Rigid vs Flexible Couplers for Trapezoidal Leadscrews?
-
Hi,
Having heard mixed views on the subject of using rigid vs flexible couplers for connecting trapezoidal leadscrews to steppers (and used only rigid ones myself so far), I'm curious about people's experience/thoughts in this context.
Basically, my question is would you recommend one type over the other and why?
Thanks!
SnowCrash -
I dislike the flex and spring of the flexible couplers. I preferred plum couplers. Semi rigid.
-
I did lots of testing on the railcore with various leadscrew couplers. (Two brands of spring style, two brands of rigid, two brands of the 3 part oldham couplers with the red plastic, and one brand of double disk diaphragm couplers).
I tested using my straightest leadscrews & some that were far less straight, in varying configurations. I used both printed and aluminum yokes (that tie our bed to the leadscrew & Z linear rails)
I found the following:
Spring Style - loudest. Cause Z artifacts on rapid z-hop, least consistent Z probes.
Rigid couplers - Worst transmitted artificats if you have a bent leadscrew. Also hard to align due to most of them having the grubs opposite each other.
Oldham - Best of the "common" couplers, quiet with the least artifacts transmitted. Cheap ones separate easily.
Double Disk - Overkill for common 3d printing use. Comparable print quality to the oldham couplers.I now recommend folks go with good oldham couplers. Zyltech is a good source for them.
Hope this helps!
Edit: Oh, my test print is a spiral vase cylinder, only 30mm in diamater, at least half the height of the bed. Couplers transmitting wobble will show up as a wave pattern with the same period as your leadscrews, in my case 4mm.
Second edit: I've been told misumi sells even nicer oldham couplers, but I have not spent the $ to try them. Part # MCO15-5-8
-
@kraegar Can you share a source for the Oldham style (plum) couplers?
-
For a source of the inexpensive ones Zyltech seems to sell the better quality: http://www.zyltech.com/flexible-plum-coupler-shaft-various-combinations-from-5mm-to-12-7mm/
Or the misumi ones I gave the part number for above.
-
It is hard to answer to that question without knowing the full geometry of your leadscrew context: it depends how it is mounted. Is there a longitudinal stop? Is it guided at the top, at the bottom? Are you using an anti-Z-wobble system under the bed?
There are many types of coupler, each one having different degrees of freedom (some have none, some have one, some have more). A lot of 3D printers setups are wrong, and users fixes totally insane (over-constraining things).
-
@snowcrash The approach I've taken is to think about the structure of the printer and internals of the stepper and then decide.
The stepper inside is a shaft with two ball bearings, one at one side, one at the other side of the stepper. The coupler prolongs the shaft to another shaft. So think about whether the angles of the shafts are near zero, then you can use a fixed coupler. If there are angles between the shafts, you want to protect the ball bearings of the stepper and use loose coupling.
Another thought is to unburden the stepper bearings by supporting the long shaft with an additional bearing. If you use spring style couplers, changing direction can change the difference between stepper and long shaft and the shaft can wave. The bearing, if fixed at the long shaft, will protect against this.
A next point to think of is backlash: if you change direction, is there any? Oldham seems good, but I bought a cheap one where the axes are crooked, so don't buy too cheap. ( i learned below that I bought a jaw coupling, not Oldham
)There is a fifth type of couplers, bellows coupling, but thery are even more expensive. They are said to have near zero backlash.
-
To avoid Z-banding it's important that the leadscrews are rigidly mounted in one end. Both ends is a bad idea, especially if the leadscrew or coupling isn't perfectly straight.
I have 3 printers:
- Prusa I2 style: Leadscrews hanging from a flexible coupling on the motor, no support in the bottom. -> Z-banding
- Prusa I3 style: Flexible coupling in the bottom to the motor and a bearing with set-screws on the top. -> no Z-banding
- Hypercube Evolution: Rigid couplings on the bottom by the motors, floating in normal bearing in the top. -> no Z-banding
-
@joergs5 said in Rigid vs Flexible Couplers for Trapezoidal Leadscrews?:
A next point to think of is backlash: if you change direction, is there any?
Same, with spring couplers, if you exceed maximum rated torque (under strong accelerations, for example), motor shaft and leadcrew may not remain in sync (maybe not really a problem for Z axis, but can be for X/Y).
-
@lhelge said in Rigid vs Flexible Couplers for Trapezoidal Leadscrews?:
To avoid Z-banding it's important that the leadscrews are rigidly mounted in one end. Both ends is a bad idea, especially if the leadscrew or coupling isn't perfectly straight.
Unless you have a floating coupler between the nut and the bed; this is my current config (as it was on my previous printer), and it work fine.
Look for (good) anti-Z-wobble, where the nut only push the bed up, but is free to move in XY plane. There is an anti-rotation system so the nut does not turn, but only translates:
https://www.thingiverse.com/thing:2646220
And an example of a bad anti-wobble system:
https://www.thingiverse.com/thing:2121017
There, I guess top, bottom and nut are all rigid, and as you said, this is not good at all.
-
@kraegar those zyltech ones aren't actually oldham couplings at all whereas the Misumi ones are
-
Ah, thanks - guess I was mis-using the term for those. They're the better of the cheaper conventional offerings, if you buy good ones.
Agreed on the points here about having enough constraint without overconstraining them. We mount the stepper rigidly on the bottom. We use a yoke to tie the leadscrew & linear rail together, and attach the bed. We do not constrain the top of the leadscrew in any way.
-
I'm using the flexible aluminum couplers with a 4mm ball bearing inside to eliminate any compression, but still allows bending.
