McWire Successor

After coming across a number of people and a number of posts feeling frustrated with the McWire, I felt like there should be a single thread to discuss the problem, and how to deal with it. In addition to this thread, I've created a wiki page:

Development: McWire Successor

Any and all discussion about a replacement robot are welcome, even if it restates stuff from other forums and threads.

I have a couple of things to get us started:

-The robot should be made from materials that can be purchased locally.
-The robot should be able to be made with a minimum number of tools.

What are all the problems that the McWire suffers from? What are the advantages it has?

(Edit: I renamed this thread, and then reverted the name change when I realized it was silly. -Sebastien)

Edited 2 time(s). Last edit at 06/22/2010 01:45AM by SebastienBailard.

I really don't think local is important, I think if they are available from mcmaster (or other major distributors) in the US (no idea for europe) that should be fine. Mcmaster has next day UPS ground shipping to at least 30% of the US (place by 7pm you will have it next day by 4pm for cost of regular ground, mcmaster has its own fleet of trucks I think). That would also open up about 300,000 more parts that home depot and lowes don't carry...

For tools I think a circular saw or table saw and a drill press are reasonable tools to require.

Threaded rod drive for the X and Y axes is a big issue, it moves too slowly for the pinch wheel extruder and even if that is resolved it makes the print time at least 200 hours. I think belt or rack and pinion is needed.

Tim/Bothacker seems really close on his, bothacker.com if we can just get him to list the parts.

We've got some decent-sounding RepStraps coming online soon.

[objects.reprap.org]
Tim said he'd document according to an email exchange ~1 month ago, but I think he's having fun dreaming up new and better TimBots. I'll remind him.

[objects.reprap.org]
Sam said he'd document, according to an email exchange 2-3 days ago.

[objects.reprap.org]
Being developed and documented.

[objects.reprap.org]
Being developed and documented.

[objects.reprap.org]
Being developed and documented.

[objects.reprap.org]
Stage: concept drawing. I need to dig out my scanner.

[objects.reprap.org]
Stage: concept drawing on the wiki page.

---

I'm highly in favor of people who use mcmaster documenting their parts list.
But mcmaster only ships to people in the us. Long story, and a topic-derail.

Also, people who design and build in metric get to document in metric, likewise,
build in imperial -> document in imperial.

I even have seen folk say "I'll only build what I can source at my
hardware store". Frankly, let's ignore them until hardware
stores sell belt, rack and pinion, or stepper motors.

>
> Tim/Bothacker seems really close on his,
> bothacker.com if we can just get him to list the
> parts.

I really like the BotHacker design and think I would like to build one when the information becomes available. However, I think the price will be high because of the 80/20.
I am in hot pursuit of a less expensive McWire alternative I have the X and Y stage designed in Sketchup and built. I am working on the drive mechanism which I think is going to work. It is a little "out of the box" so I thought I would hold off until revealing my ideas. I have prototyped it in wood and after a few attempts came out pretty darn good. I just got some raw materials from McMaster Carr last night so this weekend I will be attempting to make the final X and Y drive. After my build efforts I will reveal it to the community (be it good, bad or otherwise) as I will be at a point for some input by all the great minds around here.

I will be reusing the Z-stage from my McWire to prove out the X and Y and then, once that is under control, I will redesign that using similar contruction methods for it.

---

B^2 : [replibot.blogspot.com]

~~ We Are The Factory ~~

I really like the BotHacker design and think I would like to build one when the information becomes available. However, I think the price will be high because of the 80/20.

I think the photos are good, and the build quality is stunning on the 8020 and the square tube, but he's giving up soooo much build volume by using a 2-axis moving table. I think he should look at moving his extruder around more, or make a much longer table.

After my build efforts I will reveal it to the community
prospective wiki page:
[objects.reprap.org]
using the example page as a template?
[objects.reprap.org]

Sorry. Reflex.

---

-Sebastien, RepRap.org library gnome.

