Synchromesh Cable Drive

[Ryan Carlyle]

This is a thread fork specifically for synchromesh cable. I ordered some synchromesh cable from SDP/SI and I'm going to try to build it into CoreXYZ. If it doesn't work there, I have some other projects (like CoreXYAB) that it would be well-suited for. 

So far, the only mainstream bot I've seen using it is the Bukobot. But people say it works well. (WhosaWhatsis, you feel like chiming in?)

Drive pulley: http://www.thingiverse.com/thing:16088

Smooth printed Idler pulley: http://www.thingiverse.com/thing:15923

Double-bearing idler pulley: http://www.thingiverse.com/thing:448977 (how well does this work?)

Cable ensioner: http://bukobot.com/_media/image007.jpg?w=600&tok=6293bf

My main concern is the pitch accuracy. SDP/SI reports a pretty significant error bar on the pitch because the spiral "teeth" accumulate error per wrap, unlike a traditional molded belt. My thought is that consistent pitch error (precise but not accurate) can easily be tuned out via steps/mm. But inconsistent pitch makes life difficult. Specifically for CoreXYZ, the four over-constrained motors require precise synchronizing to avoid racking the gantry. So the four drive cables need to have the same pitch. It doesn't matter what the pitch is, but it needs to be constant through all four cables.

Another aspect I want to look at is torsional pitch change. If you twist the cable, it should slightly alter the pitch of the spiral wrap. It LOOKS like the inner core is wrapped opposite the outer spiral so the cable is torsionally-balanced. But I'll have to see what it does when it's tensioned up.

If it does have good resistance to torsion, as the internet suggests, that means eye-lug crimp end terminations may require pre-alignment to avoid twisting. I don't think that'll be a problem but I haven't 100% decided how I want to terminate the cable yet. 

[Whosa whatsis]

The synchromesh works quite well. I think the most impressive characteristic is how quiet (you don't realize belts are making noise until you get rid of them) and smooth it is, like replacing a spur gear with a helical one. I find that the best idler is either two 625 bearings with a washer between them to make a gap for the cable to rest in, or a V-grooved bearing like this one: http://www.robotdigg.com/product/110/V625ZZ-V-Groove-Ball-Bearing. I was very skeptical of the synchromesh at first and put it through extensive testing, and was not able to measure any accumulating pitch error or backlash with the pieces I had. The cable resists twisting without coiling pretty strongly. For purely inertial loads, it's great, though I might not trust it for precision applications with a gravitational load (Z axis). I did use it for the lift mechanism on my 3d scanner, but it takes a lot of tension to keep it from skipping in that configuration, and of course it's not a precision application. I'm thinking about replacing it with a belt, though it would be much harder to fit the type of large-toothed belt that I really want to use.

[Whosa whatsis]

For the record, I don't particularly trust belts for Z either.

[Ryan Carlyle]

The 625 V bearings work? That's good to know. I built all my spectra routing around 624 V bearings (smaller) and ordered the smallest synchromesh cable (smaller than Bukobot's), so maybe I'll be able to get decent results out of the existing bearings and not need to immediately redesign everything.

[whosawhatsis]

They work, though with a complicated belt path with lots of idlers spread around, it might be tricky to keep enough tension on them to keep them in place on V grooves while installing them. Once they're in place and tensioned, they'll stay though.

On Monday, October 20, 2014 at 7:41 AM, Ryan Carlyle wrote:

The 625 V bearings work? That's good to know. I built all my spectra routing around 624 V bearings (smaller) and ordered the smallest synchromesh cable (smaller than Bukobot's), so maybe I'll be able to get decent results out of the existing bearings and not need to immediately redesign everything.

--

You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.

To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/cf233bb2-4c81-4e82-9c78-562c6a4bbc7e%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

[Ryan Carlyle]

Has to be easier than running spectra line through my pulley setup. That was a PAIN. At least synchromesh cable has a little stiffness and width to it, so it won't fall completely into every one-washer-wide crack.

On Monday, October 20, 2014 1:26:01 PM UTC-5, Whosa whatsis wrote:

They work, though with a complicated belt path with lots of idlers spread around, it might be tricky to keep enough tension on them to keep them in place on V grooves while installing them. Once they're in place and tensioned, they'll stay though.

