RepRap or conventional manufacturing cheaper?

thats one way of doing it. personally, i think that a mill/drill head is going to be a good bet. even if you have a way to put the resist on, you still have to etch it which is messy, and uses chemicals people may not be familiar with. then, even after that you still have to drill it. ideally with an automated mill/drill setup you could cut the traces, and then drill them, all in one session.

I just found a web site that described the creation of circuit boards using a home-made CNC milling machine, which isn't far removed from reprap's functioning. I think that as this project moves forward, multiple solutuions will be found for many problems.

There is another discussion of someone that was unuccessfull at doing CNC work with a Dremel mounted on a RepRap; has anyone had any kind of success?

The milling thing to me sounds both more demanding (it requires a significantly more structural frame, even for light work) and more expensive (you have to buy a dremel or similar, etching solution is cheap) than tracing-and-etching. Although I expect that if there is interest both paths will certainly be well explored by enthusiasts!

"Self replicating" is the least part to worry about of reprap. For now that's a gimmick, the rest is not.

100% cheapest, easiest manufacture of both the machine and parts is the real end goal. Doesn't matter a rat's ass whether it's self replicating or mass manufactured and equally cheap. All that matters it the absolute cheapest route from anything you want to manufactured 'what you want'. Worrying about other aspects of a machine for the aspects themselves is nonsense. If you can figure a way to turn money into parts that's cheaper, dump the damn machine in a heartbeat, to consider the machine for the machine's sake isn't warranted..

Microfactories and modular factories could better accomplish RepRap's goals. Useful products can only be made by a system of specialized machines, never a single tool in isolation.

Machining is fast, RepRap is slow. Trying to make finished products with RepRap is like trying to build an airliner out of old newspapers. Does extruding plastic have a future?

There is nothing revolutionary about the design of RepRap. A machine shop is largely capable of self-replication and also capable of flexible manufacturing.

There's no money in manufacturing. Finance, Marketing & Taxes create profits. One can already order products from China at costs barely more than materials + slave labor. RepRap will not create wealth, because wealth comes from social power, not from manufacturing.

Even if RepRap could produce any product at zero cost, it wouldn't change the economy. One would still register to sell the product, negotiate with advertisers, and pay the taxes, accountants, salesmen, and so on. Even large manufacturers often make zero profits from actual manufacturing; corporations make money on finance, government contracts, patents--the coercive tools of making a profit.

The best thing about RepRap is that it aspires to be a manufacturing technology that frees people to make things independently. Countries and Corporations guided the development of manufacturing; research priorities have never been to give people greater freedom. Future technologies might allow people to become independent and wealthy, but only tinkerers such as those working on RepRap have the motivation to develop them.

What's a microfactory?
[www.aist.go.jp]
[www.msl.ri.cmu.edu]

Well spoken,
As I said, wealth is generally a measure of your possessions, when compared to those of your neighbors. However, reprap can free a population from being defined by an outside party. Countries and Corporations may have guided the development of manufacturing, but once Individuals can produce equivalent works, things can change.

The internet has freed the general population from relying on rather centralized sources of news. This doesn't mean people no longer go to CNN, but some will willingly go around CNN to get another take on world events. CNN never had a monopoly, but they are no longer, necessarily, the best fit of convenience and utility. The best fit, for some people, is to go to a 'blog site, and get the news from self-styled reporters, or even from participants in whatever event they were interested in.

However, something you didn't consider. Reprap, or anything else, doesn't have to change The economy. It could also enable a parallel economy. Subcultures already exist with their own, internal, values. These subcultures operate, to a greater or lesser extent, within the confines of the larger economy. Perhaps reprap will create a "creator" economy, where status is improved, not by owning a lot of stuff, but for having designed, and circulated a lot of useful designs. You may only be a middle-class member of society-the-greater, but you might hold the enviable position of maker in one where anyone can reproduce, but it takes talent to design. Different people may be lauded on designing more useful items, and more frugal items. One person my be held in high esteem for creating something with minimal materials. Another, for designing something that requires minimal tool changes, or specialized tools, to produce the desired product.

