New arm design for Simpson style printer.

One other note. The arms still need to be geared together as they are in this design:
[forums.reprap.org]

So there would also be some play in those gears (so potential backlash from that).

Nicholas,

Watching your very cool video, I heard you mention that 5mm bolts are hard to get here in the US. If you have an Ace Hardware store near you, they might have a good supply. I have one about a mile from my house, and their selection is far better than the big box stores. They have most of what I need.

@JohnSL: I have tried ACE and every other place in town. They have M5 bolts but not much selection and they are way over priced. I have found them to be up to 10x the price of McMaster Carr. In this particular application, I needed partially threaded bolts so the string could slide over the shaft. (I wanted hex head zinc plated bolts but ended up with SHCS with a rough black finish.) The main problem was I designed for hardware I can buy from McMaster Carr but I am stuck with whatever we have locally.

@BloodBlight: I really like the double rack drive that you put in your video. I think with correctly sprung idlers you could have near zero backlash. I think it is worthwhile to work with it a bit to see what kind of advantages and disadvantages it has. For instance, I thought the gear drive was a horrible idea but thought I should at least try it before I threw it away and it turned out to be better than I expected and needed. The things I am concerned about is the part count, picking the mechanical advantage, idlers, etc. The only concern that I have that I don't think can be engineered away is the part count. That is not necessarily a deal breaker. My first thought is to drive the spur gear with a worm gear on the stepper. However, that adds backlash so I then revert to some kind of belt or string drive. BTW, a tensioned belt or string drive will have near zero backlash.

Hmmm... I will think about part count. Ether using loner gears and a clip at the end (one printed part and a screw is all it would be) would be easy, but I think a I can come up with an easy and low part count way to use springs (let me sit on it for a while).

My biggest concern would be the hinging point with a metal gear. I have to think about that as well.

In addition, it may be better to use a very small gear on the motor and direct drive it. That will keep the points as close to each other as possible. This would keep accuracy high, as well as torque and even increase the movement range (as this design will have a limited extension).

Here is a question, Is there any reason NOT to use a metal rack gear? They are cheap and fairly available.

How many bearings are in your design?

Nothing is wrong with a metal rack. (It isn't a reprapy solution but that may not be your goal.)

By my math you have the same mechanical advantage regardless of how big or small the spur gear is. There is no way to direct drive it.

Also, I don't know if I am seeing what you are talking about but accuracy and repeatability will increase as the racks are moved apart. As a thought experiment, what happens if they get really really close? I think you will find that the stability goes down.

I have 22 bearings in the current Simpson design.

Per arm or the whole design?

Hmmm, I don't see why this can't be a direct drive. Is it a clearance issue with the other arms?

As the gear gets smaller, the movements from the motor get smaller and should increase accuracy. However I am assuming you mean this is countered because of the geometry of the four points change (having a hard time picturing it). That could also be adjusted for by moving the four points farther out, but that would come at the cost of movement range.

It might be a good idea to start with the required movement range and go backwards from there....

Ok, I see, you can't move the pivot points.

Sorry, I am not a CAD guy, so this is primitive. Here is a Sketchup file.

@nicholas.seward

At the local hardware store I found a lot of things that might be useful for the MA shafts within the gear drive, such as nylon/steel bushings, spacers, roll pins, cotter pins, dowels, etc that could be used as a shaft and roller bearing for the gear drive. Thin walled hard nylon bushings on a dowel like surface would work well as a bearing surface for the string. But I worry about the string falling between the bushing and inside wall. I guess you could epoxy a washer on the ends to capture the string.

A chrome bolt will have less friction than the cadmium or black oxide coated bolts. If you're using a black oxide bolt, I would consider using a spring roll pin, as it will reduce your part count, and weight. You will need to orientate the gap away from the string. You could also have SS roll pins electropolished, or chrome plated etc.

Nothing simpler than a polished surface, if it's some thing that you can get away with. I prefer to have a thin wall rolling bearing if it can be made to fit.

A2

Edited 1 time(s). Last edit at 08/23/2013 06:46AM by A2.

I would like to test zinc vs chrome plated bolts. I suspect that they are very similar. Chrome bolts are much more expensive. (My guess is that both will be great and better than black oxide which is what I tested. My test setup seemed to indicate that black oxide was good enough.)

I think since different strings are moving at different speeds that nylon would get worn away.

@nicholas.seward

Spider wire fishing line can cut through plastic, and your hand like a razor blade.
Bushing would have to roll for a plastic to work.

"The metal Line Roller was sliced into 2 by the braided line. On post mortem, it was always because of the plastic bushing inside the Line Roller that had gotten a little sticky, thus the roller could not keep pace with the line speed."
[www.fishingkaki.com]

In this case the speed of the line was causing the failure, but I don't think you are achieving speeds any where near what a fish can strip line off of a reel

I think that the softness of zinc, and it's oxidation might not be a long term solution.

