in most Hobby CNC machines. A huge benefit to the lovejoy and the oldham is the coupled shafts can be taken apart without having to remove any coupling components. The backlash is not that great for control systems driven by external linear encoders where settling time and instabilities in your control system will be caused by the backlash.
Disc Couplings: I haven't used these in any recent applications but they have always appeared to be a great application to motion control. Apparently they are somewhat delicate compared to the other designs, but have much better torsional rigidity. We replaced a disc coupling with a lovejoy coupling with a much higher load rating for a lesser price - our system does end point correction and can tolerate some backlash.
Bellows Coupling: Probably the cream of the crop for any motion control application (check out ebay, you can score on some cheap ones sometimes).

Hard couplings: The only way I would use a hard coupling is if I used an indicator and probed each shaft and made sure they were dead nuts to each other.

I've used Ruland manufacturing for couplings before - in fact, they sent me a bunch of samples for free (to my place of work). I believe McMaster sells the same couplings:
Shaft Collar, Rigid Coupling & Flexible Shaft Coupling Manufacturer

And here's an article stating everything I just said above but better:
What to look for in a Servo Coupling | Ruland

[JerryBurks]

Just to throw in an out-of-the-box solution that I have used on my machine for some 3 years now:
- the ball screw has a pre-loaded double angular contact bearing only on the far end of the screw, i.e. opposite the stepper motor. This takes up all the axial screw loads.
- there is no separate bearing on the motor side
- the stepper motor is connected with a rigid coupler (axial alignment on my machine is pretty good but not perfect).
- The stepper motor is not mounted rigid but on a slightly flexible aluminum plate. This plate allows the motor to flex axially (compensate screw/machine temperature expansion) and some torsion flex perpendicular to the shaft (compensate the coupler inaccuracy) but is rigid rotationally to take up the motor torque. Practically it is like a disk coupling but not rotating.

Basically it means to relocate the flexible joint from the rotating screw coupling to the stationary motor mount. Since my couplers have a tiny angular error of maybe 1/2 degree, you can actually see the steppers wobble when rotating but it does not matter for the accuracy. One could suspect the aluminum plate might fatigue eventually but it is quite large and the motion is minimal. So far so good.....

[louieatienza]

Aluminum does have a fatigue limit, every material represented on an S-N curve has a fatigue limit - it's how the line gets drawn that represents that limit.

I know what you're trying to say - steel has an asymptotic fatigue limit at a given stress where no number of cycles will cause it to fail, while aluminum will continue to decrease in strength as the number of cycles increases until it fails. Essentially we're talking about the 'toughness' of a material. 7000 series aluminums are probably your best choice for fatigue resistance, 7075 and 7050 being very good choices. Polymers and composites have fairly poor fatigue properties as well - this is why steel dominates in the structural engineering field.



Back to the issue of couplers -

Oldham couplers: I will only use these under certain circumstances; those circumstances being when extremely small packaging/clearances are needed, and when I cannot guarantee shaft concentrities due to some overconstraint. These actually do a great job at dealing with parallel offsets, angular offsets and axial misalignments if the disc portion is of sufficient lubricity (ie Delrin/Acetal) - but they are awful if you are trying to line up two shafts concentric before bolting something else down (like an encoder). In fact, aligning shafts with an oldham coupler is painful to do; I have used an indicator to get the oldham couplings concentric before bolting down an assembly.

Attachment 267994

Helical couplings: Really should be used forangular and some parallel misalignments only. Introducing axial misalignments will stress that right angle where the slit is relieved into the shaft coupling ends, right angles are bad for fatigue.
Lovejoy couplings: Can accept all 3 misalignments to a degree, and maybe the best for axial misalignment (speculation). The typical soft polymer spider piece allows for some backlash, which is probably fine in most Hobby CNC machines. A huge benefit to the lovejoy and the oldham is the coupled shafts can be taken apart without having to remove any coupling components. The backlash is not that great for control systems driven by external linear encoders where settling time and instabilities in your control system will be caused by the backlash.
Disc Couplings: I haven't used these in any recent applications but they have always appeared to be a great application to motion control. Apparently they are somewhat delicate compared to the other designs, but have much better torsional rigidity. We replaced a disc coupling with a lovejoy coupling with a much higher load rating for a lesser price - our system does end point correction and can tolerate some backlash.
Bellows Coupling: Probably the cream of the crop for any motion control application (check out ebay, you can score on some cheap ones sometimes).

Hard couplings: The only way I would use a hard coupling is if I used an indicator and probed each shaft and made sure they were dead nuts to each other.

I've used Ruland manufacturing for couplings before - in fact, they sent me a bunch of samples for free (to my place of work). I believe McMaster sells the same couplings:
Shaft Collar, Rigid Coupling & Flexible Shaft Coupling Manufacturer

And here's an article stating everything I just said above but better:
What to look for in a Servo Coupling | Ruland

I'd like to add that there are different rated spiders that are used for different situations. I believe Ruland color codes theirs. I've uaed their couplers on a few projects, and yes I bought from McMaster.

Also, Lovejoy makes curved jaw spider couplings as well. Their most common one is the L style, which can have some noticeable play in them. I noted the backlash problem in a YouTube video, and actually had the VP of Lovejoy International respond (and subscribe to my channel!)

[Somun]

Just to throw in an out-of-the-box solution that I have used on my machine for some 3 years now:
- the ball screw has a pre-loaded double angular contact bearing only on the far end of the screw, i.e. opposite the stepper motor. This takes up all the axial screw loads.
- there is no separate bearing on the motor side
- the stepper motor is connected with a rigid coupler (axial alignment on my machine is pretty good but not perfect).
- The stepper motor is not mounted rigid but on a slightly flexible aluminum plate. This plate allows the motor to flex axially (compensate screw/machine temperature expansion) and some torsion flex perpendicular to the shaft (compensate the coupler inaccuracy) but is rigid rotationally to take up the motor torque. Practically it is like a disk coupling but not rotating.

Basically it means to relocate the flexible joint from the rotating screw coupling to the stationary motor mount. Since my couplers have a tiny angular error of maybe 1/2 degree, you can actually see the steppers wobble when rotating but it does not matter for the accuracy. One could suspect the aluminum plate might fatigue eventually but it is quite large and the motion is minimal. So far so good.....

This is very interesting. Could you share more details? Is this for a mill or a router?
Do you have some pictures for the stepper motor mount?

[Eldon_Joh]

Just to throw in an out-of-the-box solution that I have used on my machine for some 3 years now:
- the ball screw has a pre-loaded double angular contact bearing only on the far end of the screw, i.e. opposite the stepper motor. This takes up all the axial screw loads.
- there is no separate bearing on the motor side
- the stepper motor is connected with a rigid coupler (axial alignment on my machine is pretty good but not perfect).

Basically it means to relocate the flexible joint from the rotating screw coupling to the stationary motor mount. Since my couplers have a tiny angular error of maybe 1/2 degree, you can actually see the steppers wobble when rotating but it does not matter for the accuracy.

My steppers have wave washers on both sides internally, which means the rotor core of the stepper can travel axially perhaps as much as 1mm in both directions. this is plenty for thermal expansion.

i'm mentioning this so anyone reading it can simplify their design.. however you must verify that you have wave washers on both sides of the stepper.

alternatively you could replace the bearings in the stepper with AC bearings and bypass the second bearing block completely. this might actually be worth a try..