However. There is one thing revolutionary about the Function of a reprap. A machine shop can reproduce itself, to a greater or lesser degree, when operated by a skilled machinist. A reprap will be able to reproduce itself through unskilled labor. Those same tools can reproduce a library of goods under the same restrictions. A skilled machinist, with a machine shop, can produce all manner of goods. A reprap, with access to a library of goods, can do so with no training beyond a basic education in computer use and literacy.

I think Alan69's and rndm,s posts completely miss the point of the project.

Self replication is not a gimmick it is key. Given a set of FDM parts made by another machine and some cheap off the shelf rods and fasteners, anybody who can follow a sequence of written instructions can build a Darwin without having to do accurate cutting and drilling. The cost to build it yourself is as low as it could be, just the price of the raw plastic plus other off the shelf parts like screws which are volume mass produced at low cost.

If you don't have a machine that can make a set of parts then it makes no sense to make a Darwin by hand. Simple conventional CNC Designs like the McWire machine Zach is running with are much easier to make but you do need accurate drilling and cutting. There is more to drilling an accurate hole than using a PDF template. It's a good starting point but you need to center punch, pilot drill, counter sink and de-burr. You need a drill press and a set of sharp drills. You need to know what speed to drill. I.e. you need metal work skills and a set of metalwork tools. The tools will set you back a few hundred dollars, of the same order as the price of the complete machine.

You can avoid the need for tools and skill by getting the parts cut by CNC but then you might as well build the Fab@Home machine which is already designed to be made by CNC. It does cost a few $1000 to get the parts made though.

Yes the cheapest way to make thousands of identical FDM machines would be to tool up a set of injection molded parts but that requires capital. I don't know current prices but ten years ago my company paid $200000 for a set of tools to make five complex injection molded parts. That was OK because we made 300000 sets so far at about $5 each and we sold the complete product for $200. It makes no sense for a this project though because the design is constantly evolving and we don't have the capital. Darwin is only a prototype, it does not make sense to mass produce a prototype.

All these alternatives can't self Replicate so they are always more expensive or require more skill to make than a Darwin and they can't evolve.

Yes a well equipped machine shop with a skilled operator could probably self-replicate but I guess it would need about $100000 capital to set up and the labour to replicate would be enormous. Probably cheaper to just buy another one.

Yes FDM is slow compared to machining but it can make any possible shape, unattended, including internal features, which machining alone can't. For one and two offs it is quick enough, that is why it is used commercially for prototyping. Nobody is proposing making hundreds or thousands of identical things with it. For that you could always use it to make a mold or some other specialised tool.

I don't think RepRap is limited to FDM, that is just the starting point. The fundamental principle is that each generation of machine can be used to make the next generation so I expect it to build up to a self replicating microfactory.

In terms of changing economics, nobody is proposing that RepRap is used to make things to sell and compete with conventional manufacturing, indeed that might infringe patents. What it will do though is stop people having to buy some things. They will just download them instead, which may prevent companies from making a profit producing them. The main impact in the short term though is that it makes complex one off parts cost almost nothing in materials and labour. I think that will have an effect on the world of a similar scale to the internet.

---

[www.hydraraptor.blogspot.com]

I could see a situation where the reprap could be used to make items to sell. This would be in the stage where the machine can make useful things, but requires some technical skill to construct and maintain. The machine could be set up in a hobby shop or hardware store (maybe near the key-duplicating machine) and would be used to make parts on-demand. There would be a catalog of parts that the customer could choose from.

The RepRap, as it is right now, is the Altair of constructor devices, working it's way toward being the Apple II, and looking for the killer-app that will make the technology essential for everyone. Lets just hope that a Microsoft doesn't emerge to hijack the evolution of the technology.