Any smooth hard steel surface will work well with Spider wire fishing line. One of it's weakness is it's propensity to slip, (knots don't hold), special knots have been developed specifically for Spider wire fishing line. You can also draw ideas from the eyes of the fishing, some are just plain steel.

Creep is another problem, the guitar tuner would be hand as it's easy to back off the tension when the printer is not in use.

Edited 2 time(s). Last edit at 08/23/2013 12:59PM by A2.

BloodBlight Wrote:
-------------------------------------------------------
> Sorry, I am not a CAD guy, so this is primitive.
> Here is a Sketchup file.


Thanks a lot, that clarifies your idea.

It is ok if you want to use metal racks if that gets you closer to your goal. In my case, I wanted to improve Simpson, making it more easy to replicate, less vitamin hungry, more robust and so on, on that way of thinking, which I believe has been partially achieved by the geared joint, an idea born in this thread. So, all ideas are welcome.

Getting back to the new arm design...


Nicholas, have you considered using a printable ratchet like this:



instead of the tuning peg? I know in many ways a tuning peg might be unbeatable, because has been designed to mantain a high tension on a string, but probably the ratchet could work, with the added benefit of being printable.

My idea is to embed it on the arm, so you can only see the peg coming out of it. Of course, that has to be designed and corrected, I'm only showing a rough concept and I don't know if PLA will be flexible enough, probably ABS would be better. Also, if you embed it, removable support material would be needed, but I see no problem, as it could break easily when first using it, we only would need to consider enough voids to let that material go out.

Guizmo,
I like the ratchet idea, and I can see how each part could be printed, and integrated.

But because the string is inelastic it will require very fine detentes to achieve the correct tension, this could prove difficult.

A2

A2 Wrote:
-------------------------------------------------------
> Guizmo,
> I like the ratchet idea, and I can see how each
> part could be printed, and integrated.
>
> But because the string is inelastic it will
> require very fine detentes to achieve the correct
> tension, this could prove difficult.
>
> A2

Yeah, I devised that. That could be corrected using aditional layers with an angular offset, so you can stop at half, or less, steps of the main ratchet. Another solution is a mechanism based on the Vernier, that can stop in even less fractions of a turn. We could make this by making the stationary part and the movable part with a different quantity of stoppers, and then a second layer with a radial symmetry of 180 degrees to assure at least 2 point of contact.

Well, it is just an idea towards a less vitamined Simpson.

I bought a 4 foot driveway marker to use as a pulley shaft between the gear drives.
$1.99 USD, fiberglass rod.

If the O.D. is not an odd size, I will use a Nylon spacer or bushing as a pulley for the string to slide over.
Surgical SS tubing would probably work well as a pulley, and it has an electropolished surface.

A2

Edited 1 time(s). Last edit at 08/31/2013 04:44AM by A2.

@A2 - do you have a build log anywhere? I have a large format Simpson in the cards and am a fan of the hardware store approach where possible, with possibly thin-walled steel extrusions or laser-cut sheet metal for the arms? My goal is to be close to a 1m build area with a 4' X or Y length, but I might fudge the X by placing the bed on wheels on the ground, driven by a long steel rack gear. (think Lego Monorail-type) -- I would like to get at least two large format machines running, with multiple beds that can travel between them. (different materials/processes/nozzle diameters?)

[www.amazon.com]

@jason.fisher

Build log:
What should I do, start a new thread? You'll have to point me in the right direction.

Belt/Conveyor:
I was thinking along the same lines as you. Placing a belt/conveyor under the print head would allow me to run the machine in a production operation. There are examples on YouTube of this already with delta robots.

With a belt/conveyor you can dump the parts into a bin when the printing operation has completed. You can run 24/7 3D printing! It's on my bucket wish list I am also interested in open-CAM and 3d printing solutions.

I think a simple cable to move the table between two print heads would work fine. Camera dolly's are a good example of a simple long linear solution. The other day I found 3 inch plastic pulleys for clothes lines at Home Depot for cheap, they would work for the cabling (but they need bearings). Again on YouTube there are examples of cable/trolley robots suspended from cables performing fairly accurate work in free space.

Robot arms:
The power of the block and tackle, stepper motor combination, affords you many material choices to construct the arms from. Nichloas drive gear demo is very impressive, looks to be a lot of execess torque available. Look at truss, and crane boom designs for ideas that are light and ridged.