---

Remember, you're all RepRap developers (once you've joined the super-secret developer mailing list), and the wiki, RepRap.org, [reprap.org] is for everyone and everything!

I will start with a disclaimer, I have not yet built the Mcwire cartesian bot as I am still griefing with electronics and software but I do have some opinions.

From reading the instructions, forums posts and looking at other builds I feel that Mcwire has many good points and one major bad point. The Mciwre structure seems cheap and easy to build with a large margin for error and ample opportunity for substitution. A reasonable solution might be to design a drive upgrade for Mcwire and have it published on the wiki. I am thinking either belts, ball chain, or rack and pinion. I personally like ball chain as it is cheap and easy to get. I have some ideas on how to make a ball chain gear by hand and will post about it if I make some progress.

Here are some links to video of Mciwres that have been upgraded.
belted Mcwire 1
belted Mcwire 2
bleted Mcwire 3

They seem capable.

Edited 2 time(s). Last edit at 02/26/2010 12:11PM by spaztik.

After my build efforts I will reveal it to the community
bjbsquared That's just Cruel! Not even a hint like the page from the McMaster catalog of a significant part? Come on!

I started a build of a McWire and when the lower stage became warped from having the screws holes on the pipe be off the top surface I just stuck it all up on top of the cabinet and went back to the original one I had started 3 years ago but rebuilt it bigger like [buildyourcnc.com] and then discovered the too slow problem and the high price of belts and pulleys put me on hold again.

Now Vic has me wondering if ball chain would work but I don't want to get the pokono laser cut pieces he recommends.

I'm not being cheap. I just want to come up with a design for an "entry level" machine that is cheap!

Edited 1 time(s). Last edit at 02/26/2010 11:33AM by Arvin.

I don't think a RepStrap using 80/20 and or parts from McMaster should be the official recommendation. They are only available in the US and they make the price of the repstrap more than the Mendel. Fine if you prefer you time over money, but not good for people with the opposite priority. To some extent the McWire is also US centric because the steel pipes are not common in the UK. Rather than a steel frame I would make an MDF box. In the UK you can get MDF cut to size, so you can make a frame like HydraRaptor's for a few pounds. Much easier than drilling holes in steel pipes.

A McWire can be made to work if it has a closed loop extruder. An open loop DC motor can't go slow enough and doesn't produce very good results even on a MakerBot. A stepper based extruder or a DC motor and shaft encoder are the minimum requirements for an extruder IMHO and if you want results like mine, a geared stepper is the way to go.

With the right motors and an acceleration scheme like I use I think it will do at least 8mm/s and I have seen somebody quote that they got 8mm/s. That is only half the speed that HydraRaptor ran at when it printed two Darwins, so I think it is reasonable performance from a cheap RepStrap. You print a Mendel with it and move on.

If 200 hours of watching it is more than you can stomach I think you are going to be disappointed with FDM in general.

But for the impatient a more fine grained bootstrap process could be used: once you build your McWire you slowly build parts to make it go faster. You can print Mendal style pulleys and buy belt. You can print Vik's ball chain pulleys and buy ball chain or you can print Forrest's rack and pinion and buy nothing. All these parts and a few brackets to mount them will not take too long to print, even slowly. Then you can print a Mendel faster than I can.

It would be an interesting investigation if somebody local to me has a completed McWire to hook it up to my electronics and firmware and see what it is actually capable off. I can't see any reason why it would not be far from what HydraRaptor achieves.

Another aspect to booting is that McWire was designed for light milling, so it should be possible to use it to make its own extruder, as I did with HydraRaptor.

---

[www.hydraraptor.blogspot.com]

I personally think that the way forward is to reduce the Mendel STLs to an equivalent DXF file of laser-cut parts. Hand out the DXF and let people get the parts cut at their local laser cutter shop. That gets them their Mendel built right off the bat.

Once they have it built they're going to need to start printing replacement parts. This is exactly what is happening with the Rapman 3 series.

