McWire Successor

Yeah, you pay a premium for those breakout boards. They're good to prototype with, though.

---

-------------------------------------------------------

Hell, there are no rules here - we're trying to accomplish something.

Opportunity is missed by most people because it is dressed in overalls and looks like work.

Thomas A. Edison

How much attention has the micromark XY table gotten from the McWire community?

[www.micromark.com]

You're looking at around $110 after shipping for a turn key positioning system requiring only a gantry, electronics, and an extruder to get up an running. Add a dremel and it doubles as CNC router which has potential for replacing some of the dependance on commercial lasercutting outfits.

JohnnyCooper Wrote:
-------------------------------------------------------
> How much attention has the micromark XY table
> gotten from the McWire community?
>
> [www.micromark.com]
> MENT,8009.html
>
> You're looking at around $110 after shipping for a
> turn key positioning system requiring only a
> gantry, electronics, and an extruder to get up an
> running. Add a dremel and it doubles as CNC router
> which has potential for replacing some of the
> dependance on commercial lasercutting outfits.

First if you look at the specifications you see that you don't have sufficient Y direction movement to have a useful build volume.

Specifications (inches):
Table Size: 8-1/4 L x 2-3/4 D
Overall Size: 10-1/2 L x 7-1/2 D x 1-11/16 H
Base Size: 3-3/4 x 5-1/2
X Direction Travel: 5-1/4
Y Direction Travel: 1-3/4

Dial Graduations: .001; .050 per revolution

Also you would require some gearing to obtain the needed speed otherwise you only would get .4 inches per minute; and don't forget the z-axis and a platform to connect them.

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

Hey folks, I'd like to chime in here on my thoughts for McWire. I have been working on a McWire for a couple of months on weekends and evenings. I *just* got it to the point that it is printing anything at all, and I've noticed the slow threaded rod problem, too.

I was actually looking around for belt drive solutions when I came across this thread. I had an idea to use ball chain, but sourcing the gearing was a problem.

After a couple hours googling, I think I've hit on the right search combo: "Vertical Chain Control Gear", this is what's used in curtains and blinds when ball chain opens and closes the things. Depending on the type, these gears are $2-$4 each. I think these can be implemented easily to turn the McWire into a chain drive. I'm heading to the hardware store later to try to pick some up. This gives the benefit of bootstrapping the ballchain drive without needing to first print Vik's gears.


---More thoughts---

My intention with my McWire was minimizing total cost and special pieces. I also tried to source things in local shops to minimize shipping costs (which always seem to be neglected and which can be significant). McMaster is great, but for some things there's a minimum order quantity that for a single build, can often lead to severe cost penalties. Mouser also has a similar issue with shipping of electronics components that makes tinkering and experimenting really expensive.

I bought a set of high-torque steppers from alltronics' ebay store. These were Lin Engineering 4218L-01, in a set of five for about $8/each (including shipping).

Instead of the laser cut acrylic pieces, I used $2 polypropylene 8.5X11" cutting boards from WalMart. They work ok, but the Z-axis tends to sag under the vertical spring arms. More teflon runners in the middle would help. I was really reluctant to use laser cut acrylic because I don't know anyone who can do this for me (also to minimize cost).

One big problem I found with the McWire build info to date was the extruder. The whole assumption is that one doesn't have access to rp parts, but then the success of the whole thing relies on using the final mendel (or makerbot) extruder design to make it work. I ended up making a stepper driven pinch wheel based on binarycontruct's design. The Lin stepper seems to work well for extruding so far... Very noisy with the extruder board and a low step rate.

The nozzle I made is based on Renoir's idea to use MIG welding tips. I took a 0.5mm tip and drilled it out to 3.5mm with a drill press. Then I made the PTFE barrier way shorter (approx 30mm) because the threads of the welding tip are only about 10-15mm (also with a drill press). Next, I took the idea of makerbot's insulator retainer: I found an e-clip that fits right around the neck of the welding tip; it's very very thin. I can then grab it with a fender washer so that there's no load on the ptfe to pop the welding tip out.

Nichrome and kapton and insulation from makerbot, wrapped very close to the tip of the welding nozzle. 1mm thermistor. So far it seems to print PLA fine, but I've only got about 3 or so print-hours into it. Heats up to 220C in about 1 minute.

(big dump of info, I know. I'll try to get pictures of my setup. I've not really got into the habit of documentation yet.)