On Monday, October 20, 2014 at 7:41 AM, Ryan Carlyle wrote:

The 625 V bearings work? That's good to know. I built all my spectra routing around 624 V bearings (smaller) and ordered the smallest synchromesh cable (smaller than Bukobot's), so maybe I'll be able to get decent results out of the existing bearings and not need to immediately redesign everything.

--

You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.

To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[Whosa whatsis]

Yeah, probably easier than that. For a complicated belt path, you might consider putting a guard loop around the pulleys that you need to thread the cable through and will prevent the cable from falling off the end entirely. Getting it from the flat of the bearings into the groove is easier than getting it onto the pulley in the first place and getting it to stay there. 

On Monday, October 20, 2014 12:29:14 PM UTC-7, Ryan Carlyle wrote:

Has to be easier than running spectra line through my pulley setup. That was a PAIN. At least synchromesh cable has a little stiffness and width to it, so it won't fall completely into every one-washer-wide crack.

On Monday, October 20, 2014 1:26:01 PM UTC-5, Whosa whatsis wrote:

They work, though with a complicated belt path with lots of idlers spread around, it might be tricky to keep enough tension on them to keep them in place on V grooves while installing them. Once they're in place and tensioned, they'll stay though.

On Monday, October 20, 2014 at 7:41 AM, Ryan Carlyle wrote:

The 625 V bearings work? That's good to know. I built all my spectra routing around 624 V bearings (smaller) and ordered the smallest synchromesh cable (smaller than Bukobot's), so maybe I'll be able to get decent results out of the existing bearings and not need to immediately redesign everything.

--

You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.

To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+...@googlegroups.com.

[Ryan Carlyle]

Yeah, most of my pulleys are enclosed. Ok, maybe half. The static frame routing pulleys are all open, but that will change if I end up needing to switch to bigger pulleys for bend radius improvement.

[Ryan Carlyle]

Hmm. Why do people consider this stuff anti-backlash? I haven't put my hands on the SDP/SI pulleys yet, but the Thingiverse knock-offs look like they would have backlash from the spiral shifting in the pocket.

[whosawhatsis]

I've never tried printed pulleys with it, but the SDP/SI ones don't have any room for it to shift. If I was going to print pulleys, I make sure that the slots for the spiral to sit in were thinner than that cable, so that it rids on top of them and only makes contact on the rounded edges of the outer strand.

On Tuesday, October 21, 2014 at 4:32 PM, Ryan Carlyle wrote:

Hmm. Why do people consider this stuff anti-backlash? I haven't put my hands on the SDP/SI pulleys yet, but the Thingiverse knock-offs look like they would have backlash from the spiral shifting in the pocket.

--

You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.

To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/b0125e10-b754-407f-a919-6714aadec7c8%40googlegroups.com.

[Chris P]

Unless you have a 3D printer that can print perfect circles, printing the drive or idler pulleys will introduce runout in the system since the running surface of the pulley is effectively non-concentric with the axis of rotation.  I would stick with manufactured parts in the drive system for this reason.

That said, SDP has sent me timing belt pulleys with the outer sleeve not pressed on the center bore squarely, so even here there are no guarantees.  THAT was fun to chase down...

[whosawhatsis]

While I suppose that's possible, we've had thousands of synchromesh pulleys ship in and out of Deezmaker, and never encountered that issue. I have seen two or three that they somehow forgot to cut teeth into, though, just a smooth groove.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/e21afe8c-d691-4c84-addf-ab720f1f2d2e%40googlegroups.com.

[Ryan Carlyle]

I did buy enough proper SDP/SI pulleys for my CoreXYZ drive. But for some of the other applications I'm looking at, a smaller pulley diameter would be advantageous and SDP/SI doesn't have a lot of diameter options. (For proof-of-concept prototyping, I'm more worried about envelope constraints than fatigue life with respect to minimum bend radius.)

I'll get a better feel for it when I put my hands on the stuff, but I'm wondering about the best idler pulley profile for smooth motion. The obvious answer seems to be treating the stuff like it's a round belt / o-ring to spread contact area over a longer arc length of the spiral wrap line. Using V-bearings intuitively seems like it's going to cause vibration as the effective line diameter changes.

