Tri-Polar Bot Concept -- Meet Wally

woo Wrote:
-------------------------------------------------------
> thats because when powdercoating powder bakes in
> owen on 170+ degrees celsious.so every moisture in
> wood will results with bad powdercoating.
>
> i have akso powdercoating on my working place, but
> we are dealing only with metals...

We worked with powder coating 1.25" MDF for desktops at one point and you can buy a lower temperature powder specifically for wood. There's also machining tricks you can do to minimize moisture offgassing, which causes pinholing in the edges.

owens Wrote:
-------------------------------------------------------
> @glennswest - This is based on my personal
> shopping list, which is completely UNOFFICIAL. ...
.
>
> Electronics: Azteeg X1 with SureStepr SD8825 on
> the beta, set up for 1/32 microstepping. Any
> controller should work fine, but if you use
> stepper drivers that only support 1/16 microsteps
> you'll have to keep track of the difference for
> your calibration values. One reason for choosing
> the X1 is because it can handle the 24 volt power
> supply and drive the bed heater; some other boards
> won't. If you run with an unheated bed or a 12
> volt heater that won't be an issue.

I am wondering if an Arduino Nano, Uno or Mega can be used instead of the Azteeg X1, and if so what would I need to use in addition to the board that I don't need with the Azteeg X1? I ask because I am trying to keep costs to a minimum if it doesn't add too much complexity. You can get an Arduino Nano, Uno or Mega for about $8.50, $9.50 or $16 whereas the Azteeg X1 is $70.

MightyMouth Wrote:
-------------------------------------------------------
>
> I am wondering if an Arduino Nano, Uno or Mega can
> be used instead of the Azteeg X1, and if so what
> would I need to use in addition to the board that
> I don't need with the Azteeg X1? I ask because I
> am trying to keep costs to a minimum if it doesn't
> add too much complexity. You can get an
> Arduino Nano, Uno or Mega for about $8.50, $9.50
> or $16 whereas the Azteeg X1 is $70.

The short answer is no - an Arduino by itself can't drive a 3D printer.

The long answer is that most of the printer controllers are based on Arduinos, with additional electronics to allow them to drive stepper motors, power the heaters in hotends and heated beds, and read the thermistors. For Wally, you have a number of choices:

- RAMPS, a shield that fits on an Arduino Mega and takes add-on stepper drivers
- Sanguinololu or Azteeg, Arduino-like boards with sockets for the stepper drivers to be installed,
- Printrboard, Brainwave, Melzi, RUMBA or RAMBo, all-in one boards with drivers soldered on,
- Smoothieboard or 4pi, advanced boards with more powerful CPUs.

The general difference is that you can save a little money by buying a board with the stepper drivers soldered on, but if you damage one the board becomes useless. There are other differences in terms of capability; some boards will only drive four steppers and some will do five or more, for example. Wally only needs four, so any of those boards will be fine.

Personally I like RAMPS, and I buy mine from RepRapDiscount because they sell a nice kit. I like the Azteeg X1s but I don't see that they really have any advantages over RAMPS. Printrboard is nicely integrated and a little bit less expensive but the firmware isn't in the mainstream distribution, so it's difficult to build a new version if you want to change something. I have a RAMBo but it's sitting on the shelf right now; I definitely wouldn't recommend it to a beginner, nor the Smoothieboard or 4pi.

One more point on costs - the $70 Azteeg X1 does *not* include the stepper drivers, which cost $11.50 or $12 each from Panucatt. You need four, so the total price ends up around $120.

Bill.

Thanks Owens, I think we will go with the arduino ramps combo.

Hey Nick
Can you post some of the dimensions on the .dxf files
I can't be sure they are getting into sketchup correctly

Tnx

Here are the outside dimensions if you are milling 12.5mm boards. Dimensions will change for different flatstock thicknesses.



Edited 2 time(s). Last edit at 10/04/2013 05:07PM by nicholas.seward.

I am thinking that the Big Blue Saw would do a good job cutting these (waterjet cutter). Has anybody used them? [bigbluesaw.com]

I am going to use the
Radial Arm Saw, Band saw, router, drill press
on some MDF.

Has manual woodcraft died?

The starter files were out there 22 Sept
Where is

BOM
Assembly drawing
missing parts

???????

My attempt at a BOM attached (excel file)

Getting a sketchup workout trying to put together assembly drawing!!!

Tnx

RussNelson Wrote:
-------------------------------------------------------
> I am thinking that the Big Blue Saw would do a
> good job cutting these (waterjet cutter). Has
> anybody used them? [bigbluesaw.com]


Looks like they offer many materials, but no wood thicker than a quarter inch. Wally needs half-inch, unless you're going to double everything up and laminate them yourself. . .