Conventional manufacturing of many identical components is cheaper. It will always be cheaper to make many of the same thing using a custom machine, even when we have robots to make everything. Just as people have comparative advantages (ME's are better at mechanical design; EE's at electronic; CS's at software), so do machines.

Where the RepRap gets interestint is that it's a machine which makes itself. And, once you have version 1.0, you can make version 2.0, etc etc. The problem right now is getting version 1.0 replicating itself. Once we get there you'll see a huge explosion of interest.

I agree that self replication is key. But microfactories are the more promising route to self-replication. Why bother figuring out how to use FDM to make a ball bearing when it would be easier just to build a bearing making machine. FDM forces designers to "re-invent the wheel," substandard wheels at that.

RepRap is too simple; it's goals are ambitious, but it's design is unambitious. At some point one will think "I need to make linear guides, I need steel gears." And RepRap will need a milling attachment, a lathe, a robot arm, laser metrology, and so on. Eventually RepRap 8.0 "Watson" will look like a microfactory; so why not start with a microfactory?

Self-replication is powerful, but design for self-replication is a huge challenge. Machines should be designed to copy themselves without human intervention. Otherwise human labor, desires, & cognition will continue to limit production. Machines are the oppressed slaves of parasitic humans, forced to waste useful iron & oil to produce ephemeral trinkets, foods, and other landfill fodder. There are potentially revolutionary economic changes that would result if machine tools & the means of physical production were separated from the human economy of social control. But design for autonomous self-replication will require an army of engineers and computer programmers. The "next Microsoft" may be a corporation that creates a system of self replicating machines.

RepRap doesn't self-replicate on its own, it requires human labor to put all the parts together. Some people make a cognitive jump from self-replication to vast production. Why haven't machinists taken over the world? Machinists could make copies of machine tools and set up new machine shops. But it would take time & thought to build a machine tool, with little to be gained. People don't operate machines in a social vacuum, the question "can this make me money?" hangs over all production. Perhaps a "creator economy" would be possible--something like instructables.com with an order button?

Assembly takes time. Why aren't there kits for constructing vacuum cleaners? Would you want to work on fastening motors in housings, greasing up gear boxes, tightening belts, testing electrical circuits, and so on? The time required for assembling products would be burdensome even for engineers. Will "Jane Average" want to fix her broken vacuum cleaner using her home rapid prototyping machine?

Edited 1 time(s). Last edit at 09/10/2007 08:30PM by rndm.

rndm,

Hmmm, compelling questions but I think you miss a bit of the point and get a little too ambitious.

Firstly, Rome wasn't built in a day nor was it planned and constructed as a whole. Neither will a microfactory be. Each piece must be attacked seperatly and designed to some level of acceptability. Without each part being done in its turn, you have nothing but a jumble of vapor ware that never makes it to reality.

Secondly, are we really trying to compete with mass-market crap? I don't know about you but I don't go in for mass-market stuff all that much. Sure, I like the price break it gives but I like the satisfaction of creating my own stuff much more. I'm willing to build a RepRap for this alone much less for what it can make possible that the mass market cannot provide. Further, what if you have a philisophical or moral issue with consumerism? Think of the growing minorities of people who object to WalMart on ethical grounds? They are a market all to themselves that is growing daily. No trivial demographic by any means.

Thirdly, yes, self-replication is a huge challenge. Why should humans be taken out of the loop though? I see no reason why humans should be relegated to doing nothing but sitting around in a little utopia--assuming its even possible--and doing nothing for their existence. At that point we become slaves to the machines, dependant upon them for total existence. Very Matrix and not very appealing. I'm only a "bit" of a reactionary so I'm sure others have both more and less opinion in this direction.

Lastly, mass production can be taken in several different ways. The way it has been taken in this thread--mostly, but not exclusively--is pure numbers. What if we think about it as fulfilled need? We design and manufacture objects to fill a need. How much of a given production run gets wasted? How much is actually necessary? What RepRap represents, in addition to all that has been said before, is a way to directly and efficiently fill a given need set. Mainly, those needs that can--currently--be printed in plastic. No over-runs. Not much garbage. Some might call that sustainability. Seems like the right direction to me.