[www.bitsfrombytes.com]

The printed parts are more robust than the acrylic. This gets users into the habit of printing spares and into making new parts sets.

The other thing that needs to happen is for Ian at BitsFromBytes to finish converting his Rapman MCU to a more general MCU usable by Mendel builders. He said that he would do that as soon as he gets his professional model printer into production.

The Rapman MCU is a much simpler to use control system than the mainstream Reprap electronics. It's also completely stand-alone. Having two alternative MCU systems for Mendel is an unalloyed good, imo.

---

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

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

Although I'm impressed by the t-slot construction, and it seems to offer a really simple platform for expansion, I lean towards Nophead's suggestion of building on MDF. It's easy to come by, and easy to work with. One of my original designs focused on replacing the pipe base with MDF, but keeping the neck and simply screwing it to an MDF base with another flange.

T-slot is relatively expensive. It's a balance between ease of construction and cost of materials. I imagine more people have time than money.

Re: Tools
I think a table saw and a drill press are not common enough to expect these to be in every home. I'm inclined to lean towards: Hand Saw, Drill, Measuring Tape, vise or clamps, etc. Tools that, if you don't have them, you're not spending a significant amount of money to get them.

If there's an easy way to, I think I'd like the machine to have a reliable means of correct positioning. One of my concerns with the McWire is that the x platform rails rely on correct movement based on the holes you drill in the pipe. This is pretty easy to get wrong IMO.

I think it is a complete fallacy to compute things as only time or money, time and money are equal. Certainly in different ratios for different people. I think a rough budget needs to be set which include hard costs as well as time-money. The most efficient solution that is proposed taking into account both those values should win out.

To me assuming a value of $10 US or $7EU per hour makes reasonable sense. Anyone that has the skills and equipment to build a mcwire could also easily obtain projects paying that rate for the equivalent duration of time.

I propose a budget of $500 US (including both time/money) . Assuming some parts are reusable (electronics, motors, extruder perhaps) that would permit the building of the mendel as well for under $1000 US total. Clearly makerbot has shown there is no shortage of folks willing to spend $1000 .

Lasercutting really does NOT seem like a solution to me. Ponoko is a good service but now has queues going well over 30 days and in my experience has had some tolerance issues. If RRRF is going to stock a lot of inventory than that is fine but it doesn't seem like that is possible. Local lasercutting services are generally quite expensive or very difficult to work with for designs like reprap, they just don't know what to do with it and are used to seeing display stands and boxes. Most that I looked at also have a min order or sheet fees that made them quite expensive.

On 8020 costs, anyone can go to flomoco.com and instant quote parts/lengths of parts/etc. I am a very frugal person and I can't see why everyone thinks it is so expensive. I have put way more into my mcwire already than if I just started with 8020 to begin with. The time-money value of a setup that offers cutting and drilling for a nominal fee I think is extremely compelling. Those people with more time can order the raw parts and cut/drill them, those with more money can order the pre-cut pre-drilled.

I think if the mcwire successor is not going to suffer from all the same problems as the mcwire these items have to be resolve first prior to developing the design:

1) Will there be 2 different US and Euro solutions, or just 1 that works in both. I think costs will be cheaper for both if there are 2 distinct designs, make similar where possible, diverge where sensible.

2) Set a firm time-money budget of $500US/$350EU with an hour cost of $10US/$7EU . Hopefully much can be reused.

3) Minimum tools. It is true that not everyone has a table/circular saw and drill press. They can be purchased for under $100 combined at least in the US. I do think that everyone knows someone with a drill press and table/circular saw. Someone could get by with a hand power drill, but I think that should be the exception. If it has been expected in the past that people needed to know someone with a 3D printer that I can't possibly see how this is unreasonable.

goinreverse Wrote:
-------------------------------------------------------
>
> Lasercutting really does NOT seem like a solution
> to me. Ponoko is a good service but now has queues
> going well over 30 days and in my experience has
> had some tolerance issues.
>

I didn't suggest using Ponoko. I suggested distributing the DXF and letting people use local laser cutting shops.