Here's the feedback from the laser cutter.
I've given them the dxf 1-to-1, and I've given them the outside diminsions, but they seem
to want everything in traditional format.

Even the Mirron Image, which should be the same on the left and right, there asking to verify.

Any Help?

Here is a bunch of detail. (Don't judge me. I just spewed dimensions everywhere. I hope I got it all right.) I went ahead and showed all the 1/4" boards you would need to laser cut. I used 1/2" but milled it. You can glue the matching boards together to get faux milled boards. (6mm flatstock will also work instead of the 1/4" but there will be .5mm mismatches on the edges but no one should be able to tell.)

Eventually, I will get around to making a true laser cut version that doesn't need the printed corner brackets. Too busy right now.

@nicholas.seward: Awesome, thanks for this! Any idea how soon the printed part files will be released?

any news about software?

@woo: This is my gcode pre or post processor depending on how you want to look at it. It isn't very pretty. I just got my Wally back from New York City yesterday so I haven't been able to do all the calibration tests that I want to verify that the code is accurate.

It will be quite a while until I get the host software, slicing, calibration, and coordinate transformation all in one easy to use package. Even without the this, it took me less than 1 minute to calibrate the z-level for Wally. I still need to do some XY calibration.

When is Wally's Beta release

Wally's like a movie star after the NYC MF Film festival


So are there CAD files for Wally?
Wally assembly drawings?
BOM?

I think you said Wally was in a CAD package?

How about all the parts? Drive drum?

---

A Wally BOM I created, text format
NOT official --- NOT from Nick

Qty Part Drawing Matl supplier

7 arm arm.stl PLA
1 xyforearm xyforearm.stl PLA
1 boardbracketleft boardbracketleft.stl PLA
1 boardbracketright boardbracketright.stl PLA
1 filamentdrivemount filamentdrivemount.stl PLA
1 spoolholder spoolholder.stl PLA
1 L_Bkt_1 boardparts.stl PLA
1 L_Bkt_2 boardparts.stl PLA
1 L_Corner_Bkt boardparts.stl PLA
1 R_Corner_Bkt boardparts.stl PLA
1 L_Motor_Bkt boardparts.stl PLA
1 R_Motor_Bkt boardparts.stl PLA
1 R_Tablet_Mnt boardparts.stl PLA
1 L_Tablet_Mnt boardparts.stl PLA
2 Drum_Drive PLA
2 Motor pulley XY
1 Motor pulley Z
1 board_back board_back.dxf MDF / Wood
1 board_bottom board_bottom.dxf MDF / Wood
1 board_left board_left.dxf MDF / Wood
1 board_right board_right.dxf MDF / Wood

Bed Surface black basalt plates

26 Bearing 608
2 Bearing 6702

Braided line Spectra fishing line

Electronics
4 Stepper Motor NEMA 17 40mm deep QU-BD

1 Power supply 24 volts at 14.6 amps
1 Azteeg X1 Printer controller
4 SureStepr SD8825 Stepper Driver
1 Bed Heater 200 mm square heater "QU-BD 200 mm round silicone,"

Extruder
1 Extruder -Bowden 1.75mm QU-BD
1 Hotend Printrbot Ubis

Endstop Z bed stop switch

Fasteners
14 M3-20
44 M3-16
1 M3-10
2 M3 NUT
2 M3-50
2 M3-5ish GRUB SCREW
2 M3-6
57 M3 NYLOC NUT
4 M4-12
13 M8 WASHERS
4 M8-55
3 M8-65
2 M8-70
2 M8-90
1 M8-105
4 M8-35
16 M8 NYLOC NUT

nicholas.seward Wrote:
-------------------------------------------------------
> Here is a bunch of detail. (Don't judge me. I
> just spewed dimensions everywhere. I hope I got
> it all right.) I went ahead and showed all the
> 1/4" boards you would need to laser cut. I used
> 1/2" but milled it. You can glue the matching
> boards together to get faux milled boards. (6mm
> flatstock will also work instead of the 1/4" but
> there will be .5mm mismatches on the edges but no
> one should be able to tell.)

One point that might not be clear if someone hasn't seen the original milled parts - the way the bed is set up. In the PDF it is shown as two pieces to be sandwiched, one of which has a large circular cutout and a small circle adjacent to it, and the other with just the small circle. It's intended to be used with the QU-BD 200 mm circular silicone heater, and the smaller circle is for the wires to feed out underneath the bed, and to give a little relief for the thicker part of the bed where the wires are connected. When the two parts are sandwiched, there will be a 6 mm deep hollow for the heater to sit in, which is only about 2 mm thick. That will mean you'll want to provide something to push it into contact with your printbed, or glue it onto the printbed with high-temp adhesive.