One more quick thought just for posterity: I used 1/2" galvanized steel pipe instead of the recommended 3/4" black steel, mainly to reduce cost by almost 50%. Same rigidity, no real need for 3/4" pipe, lower weight.

- Nik

Okay folks. A couple of things strike me from the conversation going on here, and I have a couple of thoughts to add as well:

I think we should stick with already utilized ideas within the community overall. I see the appeal of the ball+chain drive, and I have to say that I am intrigued by RepOlaRap. But in the case of both, there is only one or a few people doing it. I want to err on the side of "Lots of people have done this."

I think $500 is a decent budget for this, but I think we should also focus on *just* the cartesian bot. I say this because Electronics are going to fluctuate on top of printer design and worrying about electronics along with designing a new printer adds a lot for us to deal with. If we need to assume a budget, I recommend we drop $100-$200 for the Electronics and shoot for a Cartesian bot that fits within the $300 range. (Or whatever price ceiling we figure is appropriate)

I crunched a lot of numbers to figure out the cost of selling MakerBot's Electronics. At the old price, in a fairly decent volume (around ~50 boards assembled) they were cutting *very* close, labor included They were making a very poor profit. The adjustment in price isn't great for us, but I see the reason they did it.

I'm in favor of an MDF structure, The "L" shape being my favorite. I think it's pretty easily built/designed. I think a board + angled brackets would work very well, including round steel rod + bearings for travel. The rails on the current McWire are clever, but I don't think the price of a genuine rail + bearing system is that much more costly.

sircastor Wrote:
> I think we should stick with already utilized
> ideas within the community overall. I see the
> appeal of the ball+chain drive, and I have to say
> that I am intrigued by RepOlaRap. But in the case
> of both, there is only one or a few people doing
> it. I want to err on the side of "Lots of people
> have done this."

Everything has a beginning. Certainly, if you don't want to wait for one of the new developments, or until 3D printing is common place, trying to use an existing repstrap is certainly a good thing. Only thing being, it seems there is great frustration with the options currently available.

> I say this because Electronics are going to
> fluctuate on top of printer design and worrying
> about electronics along with designing a new
> printer adds a lot for us to deal with. If we need
> to assume a budget, I recommend we drop $100-$200
> for the Electronics and shoot for a Cartesian bot
> that fits within the $300 range. (Or whatever
> price ceiling we figure is appropriate)

That is my feeling too. In fact, for RepolaRap, I dumbed down the firmware, per suggestions from NopHead, to the point where it will probably even fit on an 8K Program, 1K RAM microcontroller (At least, for amtel 8 bit cpu, if I don't need encoder logic for the XY angle feedback; 16K would be fine if encoder feedback will be required.) Anything you can get a C compiler to run on with at least this much power would work, as long as it has enough IO ports.


> I'm in favor of an MDF structure, The "L" shape
> being my favorite. I think it's pretty easily
> built/designed.

If you have a good cheap easy to build design for just a Z-Axis, let me know. I'm still thinking I'll just use a drawer slide, a few MDF panels, and a threaded rod drive. This is all I need beyond the current mechanical structure for RepolaRap. I am still uncertain about the encoder electronic requirements; and of course, since I traded mechanical simplicity for software complexity, it will take some time for me to work thru all the math on the host side.

> I think a board + angled brackets
> would work very well, including round steel rod +
> bearings for travel. The rails on the current
> McWire are clever, but I don't think the price of
> a genuine rail + bearing system is that much more
> costly.

Indeed. When I was evaluating the different designs, it came down to getting good and accurate linear bearings. You trade off ease of programming (brain dead simple for cartesian X+Y) for ease of building (no linear bearings needed at all, just a few pieces of MDF, some thrust bearings, and a radial bearing, but now you have to do fancy inverse kinematics, and be able to drive two motors non-linearly to get linear output.)

I wouldn't dismiss the alternatives too early; I suspect all current designs will very likely be obsolete in 3 or 4 years anyway (they'll be great to bootstrap to the RepRap version 4 or 5 though...)

BeagleFury Wrote:
-------------------------------------------------------

> Everything has a beginning. Certainly, if you
> don't want to wait for one of the new
> developments, or until 3D printing is common
> place, trying to use an existing repstrap is
> certainly a good thing. Only thing being, it
> seems there is great frustration with the options
> currently available.