---

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

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

once you build your McWire you slowly build parts to make it go faster. You can print Mendal style pulleys and buy belt. You can print Vik's ball chain pulleys and buy ball chain or you can print Forrest's rack and pinion and buy nothing

That is a very good solution and one that I am seriously considering myself, but I won't lie, I'm cheap and the more parts I can avoid buying (especially temporary ones) the better. A solution to build a belted Mcwire with no printed parts would be best imho, but I suppose you could buy only one threaded rod and build Mcwire with smaller print area just to print the parts required to convert to belts/ball chain or rack and pinion. At least that way you save a little money, because as the saying goes "save a little and save it often".

I personally think that the way forward is to reduce the Mendel STLs to an equivalent DXF file of laser-cut parts.

I feel this solution is better suited to some than others. My understanding is that lasercutting, while not terribly expensive, is also not cheap either. I think the cheaper that a restrap can be built the more people will be encouraged to try. Personally I don't mind spending a few hundred dollars on project I am not certain I can complete but once it starts getting over $500 I'd have to think twice. As well I am sure that there are places in this world where lasercutting is unavailable. I would like to see Mcwire modified to be a capable 3d printer then people could camp there for a few months/years while they get their feet wet with 3d printing. Then when they decide they want some feature that is unavailable on Mcwire or that they need another 3d printer for whatever reason they can print and build a mendel.

One of my concerns with the McWire is that the x platform rails rely on correct movement based on the holes you drill in the pipe. This is pretty easy to get wrong IMO.

Please, someone correct me if I am wrong but as I understand it the nature of the Mcwire design helps to mitigate errors made in measuring, cutting and drilling holes.

@Forest, in the same paragraph I also identified the problems with local laser cutting services. When I tried to get the existing mcwire parts laser cut my options were all over $150.00 and with 4-20 days lead time. I also had little confidence in the result as almost all of the shops assumed thousandths even thought the DXF was clearly in metric, or were not able to read the DXF, it was a nightmare in general. One mis-step and the whole budget for someone would be blown as this services have no refunds or anything. This was across 7 shops in the US West region.

@nophead, I think the idea of bootstrapping the bootstrap is a perfectly reasonable one and a good way to reduce costs. It just has to be clear enough, well documented and pass the common sense test in terms of time-money.

Just to reiterate by $500 I meant perhaps $300 materials and 20 hours labor or $400 materials and 10 hours labor or $100 materials and 40 hours labor.

@goinreverse Laser cutting is rather expensive, granted, and if you don't keep your eye on the ball it can certainly get away from you. Having Mendel parts printed out at a commercial prototyping shop costs many times what laser cutting does, however.

It also costs a damned sight less than buying a Makerbot or a Rapman and does not have the craziness of having to cobble together an entirely different machine like the McWire which is ill-suited to printing just to get into the game.

---

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

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

I don't want to overpromise, but the structural parts of RepolaRap XY platform so far have cost me about US$70:

$5.09 neoprene .2" pitch timing belt 1/4" x 37".
$6.99 neoprene .2" pitch timing belt 1/4" x 57".
$14.80 2x acetal pulley, 10 tooth.
$7.66 100 5mm alloy steel metric balls
~$8.00 5/8" 3'x4' mdf panel
~$5.00 1/8" 3'x4' hardboard panel
~$3.00 1x3"x24" popular lumber
~$6.00 2x bearings
- 608 bearings should work; I just bought something from a local hardware store.
$10.00 various bolts, nuts, washers and screws. Hardware needs really are quite minimal (so far).

I haven't created a more complete BOM yet, because I haven't built the Zaxis, and I'm still making sure I can drive XY in software. Building the XY structure took me about 4 hours total using power tools. The most difficult part is to cut accurate circles in the MDF for the two build platforms, required because they act as pulleys. Everything else just requires careful measurement and a steady hand with a power drill.