Therefore, if you're milling the parts out of 12.5 mm ply, as we are, you don't need to go the full 6 mm depth for the hotbed. However, you may find that you want to relieve the area just adjacent to the small circle, to make sure the thicker part of the bed isn't being squished. It's something easily done with a chisel or sharp knife, to fit the bed to the heater.

For those using plywood, also take note of the two 3.4 mm diameter holes on the left side of the bed; those are for mounting the arm that lifts the bed. The second hole isn't needed on the top piece if you're laser cutting, but both holes need to go through if you're milling, and the second screw will want to be a bit shorter to make up for the space you cut out to relieve the heater.

Instead of having to write a ton of complex math to be able to print a vertical object on a platform that moves back and forth in an arc as it goes up and down, why not take a look back in history to 1864?

Google Peaucellier-Lipkin linkage

Do some 3D spatial thinking and you should be able to visualize how to design the linkages wider, like independent rear suspension arms for a vehicle. Nest one inside the other with the ends at different levels so the build platform can't rock.

The end result is like a four bar parallel link that automatically adjusts the length of the bars to produce a perfectly straight motion.

Or if you want to be simpler and use fewer components, put the build platform on top of a Sarrus linkage.designed in 1853.

The real beauty of these two linkages is they'll *always work*, producing perfectly linear motion and no misplaced decimal point or out of range numbers will make them not work.

K-MODDL, a site with an answer (usually 100+ years old) to any mechanical motion question you may have. Before computers and stepper and servo motors and complex math to make things move in the desired directions - despite the mechanics of the machines, engineers and inventors designed a huge number of purely mechanical methods to get the motions they needed.

Nice, very interesting linkages. I've been enjoying the book "How Round is your Circle" which I discovered from a reference earlier in one of these forums. see howround.com
and shown in video at www.youtube.com/watch?v=270qEZKXAfQ

But when you consider how many bearings, or how much space, is required for them, doing small corrections in software seems to me an ingenious way to simplify things.
And, the fewer the joints and the simpler the mechanism, the less chance there is for play or wobble, so the quality is better.

Wally Sketchup

Here is a file of Wally kludged together in Sketchup-8
from released parts and drawings.

I could not post big file here so I put it over on buildlog forum.

Any Sketchup Xperts out there may want to fix it up.

Wally 3D printer
[www.buildlog.net]

Has anyone started on their own Wally build based on the already released files earlier in this thread, or are they waiting for the official ones?

Personally, I'd recommend waiting for Nicholas to release his files, for two reasons - he's a really good designer, and he is taking time to test the printer he's designed. If you want to build a working printer, that's the way to go. If you would like to design your own printer that uses similar mechanisms to Wally, you have the design skills (I sure don't), the spare time to figure out what vitamins will work with your parts, and the desire to test/tweak/reprint/test and repeat, then go for it. It's entirely possible you'll end up with something better than Wally. It's extremely unlikely that *I* would, however, so I'm sticking with Nicholas' design. . .

Galane Wrote:
-------------------------------------------------------
> Or if you want to be simpler and use fewer
> components, put the build platform on top of a
> Sarrus linkage.designed in 1853.

Or you could do this: [www.youtube.com]

> The real beauty of these two linkages is they'll
> *always work*, producing perfectly linear motion
> and no misplaced decimal point or out of range
> numbers will make them not work.

True, in the mathematically pure world of simulation. In the real world of printed components, inexpensive bearings and bolts, maybe not. Have you analyzed the variation in the output when there are errors in the parts used to build the mechanism? I certainly haven't, and wouldn't know how to, but that would be the first step in determining whether a mechanism was well-suited to this particular application. Unless you're going to pay to have someone machine them for you on a CNC mill, of course.

> K-MODDL, a site with an answer (usually 100+ years
> old) to any mechanical motion question you may
> have. Before computers and stepper and servo
> motors and complex math to make things move in the
> desired directions - despite the mechanics of the
> machines, engineers and inventors designed a huge
> number of purely mechanical methods to get the
> motions they needed.

Absolutely, many of the mechanisms they designed were inventive and elegant, and beautiful to watch. But now we have computers, and we need to use them to build 3D printers regardless of the elegance of our mechanisms. Sometimes it's better to take advantage of a simpler mechanical answer with a more complex bit of math behind it.