The reason I want to go this direction has to do with commonality. RepOlaRap is awesome, but one thing it lacks at the moment is a significant amount of experience within the community. Although our new bot will be new, it will also (I think) similar to McWire/Darwin/Mendel in terms of form. It's advantageous to be working from similar concepts.

> If you have a good cheap easy to build design for
> just a Z-Axis, let me know. I'm still thinking
> I'll just use a drawer slide, a few MDF panels,
> and a threaded rod drive. This is all I need
> beyond the current mechanical structure for
> RepolaRap. I am still uncertain about the encoder
> electronic requirements; and of course, since I
> traded mechanical simplicity for software
> complexity, it will take some time for me to work
> thru all the math on the host side.
>
> > I think a board + angled brackets
> > would work very well, including round steel rod
> +
> > bearings for travel. The rails on the current
> > McWire are clever, but I don't think the price
> of
> > a genuine rail + bearing system is that much
> more
> > costly.
>
> Indeed. When I was evaluating the different
> designs, it came down to getting good and accurate
> linear bearings. You trade off ease of
> programming (brain dead simple for cartesian X+Y)
> for ease of building (no linear bearings needed at
> all, just a few pieces of MDF, some thrust
> bearings, and a radial bearing, but now you have
> to do fancy inverse kinematics, and be able to
> drive two motors non-linearly to get linear
> output.)

I've been curious regarding drawer slides if they're really reliable. The ones I've played with (mostly in homes, attached to drawers) have an awful lot of play in them. Maybe that can be tightened up, or is a product of poor installation. The rails on BotHacker's machine are what I envision as the best option. They're adjustable on both sides, so if you find yourself traveling at an angle, you can adjust them. Properly assembled, I imagine you'd face no problems travel-wise.

> I wouldn't dismiss the alternatives too early; I
> suspect all current designs will very likely be
> obsolete in 3 or 4 years anyway (they'll be great
> to bootstrap to the RepRap version 4 or 5
> though...)

I wouldn't expect a design to last that long as the rate of improvement increases. The McWire has been around for about 2 years, and it's weaknesses have been toiled over. Hopefully we can make a better machine, and maybe in a couple of years someone will say "Hey, we should make a better repstrap" Of course, by then plutonium reprap parts will be available at every corner drugstore.

Here is my hat toss into the ring for the McWire Successor. It is still very early in the design stage but I am hopeful that it will work or be a stepping stone to something that will.

http://replibot.blogspot.com/2010/03/introducing-new-repstrap-design-auger.html

---

B^2 : [replibot.blogspot.com]

~~ We Are The Factory ~~

sircastor Wrote:
> > Everything has a beginning. Certainly, if you
> > don't want to wait for one of the new
> > developments, or until 3D printing is common
> > place, trying to use an existing repstrap is
> > certainly a good thing. Only thing being, it
> > seems there is great frustration with the
> options
> > currently available.
>
> The reason I want to go this direction has to do
> with commonality. RepOlaRap is awesome, but one
> thing it lacks at the moment is a significant
> amount of experience within the community.

I expect that to change very quickly once someone gets an easy to build, inexpensive, and well documented repstrap. I suspect people will gravitate to the 'easy and cheap' when made available.

This isn't to suggest others shouldn't pursue the approach you suggest, afterall, it sounds you have the same goal as me -- provide the community with a cheap easy to build printer. McGuire and such has about 2 years lead time, but still seems to have a lot of issues preventing others from successfully using them. Wooden mendels, from my personal experience, cost about as much as mendel, but have a large tooling cost or large time investment cost.

> Although our new bot will be new, it will also (I
> think) similar to McWire/Darwin/Mendel in terms of
> form. It's advantageous to be working from similar
> concepts.

The RepolaRap software design actually could have more in common than my current approach; I have more fun trying the spline approach to improve performance, when really, I probably could just plug the inverse kinematics into the current Mendel firmware, calibrate the polar constants, and use everything standard except for the inverse kinematic patch. I might do this if I'm not able to progress further along my current strategy of spline programming.

As far as hardware differences, I'd say so far that the RepolaRap design has the advantage, but then, I'm biased. ;-)

> I've been curious regarding drawer slides if
> they're really reliable.