On Tuesday, October 21, 2014 9:51:57 PM UTC-5, Whosa whatsis wrote:

While I suppose that's possible, we've had thousands of synchromesh pulleys ship in and out of Deezmaker, and never encountered that issue. I have seen two or three that they somehow forgot to cut teeth into, though, just a smooth groove.

On Tuesday, October 21, 2014 at 6:13 PM, Chris P wrote:

Unless you have a 3D printer that can print perfect circles, printing the drive or idler pulleys will introduce runout in the system since the running surface of the pulley is effectively non-concentric with the axis of rotation.  I would stick with manufactured parts in the drive system for this reason.

That said, SDP has sent me timing belt pulleys with the outer sleeve not pressed on the center bore squarely, so even here there are no guarantees.  THAT was fun to chase down...

On Tuesday, October 21, 2014 7:58:36 PM UTC-4, Whosa whatsis wrote:

I've never tried printed pulleys with it, but the SDP/SI ones don't have any room for it to shift. If I was going to print pulleys, I make sure that the slots for the spiral to sit in were thinner than that cable, so that it rids on top of them and only makes contact on the rounded edges of the outer strand.

On Tuesday, October 21, 2014 at 4:32 PM, Ryan Carlyle wrote:

Hmm. Why do people consider this stuff anti-backlash? I haven't put my hands on the SDP/SI pulleys yet, but the Thingiverse knock-offs look like they would have backlash from the spiral shifting in the pocket.

--

You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.

To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/b0125e10-b754-407f-a919-6714aadec7c8%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[whosawhatsis]

Are you talking about V wheels, like what you run on extrusions like Makerslide? Those could have that effect. The bearings I've used have much more shallow V grooves. We've also been experimenting with some that have a nylon wheel with a rounded groove (I believe they're intended for sliding doors or something). As long as each idler bearing has at least 2 or 3 teeth of the spiral wrap in contact with it at a time, the effect should be minimal, though it would get worse if you use a small pulley, a large pitch, and a shallow angle (our idler pulleys always turn the cable 180 degrees).

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/04fefa16-608d-4ff5-a07e-5dc5ce0aaeba%40googlegroups.com.

[Ryan Carlyle]

This is what I have right now for spectra drive, 624zz V bearings:

My thought is that if the bearing is too small, the line center will wobble side to side and up and down as the spiral wrap contact points change. Probably won't matter a huge amount to print precision unless you hit a resonant frequency. 

[whosawhatsis]

Yeah, besides the double bearings, we've only tried these two kinds:

Very shallow v-groove: http://www.robotdigg.com/product/110/V625ZZ-V-Groove-Ball-Bearing

U-groove: http://www.aliexpress.com/item/10pcs-5mm-U-Groove-Plastic-pulley-bearings-0-197-inch-ugroove-0521-5-21-7-bearing/1262881744.html

The U-groove might be better for what you're doing. For our application, I'm not crazy about it because it increases the diameter.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/896b2a2c-2120-4399-b1c1-76e17867366f%40googlegroups.com.

[Ryan Carlyle]

I'm thinking a shallow U-groove is optimal, but that probably means printing it. The V-groove bearings would have to end up being pretty bad for it to be better to go with a printed option that's slightly out of round.

I think my SDP/SI package is being delivered today -- time to stop talking about it and go see if it's arrived  :-)

[Ryan Carlyle]

Oh, this stuff is just lovely. I knew it was 1.6mm diameter but in my head that didn't properly translate to "really freakin' small." It's only 3x larger diameter than my spectra line. (And 1/3rd the nominal breaking strength, but no biggie.) My 624zz V bearings are going to work great, at least to start.

[Ryan Carlyle]

Synchromesh cable is just awesome to work with. The little bit of stiffness makes it much easier to handle than spectra. I think this is going to work really well, once I get some time to string everything up and get the tensions balanced properly. 

Definitely need to design some new "belt clips" for adjustable cable tensioning though. And these 624 V bearings seem to work ok for now, but in the long run I do think I'm going to want a larger bearing.