-
I Have ordered 3 of the misumi Oldhams so will be interesting to see how good they are but I will also be useing thrust bearings between the bottom of the oldham and the face of the stepper motor so that there is no axial load going into the motor bearing as I feel that would lead to a fairly rapid failure.
Doug
-
I might need to pick up some of the Misumi oldhams to try and compare, too.
-
Did some digging and turned up this:
https://www.ruland.com/resources/technical-articles/technical-article-servo-coupling
Indeed, I was using the wrong term for the ones from Zyltech. I'll leave my error to avoid confusion. (also, helpful article)
-
I'm always amazed by the complexity behind even the most seemingly simple things
Thank you to everyone who contributed to this discussion!
I learned a lot (mainly about how much I don't know...), and took my time digesting the various advice and excellent leads (thanks for the article link, @kraegar, very helpful indeed!)
@fma, you're right of course about the difficulty of offering advice without knowing the specifics, and indeed I was initially inclined to provide a detailed description within my original post. However, I eventually decided against it precisely because I wanted to keep the discussion open-ended and wide as possible, rather than narrowing it down by focusing on a particular setup or application. I hope this makes sense.
@JoergS5, many thanks for the methodological breakdown of the subject-matter into logical steps. Very helpful indeed in tackling this!
@Phaedrux, @kraegar, @Stephen6309 & @LHelge, thanks for sharing your practical knowledge and research conclusions.
@Dougal1957, @kraegar & anyone else who might engage in testing different solutions, please do come back and share your insights and discoveries. I - and I think others too - would certainly be interested to hear about your experience.
As for myself, from what I've read so far, the 'Zero Backlash Jaw Couplings' look like they could offer the best balance between accuracy of torque transmission and allowance for misalignment. These cuties aren't cheap, though, so I'm currently searching for the best deal on quality parts. I'll report back once I have them on hand and assembled.
Thanks again everyone!
SnowCrash -
@snowcrash Thank you for your nice words!
I started building a printer from scratch about 6 months ago and still learning something new about the physics of the planned printer every day. Besides the 3D printer forums, I am more and more in the CNC forums also, because they are used to take into account effects of 1/100 mm and below. Understanding the CNC issues help improving 3D printers also.
One example: thermal expansion is a widespred issue in the CNC area (and how to solve it), because mixing aluminium with steel is used often. In 3D printer construction it is often neglected, but without a reason, Because you mix materials here also, and mixing and at the same time fixing will cause bending or breaking. One example is the heat bed.
I wish you the best for improving your printer and much fun.
-
Thanks, @JoergS5, a CNC is definitely on my list too
Which CNC forum/s would you most recommend? (I promised myself I won't go into the CNC project before the printer is done, but no harm in just looking, right?)
-
@snowcrash don't do it you will get sucked in lol
-
@snowcrash said in Rigid vs Flexible Couplers for Trapezoidal Leadscrews?:
As for myself, from what I've read so far, the 'Zero Backlash Jaw Couplings' look like they could offer the best balance between accuracy of torque transmission and allowance for misalignment. These cuties aren't cheap, though, so I'm currently searching for the best deal on quality parts. I'll report back once I have them on hand and assembled.
Please do so, because I would also be interested in this. I just recently rebuilt my Z axis to no longer use this kind of coupling at all but still I find this very interesting and I always keep my options to revert/change anything.
-
Looking around for these zero backlash jaw couplings I came across ones where the spider has 4 segments about equal in size (each) to one of the jaws and others where the spider has 6 segments that are considerably smaller (maybe half) than the jaws.
Has anyone any opinion on these two types?
-
@snowcrash said in Rigid vs Flexible Couplers for Trapezoidal Leadscrews?:
Thanks, @JoergS5, a CNC is definitely on my list too
Which CNC forum/s would you most recommend? (I promised myself I won't go into the CNC project before the printer is done, but no harm in just looking, right?)
The largest seems to be CNCzone.com/forums, one of the technical best I found is practicalmachinist.com/vb
Sometimes I find interesting threads on reddit.com also.
To get ideas I check openbuilds.com
-
Hi all,
@dougal1957 said in Rigid vs Flexible Couplers for Trapezoidal Leadscrews?:
@snowcrash don't do it you will get sucked in lol
Too late...
Thanks for the links, @JoergS5!
Some really nice machines out there
One of the biggest challenges I'm facing atm as I'm working on the prototype is that I often need small custom metal parts and it's incredibly difficult to find anyone to make small quantities at a reasonable price.
So if there's anyone who's based in the UK and already has a CNC and would be willing to help out (I'd of course cover all the costs of material and fabrication) - please ping me!
Back to the present topic, my search for suitable couplings led me to realize that I haven't provided enough support for the leadscrews to begin with. Apparently, without sufficient lead-screw support, there's little even high-quality couplings can do, so tackling this issue has become the top priority.
Once it's sorted (I hope
) , I'll proceed to the matter of couplings and report back.@wilriker said in Rigid vs Flexible Couplers for Trapezoidal Leadscrews?:
Looking around for these zero backlash jaw couplings I came across ones where the spider has 4 segments about equal in size (each) to one of the jaws and others where the spider has 6 segments that are considerably smaller (maybe half) than the jaws.
Has anyone any opinion on these two types?
Sorry, @wilriker, I only came across spiders with 4 segments in my research and unfortunately don't know enough to offer any help on this.
-
I Received my Oldham couplings from Misumi and jeez they are nice and they are also very compact (About half the size of the std couplings that we normally use) I also realise that using them directly could transfer quite a lot of side thrust into the stepper motor bearing's so I intend to fit some Thrust bearings between the coupling and the stepper body to account for the side thrust.
Doug