You might want to see if you can find the author of a blog post detailing a Drawer Slide RepStrap

As this was 4 years ago, it would seem that the design either lacks good build documentation, or ran into issues that were difficult to resolve.

bjbsquared Wrote:
-------------------------------------------------------
> [replibot.blogspot.com]
> ew-repstrap-design-auger.html

Wow! This is pretty awesome. I especially like your bearing system. By being based around a pipe, the bearings correct by virtue of just being there. That's very cool.
I'm not big on the DIY leadscrew, I think because it feels like a bit of a project itself. On the other hand, it's a solution to a problem we've all been tossing around, which is more than I can say I've offered. Have you figured out a good mounting method for attaching the motor? In a similar vein, I actually thought once of attaching a two-foot-long drill bit as a leadscrew for the sake of getting better travel. I don't know how that would've worked...

Hi,

ups someone the same idea like me

-Wolfgang

The author was James Wilkins and it looks like he abandoned it and bought a Makerbot: [blog.notzero.net].

I have played with some drawer slides. They are stiff in one direction but a lot of play in the other, so you need to use two mounted at right angles to each other to make a tight axis.

---

[www.hydraraptor.blogspot.com]

hi,
nophead, for a perfectionist like you is not enough... i know

but, to print Mendel it is enough, and a cheap and easy way.

-Wolfgang

Stoffel:

How has your performance been with the drawer slides? Do you have any other method of support for the stages beyond the slides, or are they sufficient enough? I ask because I think of the kitchen drawers I've used and how they start to hang down when fully extended. I imagine the slides made these days are a bit more solid, but I'm curious. Do you have a build log of your RepStrap? I like the design quite a bit, and I'd like to know more about it.

My english is not good enough to tell stories :-)


I ask because I think of the
> kitchen drawers I've used and how they start to
> hang down when fully extended.

i dont use them if they are fully extended , only a way of 150mm


> I imagine the
> slides made these days are a bit more solid, but
> I'm curious.

They are made of metal (iron), they are stable enough to extrude, but not to drill or Engrave.. when I made my Mendel parts, it is degraded (into Trash)


> Do you have a build log of your
> RepStrap?

no, but if i have Time i will make one (but only in German, sorry) my English ist TOOO bad.

Greeting from Cologne

Wolfgang

Tell story in your own tongue, that's why god founded Google.

---

repraplogphase.blogspot.com

Google tranlations...the hell in letters

> Do you have a build log of your
> RepStrap?

no, but if i have Time i will make one (but only in German, sorry) my English ist TOOO bad.

No problem, the wiki software understands German; would you like to write up your notes, and upload your drawings and photos to:
[objects.reprap.org]
using
[objects.reprap.org]
as an example?

And then some volunteer can do an English translation on another page?

We need a good simple RepStrap that works well and is easy to make.

@Stoffel15...Your printed gears and mendel parts look good.

How did you make your first geared extruder? I'm assuming that was needed to keep up with the slow speed of the threaded rod.

Hi Wolfgang,

... i'm inserting/uploading my pictures by editing the wiki-text, embedding an 'image-anchor' (the image-button in the toolbox), editing the name and clicking preview ...

In the prewiew-window is then a red link with the image-name - by clicking on the link i open a new window where i can upload the image from my PC.

Then edit the image-anchor wiki-text - e.g. [[Imageilot-laser.jpg|200px|Pilot-Laser]] for size and hint-text ... look here for some more examples ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Arvin
> @Stoffel15...Your printed gears and mendel parts
> look good.
They are good, but i have the same problem like all... bigger things warp.

>
> How did you make your first geared extruder? I'm
> assuming that was needed to keep up with the slow
> speed of the threaded rod.

My first was from Wood works but to much trouble...

i put a Picture on the "Wulfstrap"(Wulfs-trap ) Page of my secound one it was better. But now i use the Wade-Extruder. It is very good.

Greetings

Wolfgang

Edited 1 time(s). Last edit at 03/03/2010 07:50AM by Stoffel15.

sircastor Wrote:
-------------------------------------------------------
> bjbsquared Wrote:
> --------------------------------------------------

> I'm not big on the DIY leadscrew, I think because
> it feels like a bit of a project itself. On the
> other hand, it's a solution to a problem we've all
> been tossing around, which is more than I can say
> I've offered.

There's an easier way to produce the auger leadscrew. Wrap a wide strip and a narrow strip together helically around a core. Glue/screw/fasten/whatever the wider strip, when that's secure, remove the narrow strip and there's your groove.

