New arm design for Simpson style printer.

I love this printed gear bearing. Maybe not entirely helpful in this case, but still inspirational. [youtu.be]

You could print a circular inner bearing gear path and custom outer cam profile?

Edited 1 time(s). Last edit at 09/06/2013 08:27PM by jason.fisher.

3M Anti-Slip Tape will be a test substitute for the herringbone gear.
The lateral, axial, rotational grip seems adequate.
I think a coarse sandpaper could be substituted.

4” inch wide by 12” inch long, adhesive coating, cost $2.79.
@0.70” inch wide by ~9” long, you need ~2 feet, total cost $5.58.

A2

Edited 1 time(s). Last edit at 09/07/2013 03:37PM by A2.

A few more tensioner ideas:

Belt Tensioner for printrbot GO - Y axis
By ei8htohms, published Aug 18, 2013
[www.thingiverse.com]


Belt terminator
By chowderhead, published Nov 24, 2012
[www.thingiverse.com]

Edited 1 time(s). Last edit at 09/07/2013 03:45PM by A2.

Here is a release candidate that I just posted for the arm design. Let me know if you guys see anything I missed. Click over for more pictures and more info.

Have you run any Finite Element Modelling on the arms to check for stress points or model the wear?

I am in over engineer mode to make sure there is no chance of failure. Optimization will come later when I have time.

I know how to do stress analysis but I am clueless about how to start a wear analysis. Theoretically, all the contact is rolling contact so I would suspect a FEA model would predict surface smoothing but not plastic removal.

Bottomline, I am using the sledgehammer approach right now.

HP used to make a plotter that used a single sparse-fine-grit-covered light weight shell as a roller on the back side of the paper to move the paper in Y while a carriage moved the pen in X. There was a wide roller on the front side of the paper that held it against the grit..

The grit made little dimples in the paper that kept it in a repeatable relationship with the roller surface, no slip. This roller reversed itself many times in the course of a typical plot, and as I recall it worked pretty well, without loss of registration or slip.

The analog here would be to put sandpaper on one arm, and let it make little indents on the other. One might have to change the radius to account for the grit just being on one side.

Or, embed hot sharp grits onto one side, then heat the other side to soften it and roll them together to make the matching pair.

Seems like gear teeth are simpler, in this situation.

New arm look great with the motor on top


Just curious, are you cutting (swept cut/extrude) the round root/crest gear profile based on a plane that is set "normal to curve" of the helix? (SolidWorks jargon). Or is your profile sketch plane one of the default planes?

A2

I am not sure what you mean. I have used Solidworks.

Workflow:
*Using my program that I provided above I import the cycloidal profile to the xy plane.
*I extrude along the z while rotating in the z direction. The rotational pitch is pi*D where D is the pitch diameter of the gear. This gets you a 45 degree helical gear. About.

PGDJ in action. We can definitely use this tech along with DLCJ to get rid of all bearings.

It's nice to see it all come together.
I like the rolling motion, looks solid.

A few comments about curb side appeal, etc..:
1. Gear rub test: add a small amount of Prussian blue to the gears to see where they are rubbing. Maybe they are not doing what you thing they are.
2. Product validation if you dare? Spray some hair product onto the gears, then add a gummy bear, and some beach sand. If it's still rocking back and forth push a pencil in between the gear teeth, maybe cut the string on one of them too, I'm curious what it does. Make sure your camera is in focus
3. The bolt heads look out of place. It's a combination of the shinny chrome look, the scale, and the shape of the hex. Consider embedding the head of the bolt within the side wall of the gear, and/or behind a cover plate.
4. Goth arch style apertures are not rocking the industrial design boat (but I bet it makes it easier for fat fingers to thread fishing line). Consider adding symmetrical apertures on both side of the gears of a different design/style, or a cover.
5. If possible move tensioner inboard, and/or hide it under a cover. I'm surprised to see the tensioner on the forearm, care to share why it ended up there?
6. Liability, you might want to sell this only as a kit, and let the customer be the manufacturer and bearer of injury claims such as finger crushing. Consider adding a break away feature that releases the string tension if a finger like object gets caught in the gears (toggle release, might not be possible to add this kind of feature in such a low cost device). Or add a shield around the gear (clear), again it might be too expensive. Shields could look like elbow pads on a mechatronic figurine.

Little cover plates, safety shields, bolt head covers, etc., could be the first parts that the customer prints.

The bearingless Simpson is a fun challenge. Right now I'm debating if I should use plywood instead of the wood that I have for the arms. Which means I need to go hit the streets and find some curbside plywood

I like to see more videos of it moving!

A2

That is hot. When the head drops, it seems to jerk at the same spot each time. Is that a math problem? Acceleration seems like it would be very beneficial with this design also?