Here's how one line looks, strung up with slight tension:

[whosawhatsis]

We've experimented with several types of tensioners, and by far the best is the sliding 180-degree idlers that I designed for the Bukito. It uses a single screw (plus a "finger" to keep it aligned) on the slot in the aluminum extrusion, with a separate screw to hold the idler pulley (though if you want the synchromesh adjacent to the extrusion instead of around it, it might work with just one screw). Just slide the idler piece on the slot to pull it tight, then tighten the mounting screw to lock it in place. I definitely recommend using real T nuts for this, not just M5 jam nuts.

On an H-bot this should be easy. CoreXY is a little trickier since none of the idlers are 180-degree turns, but it's probably still possible if you mount the two idlers in each of the corners separately.

--
You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to 3dp-ideas+...@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/aeb573a0-7285-44ff-becc-af91e684cb54%40googlegroups.com.

[Ryan Carlyle]

Yeah, I keep looking at my cable routing to try to figure out where I can use a sliding idler on the 2020. That was my original plan with the spectra drive, but my intended location clashed with the HBP supports. There's a few places I could probably do ~90 degree turn idlers and a little bit of awkward routing. 

I think what I'm going to do is re-use my CoreXY adjustment mechanism. SDP/SI is kind enough to provide models of the cable, so it's super easy to just extrude around the cable profile DXF to get a toothed clip shape.

On Friday, October 24, 2014 2:01:24 AM UTC-5, Whosa whatsis wrote:

We've experimented with several types of tensioners, and by far the best is the sliding 180-degree idlers that I designed for the Bukito. It uses a single screw (plus a "finger" to keep it aligned) on the slot in the aluminum extrusion, with a separate screw to hold the idler pulley (though if you want the synchromesh adjacent to the extrusion instead of around it, it might work with just one screw). Just slide the idler piece on the slot to pull it tight, then tighten the mounting screw to lock it in place. I definitely recommend using real T nuts for this, not just M5 jam nuts.

On an H-bot this should be easy. CoreXY is a little trickier since none of the idlers are 180-degree turns, but it's probably still possible if you mount the two idlers in each of the corners separately.

On Thursday, October 23, 2014 at 9:00 PM, Ryan Carlyle wrote:

Synchromesh cable is just awesome to work with. The little bit of stiffness makes it much easier to handle than spectra. I think this is going to work really well, once I get some time to string everything up and get the tensions balanced properly. 

Definitely need to design some new "belt clips" for adjustable cable tensioning though. And these 624 V bearings seem to work ok for now, but in the long run I do think I'm going to want a larger bearing.

Here's how one line looks, strung up with slight tension:

--
You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[Ryan Carlyle]

I'm pretty sure synchromesh cable was invented for ridiculous line routing like this:

Here's my new cable clips, basically the same design as my CoreXY GT2 belt clips:

I need to put a lot of thought into re-routing the lines so I can get better grip on the drive pulleys, but that'll have to wait a few weeks I think.

[whosawhatsis]

Oh, I didn't realize that you were intending to contact the motor on only a 90-degree arc, or I would have warned you of that. It looks like you could add a sliding idler next to each motor to solve both the tensioning and pulley contact issues. That's how I ended up doing it on the skanbot.

On Friday, October 24, 2014 at 8:12 PM, Ryan Carlyle wrote:

Hmm. Everything is strung up, and the motions all go in the desired directions, so that's a small victory. But I'm getting a lot of tooth jumping on the drive pulleys. The cables need to be pretty dang tight to work with only a 90 degree contact arc. So I think I'm going to have to modify my cable routing pretty significantly to get closer to 180 degree contact and better grip.

Of course, this is with three-stepper drive, so the unbalanced idler line is definitely adding a lot of drive system drag.

More pics in the synchromesh cable thread, but here's the full printer strung up:

--
You received this message because you are subscribed to a topic in the Google Groups "3DP Ideas" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/3dp-ideas/gzcjamW6XHQ/unsubscribe.

To unsubscribe from this group and all its topics, send an email to 3dp-ideas+...@googlegroups.com.

To post to this group, send email to 3dp-...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/3dp-ideas/7e9a2ba5-3b5c-49c9-890c-178d8d176d56%40googlegroups.com.

[Ryan Carlyle]

Another synchromesh-drive printer: http://3dprintingindustry.com/wp-content/uploads/2015/01/voxel-8-3D-electronics-printer.png

Looks like an H-bot.