Tensegrity delta

[Ryan Carlyle]

I’m doing a drivetrain update for my Tesseract Delta next month (new rails, tweaked arm lengths, etc) and was thinking about doing something fun with the arms. It currently has u-joint arms, which work well, but they’re bulky compared to a traxxas joint or whatever. So I’m thinking about building a tensegrity arm — one central rod in compression and two cables in tension on either side. Together they effectively make a regular parallelogram linkage, but it’ll look neat.

There are a few things I haven’t figured out yet:
1) Spectra, steel wire rope, belts, or something else for the tension cables?
2) What kind of end joints on the compression rods? I want to run wiring (and maybe even filament) through them if possible, so hollow tubes with an open joint would be ideal.
3) How to tension the cables? Arm length will be fixed by the compression rod, but uneven cable length will cause effector tilt. I’m not sure whether it would be best to use six separate cables built on a jig, or one continuous cable that wraps all the way around the three arms and locks in place after tensioning, or what.

[Whosawhatsis]

I read the title and immediately thought you were proposing a deltabot with a tensegrity frame. This might actually be possible. The most basic free-standing tensegrity structure has three beams and a triangular base and top, but the beams get pretty close to one another in the middle (depending on relative lengths of the vertical and horizontal cables), which would make your build volume neck down pretty early (depending on delta arm length relative to the other dimensions), and you would have Z axis issues as the beams get closer together. This would be totally impractical, but it should totally be Nicholas Seaward's next build.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/23bb5936-cf6b-4724-8fa6-d58468852439%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ryan Carlyle]

Tensegrity frame sounds like a terrible and awesome idea :-)

[dan...@puptv.com]

You might want to consider making a printer with a triangular anti-prism:

https://en.wikipedia.org/wiki/Antiprism#/media/File:Trigonal_antiprism.png

For those that can't see the wikipedia picture, it's 2 equilateral triangles pointed opposite directions (twisted 60 deg from one-another) for the top and bottom, and 6 vertical members connecting the top apexes to the bottom apexes making a frame from triangles...  extremely strong and rigid, even when you use very light aluminum members to make a huge printer.

I've considered attaching linear rails to 3 of the vertical members, and correcting for the angle in software... but never got around to it.

I've attached a picture of the plastic anti-prism vertex piece that makes a triangular anti-prism with 15mm x 15mm Misumi  aluminum members, and the scad file that generated it.  This is from a Flying Sky-Delta printer design.

Daniel - http://www.TriDPrinting.com/

[Zatsit]

Yes...

But if the ensemble becomes a musical instrument, resonating with different frequencies, it will show on the printed objects! Tensioned belts already induce their own problems...

Our printers are implacable seismometers: they record vibrations in the printed object!

[Patrick Barnes]

By some measure we should make our 3D printers from concrete?

--

You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/5d9de89e-7a58-4e70-a107-2371fd141b46%40googlegroups.com.

[Ryan Carlyle]

Cast iron scraped ways in a concrete bed, yeah. That gets you to the tens of micrometer precision range if you do it right.

What about tensegrity delta arms though? :-)

[Whosawhatsis]

They would have the atypical (but not unique) property that the individual arms would fight twisting, even in the absence of the rest of the mechanism. This is technically an over-constraint, but because it shouldn't be able to constrain to any shape other than a untwisted plane, it's probably ok (and MAY even help compensate for some slop in other parts of the machine). BTW, I've long thought that it would probably be a good idea to put the third set of joints in the middle of a regular rigid delta arm pair to accomplish the same thing.

Keeping the two runs of tensioned line equal is a bigger issue, and if you get that wrong, you'll have problems, but that's easier to adjust than the relative lengths of two rigid arms.

One question is what type of tension line you want to use. You need something that won't stretch, but you also want something that you can bend extremely sharply, repeatedly, without causing fatigue (no steel cables). You could put ball joins on the ends of the cables I suppose, but the necessary tension would make those pretty high-friction.