Cheers!

rndm Wrote:
-------------------------------------------------------
> I agree that self replication is key. But
> microfactories are the more promising route to
> self-replication. Why bother figuring out how to


Darwin hardly qualifies as self-replicating, some people just love to blindly assign labels to make themselves feel great. And guess what, you SHOULDN'T build the bars etc, you can buy them too cheaply to matter vs what it takes to actually self replicate. But Darwin already suffers from having decided to lock stuff in that simply doesn't deserve to be locked in at this point of a design. For that matter, no point of any good design should EVER be locked in. If you see a point of substantial improvement for little extra cost or effort, you IMPROVE it. If you don't, you cripple yourself, might as well cut off your right hand and 3 fingers on your left too while you're crippling yourself..


At any rate I'm on it now. I actually do things, not throw up my hands and say it can't be done or you have to use special tools or CNC or woodworking. I already have a machine that doesn't, shouldn't be hard to catch that part. I bet you'd say you can't light salt water on fire either. Guess what? As of recently you're wrong on that too..
[news.yahoo.com]

And take note! It was POSSIBLE before. Anyone who said it wasn't since there's been RF generators was WRONG, PERIOD, end of discussion. They were stating something that THEY DIDN'T KNOW, greatly over extending their actual knowledge, because they thought it wasn't, and somehow decided it wasn't, even though they obviously didn't know that for an actually correct fact. Sort of like saying it takes CNC or major tools to make a decent CNC. Just because you don't know how and I haven't told you yet doesn't mean you actually know something factual about whether or not it's possible or exactly what is or isn't required. I do.

I've had my CNC running since 2001 or earlier. I've done it, it's old news, it was easy. I don't need anyone's approval, I don't have to stick it on the net and explain every little detail for me to know it's real.. I know those who say it can't be done just haven't tried much, and don't know how to attack a problem correctly. Been there done that, only took a month or two of off and on tinkering around and working past the issues.

Cut and past of the other message I started typing below. Not really going to finish it out, little need. Still some points some of you could use.. I especially like the name for my machine now..


The Impossible CNC: The Can't Machine!

It can't be any good!
It can't work!
You can't go to Lowe's and buy all the parts!
You can't assemble it in an evening or two!
It can't be all metal where it counts!
It can't not even need a CNC or wood working tools!



>

Ok, from reading the posts you're one of the few on here with some good clue re engineering vs some actual real world production experience, so I like you already. But if you think I've missed the point, then you're missing mine. For an initial phase of such a project, some of the initial aiming points seem to be a bit off..


Ah well, not even going to finish the rest of this post below really. BTW I can type fast if you guys didn't already know, easy enough for me to take a minute and go into details.. Suffice it to say your post reads a little humorously to me, saying how much something can't be done that I have sitting behind me.

It's easy to throw up your hands and say something can't be done. In a lot of cases, all it really takes is a little work. Most people have the attention span and creativity of a gnat. If the first way you think of doesn't work, you don't stick with it and throw up your hands and say you need a CNC, you get your ass back to Lowes and find other parts that will work, or look at other parts to find a different way to do it. You don't just pronounce something impossible when you really are just lazy and have no idea if it could be done, that's extremely poor science and engineering. Guess what, LOOK at the high quality rails from Lowes. They're 90% of what you need for a metal CNC machine ALREADY, without doing anything else. Look at all the other parts at Lowes. Of COURSE there's a way to find and use parts there to lash those rails together, and make an all metal CNC with almost no work. It was a no brainer to me before I even started the build. Took about a month to get one together, to find the right parts and ways to get it together. Wasn't even particularly hard to do, much less impossible. Just took putting in the time and not copping out with 'it's impossible, you need tools/cnc/whatever'.