@A2:

That is quite the list. To be honest, I am quite satisfied and have put lots of thought in why it had to be exactly as it is. Believe me that the design was the result of 1 month of inner mental ping pong. For scale, I designed the Sextupteron in one evening. There are so many little details to the design that all have to work in concert. All that said, I invite others to make variations and to help me test it out.

1) You have to wait on that one.
2) What? Hair spray? Gummy bears? Beach sand? It sounds like you are having one hell of a party. (Side note: I would like to see what you ask a Cartesian bot owner to do. Maple syrup on the rails? In all seriousness, I do have a larger burden of proof because this is new but if you get crap in the gears it would be reasonable to clean it out just like you would wipe off a rail. However, I look forward to doing a little torture testing.)
3) Hex bolts are cheap and I need slick plating of the unthreaded part. They are M8 because M3 are all fully threaded and I only use M8 and M3 hardware if I can help it. M5 would work but why not use M8 if that is what I need everywhere else. The cantilevered bolt head makes stringing it so so so much easier.
4) I would have left the whole side open but I was worried that I would get a mess of plastic so I put the center support. The gothness of it was unintentional. I thought about about adding matching cut outs on the other side but I didn't need to for any reason. I wanted to make sure that side was super stable because the M8 bolts cantilever from that side.
5) The hole in the tensioner dictated its placement. I put the springs and the tensioners on the forearm because it would have made the motor arm too confusing. BTW, I do 1.5 loops between each bolt.
6) This is not a finger crusher. The spring adds up to 3.78lbf. There are 6 strands holding the halves together for a total around 20lbf. This would hurt but fingers would live to type another day. (Quit sticking your fingers in my machine. :-) I guess it is hard to avoid when you are making a miniature beach scene with gummy bears with styled wigs on my print bed.)

@jason.fisher:
There was a tracking problem with the pulleys. I have mostly solved that now.

I understand the design process that you have gone through, and I figured as well that your design is pretty much complete, that's why I wanted to say a few things about the design today. I think you're doing a great job, and on the right track.

Gummy bears:
Seeing the exposed mechanics of it's operation is intriguing. I'm am curious what happens when some thing falls in between the teeth (e.g. filament). I've tested hundreds of different products to failure, performed lots of goofy test, you just never know what might happen

Prussian blue:
Do you have support material in the arch? I'm assuming you're printing with the stepper motor mounting surface facing down on the print bed.

A2

If one were really worried about crushing fingers or sand and gummy bears you could make up a acrylic bell cover.

Ralph

"Attach this for gummy bear safety"

A2: Having had a little experience with the arms so far, I can comment on a couple of things. The access holes on the sides aren't just for fat fingers, they're for maintaining sanity while stringing. It still takes quite a bit of fiddling, but without them I think you'd have to be a surgeon (or have experience building ships in bottles). Having the bolts sticking out of the sides also makes stringing a lot easier, since you can pull the line out through the window, loop it around the bolt head and push it back. You can also pull an existing loop out in order to put the next one inside it, which is handy at a couple of points. Smaller windows on the other side would be handy too, but not necessary. The arms do require quite a bit of filament, so perhaps they could be an option in the future just to save a little plastic. And the openings print fine without support, using that shape.

The guitar tuning machines are nice but I don't think they're entirely necessary. There were other suggestions made in this thread and they'd probably work; the only thing you need to do is get enough tension on the string to start extending the spring, there's no precision adjustment required. At the same time, tuners are solid, relatively cheap and easy to use. And who doesn't like some nice shiny chrome on their printer?

I can't speak authoritatively on product liability issues since IANAL, but it strikes me that the hotend is the biggest risk on any printer, and crush injuries are likely to be a lot less common. Yes, the nutcracker effect of this design is very apparent, but I wouldn't want to get my finger between the bed and the frame on a belt-driven cartesian printer either, or between the gears of any of the extruders I've used. Neither would I let little kids with poking fingers near any of my printers (or any of a lot of other things I have

@Nichloas your label suggestion is spot on!

I've written a lot of instructions for products.
For legal reasons I add a leading statement BEFORE my instructions, and it went some thing like the following:

Instructions:
1 Follow your facilities protocols.
2 “Then” insert your instructions and warnings “here”.
3 ...

Some lawyers say... the statement “Follow your facilities protocols.” shifts the liability to the user/company/facility. In addition the law in my field also requires me to provide instructions as well. So there could be/are contradictions. So the first statement takes precedence.

Adding a label/sticker or embedding a statement onto the the product might be all that is needed. What instructions come with a blender?