On June 23, 2018 at 13:20:57, Ryan Carlyle (temp...@gmail.com) wrote:

Cast iron scraped ways in a concrete bed, yeah. That gets you to the tens of micrometer precision range if you do it right.

What about tensegrity delta arms though? :-)

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/3690e22b-9bbb-4695-930b-ffbf3ec1b90f%40googlegroups.com.

[Ryan Carlyle]

Aramid fishing line should work for bending and maintaining tension. Spectra would require periodic adjustment for creep probably.

For steel wire or belts where bend radius is an issue, the thing I was thinking of was to first use a hinge type joint at each end (effector and carriage) so the parallelogram is 2D bending only. Then the line gets run around matching radii at each end, so as the parallelogram tilts, equal line is wrapped on at one end as is unwrapped at the other. For a belt, you could use a printed half-pulley or whatever.

[Whosawhatsis]

Ah, good point. Using toothed belts with half (quarter might even be enough) pulleys at the ends would also allow you to ensure equal length on the two sides (to within the tolerance of the belt pitch over that length). A straight toothed clamp parallel to the hinge axis could transition to a partial pulley of, say, 135 degrees. Unless you use real pulleys on one end though, you would end up needing to adjust the length of the rigid member to tension it. With the belts providing a known length for the tension parts, you could do this with a set screw against the end of the beam, or even put opposing threads on each end of it and use it as a turnbuckle. These mechanisms work better under tension than compression, but if you maintain enough overlap between the internal and external threads, or use a long enough guide tube for the setscrew version, you should be able to maintain sufficient rigidity.

On June 23, 2018 at 14:03:54, Ryan Carlyle (temp...@gmail.com) wrote:

Aramid fishing line should work for bending and maintaining tension. Spectra would require periodic adjustment for creep probably.

For steel wire or belts where bend radius is an issue, the thing I was thinking of was to first use a hinge type joint at each end (effector and carriage) so the parallelogram is 2D bending only. Then the line gets run around matching radii at each end, so as the parallelogram tilts, equal line is wrapped on at one end as is unwrapped at the other. For a belt, you could use a printed half-pulley or whatever.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/43190992-fafa-4a0a-bab3-48c6e42efdcc%40googlegroups.com.

[Ryan Carlyle]

A turnbuckle in the compression rod would be fine. Toothed half-pulleys could also be rotated for tensioning. That eliminates some of the length calibration benefit of the belts though.

I’m entertaining the idea of using something like a guitar tuner to tension a single line wrapped through all three arms, with clamps to lock in the line positions. I’d place the effector flat on the bed when tensioning and locking the line tension to get it all parallel. Just have to get the carriage mounts parallel to the bed then.

[Whosawhatsis]

Getting the line from one arm to the next without a change in tension when the arms move will add significant engineering complexity. You could to that for the two runs on each arm pretty easily, but one long string for all three is probably more trouble than it's worth.

On June 23, 2018 at 14:46:50, Ryan Carlyle (temp...@gmail.com) wrote:

A turnbuckle in the compression rod would be fine. Toothed half-pulleys could also be rotated for tensioning. That eliminates some of the length calibration benefit of the belts though.

I’m entertaining the idea of using something like a guitar tuner to tension a single line wrapped through all three arms, with clamps to lock in the line positions. I’d place the effector flat on the bed when tensioning and locking the line tension to get it all parallel. Just have to get the carriage mounts parallel to the bed then.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/b8ea46a4-e6d9-4786-aacf-5db9c96df07c%40googlegroups.com.

[Ryan Carlyle]

Hmm, good point, probably too much trouble to use one line.

[Ryan Carlyle]

Right now I’m leaning towards making six lengths of wire on a jig using crimped eyelets. Then use a turnbuckle or screw type arrangement on the compression rod for tensioning. This makes a hollow wire path a lot harder but that’s not really all that important to me. The wire eyelets can be bolted to a bearing or something to provide swiveling without wire fatigue.

[Whosawhatsis]

I think it's going to be pretty difficult to get even lengths when crimping eyelets...