... in combination with (laser-) welding/fusing the strip to the core this could be a good idea for DIY-leadscrews

And with the same methode you can make the corresponding nuts too ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

I am just starting to build a McWire to repstrap a Mendel. The assembly instructions and pictures are helpful, (although no mention is made of how thick the stage plastic sheets need to be if laser cut pieces are not currently available from RRRR), but there seems to be almost nothing about the electronics, software, and extruder. What do you use if the current supply of Mendel boards is out of stock at Makerbots? I made a guess and got the Adruino Mega and four Easy Driver boards from SParkfun. I have no idea if this will be enough to control the McWire, and if the existing software will work with this hardware. I do have experience programming Atmel chips in C, but this will be my first experience with an Adruino. I think the stepper driver cards can handle the current of the NEMA 17 stepper motors. I also got them from Sparkfun, although I bought four earlier for the eventual Mendel from ebay for about $50 for all four.

Mike

Mike,

The acrylic for the McWire is all .220" (sometimes called 1/4"). The pieces on the Z axis are doubled up to give it extra strength.

I built a McWire using the instructions. I can answer any questions you have about building it. Mine works fine, a little slow, but it does work.

That said, I changed a lot of the stuff I bought when I first built it.

For an extruder I am using the Makerbot extruder. Its the one part of the machine giving me the most trouble. If you can find an alternative go for it.

For electronics I built my own set of the standard electronics. I have a reflow oven and got all the SMT parts and did them myself.

I am not sure though, after having some experience with it, that the standard electronics are the best for the McWire.

For example I switched my stepper motors with motors from Alltronics and I swapped my X and Y stepper drivers with Unipolar drivers from Probotix and now my machine runs twice as fast.

My suggestion for folks doing a RepStrap and specifically a McWire, is don't get hung up on thinking that you 'have' to use the electronics from Makerbot.

I am glad to help anyone attempting a McWire build in any way I can. Just ask.

Finally, if you are mechanically inclined enough to come up with a belt driven system, well, do it. My McWire is ok but I would love it if it had the same speed as the belt driven units.

And Its not about not having the patience to wait for the thing to print. Its just that the longer it runs the higher the chance something will go wrong.

-- Chris

I'm busy playing with hardware at the moment, and I had something different in the kitty: a board that I bought off ebay before I found reprap from a Canadian company called SOC machines . It's called MC433G and it has similar features to the reprap electronics. Onboard g-code processing, four stepper drivers and provision for other add-ons like relays, joystick, USB interface and what not. Drawbacks are that it's unipolar and doesn't microstep, although my steppers are 400 steps-per-rev, but it's tidy and assembled and cost about USD 140. Atmel based.

Thanks Krafter, murd. When the rest of the parts arrive, if I have trouble I will ask. I guessed on the acrylic and ordered 3/8" polycarbonate on ebay. It was about the same price as acrylic!

I will look into belt or wire cable based movement, as I have more than the standard tools in my garage, though not a metal smith by a long shot.

Mike

---

Team Open Air
Blog Team Open Air
rocket scientists think LIGHTYEARS outside the box!

Quote
rocket_scientist
I made a guess and got the Adruino Mega and four Easy Driver boards from SParkfun

Here is an example of those electronics in use. I bought easy drivers too. You have to choose higher voltage motors to get the same torque from 750ma though.

The stepper motors and easy drivers boards arrived today. The steppers are 200 steps per revolution, and the easy drivers default to 8 micro steps per step, so that is 1600 microsteps per revolution. to Keep 1/1000 of an inch control, that means the drive pulley needs needs to be 1600/1000 inches in circumference, or 1.6 inches, or 0.56 inches in diameter. For 0.025 mm precision, it means 1.27 cm in diameter. So whether I use a belt drive or a chain drive, to keep the resolution up, I will almost need to gear down the steppers so that I can use a larger diameter drive pulley. Since this is a low cost, low tech design, I will work with tiny pulleys. A 1/2" aluminum rod, covered with silicone RTV or other rubbery material might work well. Or put a 3/4 inch rod into the drill press and try to make track for the pulley with sides. I will report back when I get closer.

Mike

Edited 1 time(s). Last edit at 03/04/2010 01:04PM by rocket_scientist.

---

Team Open Air
Blog Team Open Air
rocket scientists think LIGHTYEARS outside the box!