Go ahead and design your printer any way you want, though, and if it works the way you intend, the rest of us will build them too. . .

I saw a working Wally at NYC Maker Faire.

Why is that design not good enough to release drawings and parts?

Is there some design killing change that is in the works?

@Owen That Sarrus linkage in the youtube is beautiful
I just may want to use that lift table.

but what I have seen about the software ( and self leveling )
I like software solutions (if the common user can understand them)

Iching to buid a Wally

Wally would like the s-maker

[forums.reprap.org]

[make-it-3d.eu]

Did Wally or S-maker come out first?

Moving the extruder and leaving the weight stationary on the bed
seems real good design.

cozmicray Wrote:
-------------------------------------------------------
> Did Wally or S-maker come out first?

I first proposed the Wally configuration with a cardboard model around June 12th in an online brainstorming session with Nicholas and Billy Zelsnak:
[groups.google.com]

Since then, Nicholas has been the driving force behind the detailed design, design variations, and implementations.

The S-maker idea was about 3 months prior to that.

The S-maker has little in common with Wally though. It has much more in common with Morgan.

Wow! A lot has happened since last I looked.

@Galane: It may seem like we may be too dismissive to your ideas but let me say that they are very very welcome. People were very dismissive when I stated my goal to build a bot without rails. They were dismissive still as I started on Simpson's design. I hope you really dig in and find ways to improve Wally or to spawn a whole new bot.

I am a mechanism nerd. I particularly like straight-line mechanisms. It hurt me inside to abandon straight-line motion. I am hoping someone will come up with an elegant solution. I have plans to do some experiments with Sarrus Linkages as you can see from the video owens linked above. I am particularly fond of Hoeken's Linkage. Here is a robotics project that one of my students put together. (So proud.)

I should probably publish a lot more of my failed mock-ups. (I read in National Geographic that for every 1000 pictures that are taken for it they only publish 1. I would say that my ratio is more like 10 to 1.) I put this one together a long time ago but it was immediately obvious that it was just an over complicated space crank so I transitioned back to working on an elegant space crank.


I have been thinking about some combo of Robert's Mechanism and Chebyshev Linkage. However, I am at a lose to make them more elegant than the 4-bar setup I have now. (Actually, only 2-bars are really necessary. I am looking at that too.)


I have also looked at using this with DLCJ's instead of gears. In fact I let tigertalar know about how DLCJ's could be used here and told him that I thought we could use this to make a tripteron and he came back with this.

Blah blah blah. The moral of the story is we are looking high and low for a better way. If you go back to the original google forum, you will see we considered quite a few different designs. see3D was nice enough to distill from those ideas into what we now call Wally. I personally welcome new thought and new perspectives. Let me turn it back on you, can you see a path forward that will simplify the current Wally design or a radical shift in design that could lead to a simpler solution? I have no false beliefs that Wally is already optimized at birth.

@cozmicray: Between my day job, Wally, Simpson, and starting a new business I get very little sleep. (My priorities are wife, teaching, Wally, eating, sleep.) I could release the files that I have now but they wouldn't make sense to many. I am going over everything piece by piece to make sure they are ready for release into the wild. Beyond that I am making assembly diagrams and instructions. I apologize for not having the packet done by now but I think the wait will be worth it. Trust me when I say that I am very motivated to finalize the design. As soon as I release v1.0 of Wally and Simpson then I get to work on the Sextupteron, Quadmorgan, and a "secret" project I am dubbing Tilt-n-Twirl. I will tell you that Tilt-n-Twirl will have 4 degrees of freedom. (Side note: I used a hack saw on Wally in the hotel room the night before the NY Maker Faire. What I am saying is that it wouldn't hurt for me to make sure everything is just right.)

Edited 3 time(s). Last edit at 10/14/2013 01:23AM by nicholas.seward.

Wow
FIRM Mark II Part 2
I didn't know you were teaching post grads at MIT robotics lab
I thought you were a high school computer science teacher?
I jest --- some amazing students -- how many fathers and uncles really do the student work?

Forget about the tiltawirl , quadgodzilla, and zippity dodah and other projects for now
This is what KILLED trinitylabs A1 development.

[forums.reprap.org]

Anything I can do to help?

Is there an S-maker machine or kit to be had?

Is it open source?

Where can one get one?

Perhaps for you it not important and doesn't cost your attention! Yours the hand contains the WALLY logo. Why on "xy forearm " there is no logo? It after all too hand? I began the press, I think others too. While all didn't print - it is necessary to correct?

Edited 1 time(s). Last edit at 10/18/2013 11:15AM by Vladimir_Lukashuk.