On June 23, 2018 at 18:19:47, Ryan Carlyle (temp...@gmail.com) wrote:

Right now I’m leaning towards making six lengths of wire on a jig using crimped eyelets. Then use a turnbuckle or screw type arrangement on the compression rod for tensioning. This makes a hollow wire path a lot harder but that’s not really all that important to me. The wire eyelets can be bolted to a bearing or something to provide swiveling without wire fatigue.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/1c37efdd-e094-45d3-a06f-605cb5dbd681%40googlegroups.com.

[Ryan Carlyle]

Seems about the same as epoxying traxxas joint rods to me. Except faster.

1) Crimp one end of the wire like normal

2) Bolt that crimp to an extrusion with some kind of 3d printed socket/cradle to hold the eyelet in a known position

3) Cut the wire 2x too long or whatever convenient length

4) Slide an eyelet on the uncrimped end and put that eyelet in a second 3d printed cradle, X distance away

5) Pull on the excess wire length end with a consistent force, eg hang a few kilos of weight

6) Crimp the second eyelet

7) Trim off unnecessary wire length

Repeat for six total wires. They should all match about as close as homemade delta rods do. The only particularly complex part of this is modeling up a cradle that will hold an eyelet in a defined axial position, allow pulling the wire through, and leave room for a crimper or pliers or something. Doesn't seem too hard. 

On Saturday, June 23, 2018 at 10:08:15 PM UTC-5, Whosa whatsis wrote:

I think it's going to be pretty difficult to get even lengths when crimping eyelets...

On June 23, 2018 at 18:19:47, Ryan Carlyle (temp...@gmail.com) wrote:

Right now I’m leaning towards making six lengths of wire on a jig using crimped eyelets. Then use a turnbuckle or screw type arrangement on the compression rod for tensioning. This makes a hollow wire path a lot harder but that’s not really all that important to me. The wire eyelets can be bolted to a bearing or something to provide swiveling without wire fatigue.

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

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[Whosawhatsis]

You need to keep the cable tensioned, keep the eyelet positioned precisely (preventing both slippage along the axis of the cable and rotation around the eye), AND actually crimp the eyelet, all at the same time. Seems to me that those goals would work against one another. You would be better off tensioning the cable, clamping it, then running the end of the eyelet up against the clamp as you crimp them, though you'll probably find that the distance from the end of the eyelet to the eye (or the internal diameter of the eye, for that matter) just isn't very precise. Eyelets just aren't precision parts, and will fight you if you try to use them in a precision way.

We used eyelets crimped on synchromesh on the bukito, but I put tensioners on each axis that allowed for about a centimeter of tolerance on those crimps for precisely this reason.

On June 23, 2018 at 20:17:57, Ryan Carlyle (temp...@gmail.com) wrote:

Seems about the same as epoxying traxxas joint rods to me. Except faster.

1) Crimp one end of the wire like normal

2) Bolt that crimp to an extrusion with some kind of 3d printed socket/cradle to hold the eyelet in a known position

3) Cut the wire 2x too long or whatever convenient length

4) Slide an eyelet on the uncrimped end and put that eyelet in the 3d printed cradle

5) Pull on the excess wire length end with a consistent force, eg hang a few kilos of weight

6) Crimp the second eyelet

7) Trim off unnecessary wire length

Repeat for six total wires. They should all match about as close as homemade delta rods do. The only particularly complex part of this is modeling up a cradle that will hold an eyelet in a defined axial position, allow pulling the wire through, and leave room for a crimper or pliers or something. Doesn't seem too hard. 

On Saturday, June 23, 2018 at 10:08:15 PM UTC-5, Whosa whatsis wrote:

I think it's going to be pretty difficult to get even lengths when crimping eyelets...

On June 23, 2018 at 18:19:47, Ryan Carlyle (temp...@gmail.com) wrote:

Right now I’m leaning towards making six lengths of wire on a jig using crimped eyelets. Then use a turnbuckle or screw type arrangement on the compression rod for tensioning. This makes a hollow wire path a lot harder but that’s not really all that important to me. The wire eyelets can be bolted to a bearing or something to provide swiveling without wire fatigue.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@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/1c37efdd-e094-45d3-a06f-605cb5dbd681%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

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

To unsubscribe from this group and stop receiving emails from it, 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/3ce59fea-45a2-4fc4-982a-ad94ca0339e3%40googlegroups.com.