Once I finish with software, I'm going to evaluate a bootstrap phase that won't require encoders -- by printing gears to mount to underside of each platform -- I'm hoping I can measure tolerences accurately enough, and use slower motion to avoid belt slip. This removes any possibility of belt/platform radius innaccuracy, and slip. If that doesn't work, I'll forge on ahead and attach encoders and use closed loop control.

For Z axis, anything that can be mounted to the platform should work; I haven't purchased the parts needed here, however, from my understanding, I should be able to get away with a threaded rod drive; I expect total cost there to be less than $40. And I'm guessing another $40 or so for the extruder itself?

Electronics will be a bit more than average if encoders are necessary for the polar angle feedback (A reprapped version using gears would not need this), but other than that, anything that works for Mendel would work for this platform. Initially, I'm using makerbot electronics I had purchased for Mendel.

I'm still working on host software to convert GCode to a form the polar firmware can consume. Splines have been awesome in software, but getting them to fit the jerk/accelleration/velocity profiles to limit vibration and improve axis speed has been a challenging proposition -- if anyone has experience with this, I'd love to hear ideas; I plan on working on it another few days before I give up and try a 'slow' mode test run.

I'm hoping to have something others can build in a few months or so; as this is a primarily weekends only project, it will take a while longer to get it to alpha stage.

@Forrest, Rereading your posts now I mis-understood originally that you meant skip the whole bootstrap to begin with and just build a mendel directly from laser cut parts. I am completely in favor of that if possible. I think that would change the equation in terms of laser cutting costs and the hassle being worthwhile. Is there any good tool that exists to slice stls into DXF either commercial or open source? I guess the tricky part is that contours (screw holes, etc) would have to be flattened into 2D slices or etched/marked for later drilling.

> 2) Set a firm time-money budget of $500US/$350EU
> with an hour cost of $10US/$7EU . Hopefully much
> can be reused.

Your costs are too low. Mendel (and even micro mendel), which was designed to be made out of as much RP as possible, costs 395EU. A RepStrap, which substitutes some other Material for plastic will HAVE to be more expensive or at the very least similarly priced. You can't avoid some costs.

Electronics $150-$220
Motors $60-$90
8 bearings $10
1-2 sticks threaded $10-40
1-2 sticks smooth $10-$50
Power Supply $25
Extruder set $15
Fasteners $50

Don't think you will manage muhc cheaper than this, no matter what way you do the build (unless the Rapola works). So that puts your BASE cost at $320, for JUST the mechanicals. $120 for your support materials, and fab is not that much.

---

repraplogphase.blogspot.com

goinreverse Wrote:
-------------------------------------------------------
> I think
> that would change the equation in terms of laser
> cutting costs and the hassle being worthwhile.
>
That's rather what I had in mind. The McWire/Rapman/Makerbot hurdle is simply too high for a lot of people. Worse, the McWire doesn't look that high and attracts in people with insufficient skills to see it through. They fail and reprap bets a bad odor.


> Is
> there any good tool that exists to slice stls into
> DXF either commercial or open source? I guess the
> tricky part is that contours (screw holes, etc)
> would have to be flattened into 2D slices or
> etched/marked for later drilling.
>
Not that I know of.

You really need to just redesign the mendel parts for laser cutting in the same way that Darwin parts were redesigned. Vik tells me that he plans on doing that.

---

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

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

The electronics is overpriced. I aim to make a single board controller with parts cost < $50.

My motors were only about $50.

$50 seems a lot for fasteners for a repstrap. I used mainly wood screws, nails and glue.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> The electronics is overpriced. I aim to make a
> single board controller with parts cost < $50.
>
I think that that is perfectly possible, especially for someone with your talents.

---

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

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

Where did you get your steppers for less than $15 a piece American?

---

repraplogphase.blogspot.com

Just to reiterate by $500 I meant perhaps $300 materials and 20 hours labor or $400 materials and 10 hours labor or $100 materials and 40 hours labor.