In addition use a larger stepper motor (more power!), I'm looking at NEMA 23 even for the small one that I'm building. But I'm undecided, I might have to buy two sets of motors NEMA 17, and NEMA 23. Note: there are differences within a motor series. Pay attention to the torque, and amp ratings of the stepper motor. ... and make sure that the microcontroller, and accessories support a NEMA 23 stepper motor (amp, volt).

A2

Twine tensioner idea:
This requires two holes in the arm, and two drywall screws.

1 Wrap the twine around the first drywall screw. This is your winch.
2 Pass the tag end of the twine through, and out the other side of the second hole, and apply tension to it to hold it in place.
3 With a screwdriver winch the string tight.
4 When sufficiently taut, place the second drywall screw into the second hole to anchor the tag end of the twine.

A2

Edited 1 time(s). Last edit at 08/31/2013 04:28PM by A2.

A2: I am sure that is just a mockup but just in case I want to say that a herringbone like arrangement of the gears or some other alignment scheme is needed to keep the arms from shifting in the axial direction.

Your mockup reminded me that cycloidal gears also could be fun to play around with. The beauty of cycloidal gears is that it is all rolling contact instead of sliding contact.


Situations like this...

... make it unsuited for power transmission or even for this rolling contact situation. However, using a cycloidal profile combined with a herringbone arrangement should prevent the situation above. I think it would be better suited to 3D printing than an involute profile.

I will be curious to see if you can tension using your procedure above. It sounds promising.

Edited 1 time(s). Last edit at 08/31/2013 05:32PM by nicholas.seward.

@A2 - I think I would just start a new thread for now, use it as a dumping ground, and then when you master the wiki you can copy/paste final solutions/designs there and a link back to the thread.

@Nicholas

Correct, that model was thrown together in about 6 minutes

I envision that a few crisscrossing of the string around the winching screw would anchor it to the screw so it wont slip during tensioning. Do you think the screw will be able to pull the string tight?

Herringbone gears vs. cycloidal gears:
I like the idea of replacing the involute profile with a rolling (lobed) profile.
Positive displacement pumps use this profile.

I agree that a round profile gear tooth is more suited for 3d printing, as it would improve the lay of the material.

A2

@jason.fisher
I did a search for build log in the reprap wiki but I'm not seeing any thing.
Maybe it has another name, most likely I just don't understand the wiki

The word RepStrap came up in a search, it seems to be close to what I'm working on.
But I don't know where to go from here with it.

A2

Another idea for a string tensioner...

Printrbot Simple x-axis tensioner
by Tshawkins, published Aug 29, 2013
[www.thingiverse.com]

Edited 1 time(s). Last edit at 09/01/2013 05:44PM by A2.

Nice idea of cycloidal gears. They worth to be tried and are very well suited to all this almost frictionless DLCJ stuff. I l ike them.

Since cycloidal gears are not your everyday tooth form, I had to make a generator. (You can approximate a cycloidal gear with arcs of alternating concavity. I don't think that will be good enough for the gear joint. Near perfect constant gear ratio is needed to remove tooth artifacts from a print. I don't want to trade in lead screw artifacts for gear artifacts.) You will need python and the dxfwrite module.

You can import the dxf into OpenSCAD, Solidworks, or Inventor. Here is an example helical cyclodial gear I made in Inventor from the import.


Edited 1 time(s). Last edit at 09/03/2013 04:52AM by nicholas.seward.

If anyone wants to play around with cycloid gears directly in OpenSCAD, you might be interested in some of these:
Epitrochoids and Hypotrochoids
Roots Blower
Cycloidal Speed Reducer in OpenSCAD

A2: If you would like help setting up a wiki page, I'd be happy to give you pointers. The easiest thing to do is to pick a page you like, and use it's source as inspiration. You can create and call your page pretty much whatever you want... something like "A2_Variant_Simpson_Build_log" would probably be ok with everyone...

MattMoses Wrote:
-------------------------------------------------------
> A2: If you would like help setting up a
> wiki page, I'd be happy to give you pointers. The
> easiest thing to do is to pick a page you like,
> and use it's source as inspiration. You can create
> and call your page pretty much whatever you
> want... something like
> "A2_Variant_Simpson_Build_log" would probably be
> ok with everyone...


You could also request access to the Reprap Builders blog...

nicholas.seward Wrote:
-------------------------------------------------------

> I used herringbone gears for axial and rotational
> alignment. I offset the herringbone gears from
> side to side by 1/2 of a tooth. This allows you
> to use the same part for both sides of the design.
> You could use string or nylon tip ties to hold
> this together. It would be easy to add tension
> from the back of the arm. Additionally, this
> looks very printable. I made the helical angle 45
> degrees to aid in the printability.

NIcholas, could you do a drawing of your string path showing string, stepper and tensioner?

I am having trouble picturing the path.

Thanks,
Ralph