[Ryan Carlyle]

If it can’t be crimped precisely enough, it’s not hard to epoxy the wire to any kind of end termination I might want. Precision clevis maybe. Hell, I could epoxy the wire directly to a bearing.

Guitar tuning pegs and aramid/spectra line seem really super easy, but we’d need a clever way to keep the lengths the same.

[Ryan Carlyle]

A random comedy thought: ball chains for the tension lines. You get the benefit of 3d bends with the ability to count balls for length calibration. 

[Whosawhatsis]

The ball-to-ball distance consistency is probably pretty bad. Also, they are known to stretch over time.

On July 6, 2018 at 19:08:28, Ryan Carlyle (temp...@gmail.com) wrote:

A random comedy thought: ball chains for the tension lines. You get the benefit of 3d bends with the ability to count balls for length calibration. 

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, 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/bc3eed08-eca0-4bcc-ad4e-78bd02eadadc%40googlegroups.com.

[Ryan Carlyle]

That's why it's a comedy idea :-)

Curious if there's such a thing as a precision ball chain though. 

On Friday, July 6, 2018 at 9:15:29 PM UTC-5, Whosa whatsis wrote:

The ball-to-ball distance consistency is probably pretty bad. Also, they are known to stretch over time.

On July 6, 2018 at 19:08:28, Ryan Carlyle (temp...@gmail.com) wrote:

A random comedy thought: ball chains for the tension lines. You get the benefit of 3d bends with the ability to count balls for length calibration. 

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

To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@googlegroups.com.

[Steven Butterfield]

I think that's called marrying up. :)

[Whosawhatsis]

Well, not _exactly_ a precision ball chain, but what about posi-drive?

To unsubscribe from this group and stop receiving emails from it, 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/f494ffad-d614-449a-9ae3-10bf18511d76%40googlegroups.com.

[Ryan Carlyle]

Had occurred to me. Probably work a lot better than ball chains. I don't know what the accumulating pitch error looks like for posi-drive though. 

I'm probably just going to use GT2 belts on printed pulley-sections, with a separate hinge mechanism on each end, but I'd really like a wire/cable option that allows 3d bending to eliminate the end-hinges. They add a ton of mechanical complexity and effector footprint. 

On Friday, July 6, 2018 at 9:24:39 PM UTC-5, Whosa whatsis wrote:

Well, not _exactly_ a precision ball chain, but what about posi-drive?

On July 6, 2018 at 19:16:12, Ryan Carlyle (temp...@gmail.com) wrote:

That's why it's a comedy idea :-)

Curious if there's such a thing as a precision ball chain though. 

On Friday, July 6, 2018 at 9:15:29 PM UTC-5, Whosa whatsis wrote:

The ball-to-ball distance consistency is probably pretty bad. Also, they are known to stretch over time.

On July 6, 2018 at 19:08:28, Ryan Carlyle (temp...@gmail.com) wrote:

A random comedy thought: ball chains for the tension lines. You get the benefit of 3d bends with the ability to count balls for length calibration. 

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@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/bc3eed08-eca0-4bcc-ad4e-78bd02eadadc%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "3DP Ideas" group.
To unsubscribe from this group and stop receiving emails from it, send an email to 3dp-ideas+unsubscribe@googlegroups.com.
To post to this group, send email to 3dp-...@googlegroups.com.

[Steven Butterfield]

Ryan, you could make a simple jig to make the cables the same length when you crimp them. Pegs in a 2x4 could do the job. What type of crimps would you use?

[Ryan Carlyle]

The crimps for synchromesh work well for steel wire. They’re just a little eyelet on a lug that you mash down on the end of the wire. (I mean, there’s probably a tool, but pliers work fine.)