I am not sure if this is a response to my comment about thinking twice about a price tag over $500, but I'm gonna pretend like it is. I meant $500 in materials cost, I value my time much lower in this endeavor as this is my hobby. If I wasn't doing this I would just be sitting on my couch watching TV, which pays very poorly. Also I was writing my comment when your (goinreverse) comment was posted, so the fact that we both picked $500 is just a happy coincidence. I wonder if that says something about a $500 price tag being a limit?

I don't want to overpromise, but the structural parts of RepolaRap XY platform so far have cost me about US$70:

That is very cool. I think it would be good if the cost of non-reusable materials in a restrap could be kept as low as possible. I also think the cost of the electronics have ballooned quite a bit from previous versions. I am actually kind of glad makerbot sold out before I could place my order as I managed to find substitutes for the motor drivers and motors for less than $180 for 4 of each. Granted they aren't standard but I have seen others using them. Maybe publishing multiple versions of "known to work" electronics would be good idea as opposed to advocating one version of proprietary electronics that very difficult to get a hold of.

P.S. Does anyone else think it is odd that makerbot just rasied the price on a full set of electronis from $175 to $200 despite all reports indicating that they won't be restocking them?

spaztik Wrote:
> P.S. Does anyone else think it is odd that
> makerbot just rasied the price on a full set of
> electronis from $175 to $200 despite all reports
> indicating that they won't be restocking them?

Makerbot has indicated they will carry Gen 3 electronics till they can't sell them any more.

---

repraplogphase.blogspot.com

Makerbot has indicated they will carry Gen 3 electronics till they can't sell them any more.

Fair enough, I was misinformed on that point. $200 plus motors is still getting up there in price though.

spacexula,

[www.slidesandballscrews.com]

Actually with VAT and exchange rate of 1.52 that is $15.

---

[www.hydraraptor.blogspot.com]

BeagleFury Wrote:
-------------------------------------------------------
> tools. The most difficult part is to cut accurate
> circles in the MDF for the two build platforms,
> required because they act as pulleys.

How are you doing your cut for this? I've always found circular cuts relatively trivial, using a router compass.

nophead Wrote:
-------------------------------------------------------
> The electronics is overpriced. I aim to make a
> single board controller with parts cost < $50.

something like this? It's nowhere near $50- the motor drivers alone are about $20 each.

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

I have a McWire that I built and I've changed the motors and stepper drivers to get up to 8mm/s. I imagine that if I switched to ACME threads I could go even faster.

I am happy enough with 8mm/s though because I only built this to make parts for a Mendel. I don't really mind waiting for it to finish printing parts.

I think, when thinking about the successor for the McWire, is it is important to put yourself in the place of the person doing it. They are probably like I was. They want to build a RepRap but have no idea how to get the printed parts. Having access to the DXF files and web based step-by-step instructions was huge. This person may have a limited set of tools and might not know exactly how to build the thing. They will learn though and when they see it running for the first time the feeling of accomplishment will be fantastic

For me the biggest frustration was waiting for Makerbot to have stuff in stock. This was before I realized you can get equivalent, heck better in my opinion, parts from other places.

I think the successor should have a fairly simple design that produces a machine that is able to make the printed parts and then, if wanted, stripped for parts for the Mendel.

It should have multiple suppliers listed for each part needed including the electronics. And it should be possible to make a fair amount of progress by going to the hardware store and buying parts and going from there. I realize you have to get some parts online but, again, you can't underestimate the feeling of going to the hardware store and being able to acquire parts for the thing.

Also being able to build it bit by bit would be good. So maybe you start with just building the X/Y axis first then get that running, then the Z, then the extruder, etc, etc.

-- Chris

Quote

something like this? It's nowhere near $50- the motor drivers alone are about $20 each.

The motor drivers are only $20 because you bought boards from Polulu. The chips are only about $5, so 4 drivers are $20 plus some passives and connectors.

---

[www.hydraraptor.blogspot.com]