@Owens
I agree the windows to access the pivots are useful. My goth comments were more about the ID (industrial design) or the curb appeal. ID is important to me because I enjoy stylized products.

Hot end:
I've noticed that most of the 3D printer manufacturers have a label warning/cautioning about not touching it.

Extruder:
Another safety label, and/or cover is needed.

A2

Oh no! Somehow I just knew this would happen...

ha ha
Oh Nooooo Mr. Bill, watch out!

Nicholas,

It looks amazing, I bet there is little to be improved. The gear joints add something special to this machine, it looks like nothing else I´ve seen, like more industrial.

Congratulations to you and also everybody that has been involved. Good work!!

Thanks to you! If it wasn't for you I would be stuck with torsion springs. Let's take this all the way.

BTW, You could do all 3 DLCJ's and the 1 PGDJ for one arm with one string. It could suck to string but you would only need one tensioner and one spring. Something to think about.

Additionally, somewhere on g+ someone came up with the idea to use separate herringbones (double herringbone if you will) one on each side of the joints. This will make registration much better.

@Nichloas
Looking at Owens video, I noticed the leveling switch or the max extension switch, I'm not sure what it's called or doing. It looks to be a Hall effect switch/sensor, is this correct?

I'm not able to locate the switch in your BOM. I down loaded the BOM a month ago, so maybe mine is out dated?

To ensure stability of the sensors, I'm wondering if it would be helpful to pot the switches in hot melt glue, or print a pocket for it?

Hall effect switch hack that might be useful for the found Simpson:
Get an old Honeywell keyboard, the type with hall effect key switches, and you can open up the key switches and remove the Hall element, and it works as mentioned here. The Honeywell keyboards were the best and most reliable.
[hackaday.com]

@Nicholas, and now @Owen.
Dynamics test:
The arms look very strong. I want to get an idea of the combined strength of the 3 arms.
How much can it dynamically bear?
Remove the filament guide, then place a weight on top, secure with duck tape, press go!

It's great to see a second geared Simpson running, now we're expecting 2X the vids.

tks!

A2

The switch is just a cheapo snap action switch.

This design isn't published anywhere. I will be posting it in a week or two.

For stability the switches are screwed in.

At speeds up to about 150mm/s I could put the weight of my arm on the hub and it would pick it up without skipping. (Of course, the steppers are running a bit on the hot side.)

Just an idea.
String crossing issues:
I'm wondering if a thin free floating 3D printed plastic bearing would help with string organization?
My first thoughts it will bind up and the string will cut the plastic in half.
Steel would be my first choice.

A2

Edited 3 time(s). Last edit at 09/18/2013 03:45PM by A2.

MicroSwitch vs. Hall effect sensor:

I've read some where in these forums that the software needs to me modified just a little if you replace a microswitch with a hall effect sensor. Do you see any issues if the microswitch was exchanged with a hall effect sensor?

I've only had bad experiences with microswitches on industrial equipment. They always drift out of calibration, it's not possible to validate them to a high standard.

tks.
A2

I think you'd want independent pulleys, sharing a bolt, and you'd want the hub to be wider than the flanges so the friction between neighboring pulleys when they touch would produce the least torque.
They'd have to be pretty smooth and round or they'd create distortions in the print.

@DaveGadgeteer
You bring up some valid points.
I'm trying to think of a hack for this.

tks.
A2

A2 Wrote:
-------------------------------------------------------
> MicroSwitch vs. Hall effect sensor:
>
> I've read some where in these forums that the
> software needs to me modified just a little if you
> replace a microswitch with a hall effect sensor.

The only change that should need to be made is to disable the pullup resistors for endstops. Just this line needs commented out in Marlin:

#define ENDSTOPPULLUPS // Comment this out (using // at the start of the line) to disable the endstop pullup resistors

Possibly the invert settings of the endstops would need to be changed too.

String pulley bend radius:
Braided line can run across an acute angle fairly efficiently.

Pulley Diameter:
If the diameter of the braided line is .02” inch (.5 mm) x 3 = Dia 0.06” inch (1.5 mm).

You could use a carpenters nail as the pivot.
A piece of 304 SS IV tubing could be slide over the nail to act as a bearing/pulley.
You could cut short pieces and flare the ends to trap the line.

Or press bushings/bearings into the sidewalls and let the nail spin.

[www.samsonrope.com]


End Stop, Microswitch, Hall effect sensor:
@NewPerfection, Tks!

Nicholas said that he is using “off-the-shelf Repetier Host and Firmware”.
I would imagine that the end stop code is similar to Marlin?
Hall effect sensor are super cheap on ebay, but I don't know the cost of the supporting bits and pieces.

A2

Edited 4 time(s). Last edit at 09/18/2013 09:00PM by A2.