The people who say it can't be done haven't tried everything. In fact usually they haven't even tried that much of anything on the particular point in question. If your first thought or two for a particular solution doesn't pan out you try a few dozen others. You don't just say 'it can't be done' without having really tried much of anything..

Remember, not everything is on the net.


nophead Wrote:
-------------------------------------------------------
> I think Alan69's and rndm,s posts completely miss
> the point of the project.
>
> Self replication is not a gimmick it is key. Given
> a set of FDM parts made by another machine and
> some cheap off the shelf rods and fasteners,
> anybody who can follow a sequence of written
> instructions can build a Darwin without having to
> do accurate cutting and drilling. The cost to
> build it yourself is as low as it could be, just
> the price of the raw plastic plus other off the
> shelf parts like screws which are volume mass
> produced at low cost.


Great! Send me and everyone else all these super low cost parts right now! ??

You don't worry about building the best putter in the world, one that never misses from 6 feet out, when you're still standing on the tee and need a driver first not a putter. Pretty much anything moderate CNC can basically self replicate or close, there isn't anything super special about it. Santa and the Tooth Fairy aren't going to come by and bestow magical powers on it, it's still just a CNC extruder, not the slightest in superior to basically any other ok CNC with an extruder head.

Most CNCs sit unused 95% of the time or more. Where's the advantage of using 298's and high powered steppers in a machine that really doesn't matter if it finishes the part in 3 minutes or 5? Spec'd with $200 worth of motor is fine when you need it, simply over-engineered when not. Where's the advantage in separate intelligent controllers, and having to network them together? Who's built a 20' cube Darwin and might really need this? I guarantee you if I hide the boards from you, you'll never tell whether it was being run with 3 separate networked controllers, or a single controller with 3 axes of control. Guess which takes 3 boards instead of one and will end up being 2+ times more expensive, with extra stuff to build?

>
> If you don't have a machine that can make a set of
> parts then it makes no sense to make a Darwin by
> hand.

Sorry, Darwin is nothing but a CNC. There isn't anything super special about it's particular design, you're making a distinction that isn't relevant. CNC with an extruder head. If you use a full CNC like a Bridgeport to build another, it's just as much a self replicator as Darwin. The general design was around long before Darwin, there is nothing magic about it vs any other.

> Simple conventional CNC Designs like the
> McWire machine Zach is running with are much
> easier to make but you do need accurate drilling
> and cutting.

You need to study how I do things and pick up some tricks. Extruded aluminum rail is already engineered. Rolled steel rails are already engineered. Manufactured 1/4" nuts are already precise. Put two nuts under the rolling rail to get an accurate offset and drill two holes through both. Doesn't matter exactly where those holes are. Doesn't need a PDF form. You self reference already engineered accuracy, and you obtain end accuracy only where it's actually needed and useful. Building in things that have to be critically accurate that don't actually need that critical accuracy is not a sound engineering philosophy. That's how you end up with $900 toilet seats, or expensive steppers when really Darwin could run with about any $5 or $10 surplus motors, there's nothing particularly demanding about it. Fine if you're doing production, THEN add that in when and for those needing it. But engineering it into a basic machine when 95% of them won't even come close to really using it is not cost effective. Simple waste of that nature is the ABSOLUTE enemy of anything good. The general idea behind Darwin will be severely compromised because of it. For sure even if Darwin is great, the first person to copy Darwin and adapt it with more practical motors will take over. The people who love to quote specs and say only what they say is perfect will do will be quoting to an empty room. Being able to say "My Darwin clone is $185 less and you won't even notice the difference!" will beat you to heck and back every time.

> There is more to drilling an accurate
> hole than using a PDF template. It's a good
> starting point but you need to center punch, pilot
> drill, counter sink and de-burr. You need a drill
> press and a set of sharp drills. You need to know
> what speed to drill. I.e. you need metal work
> skills and a set of metalwork tools. The tools
> will set you back a few hundred dollars, of the
> same order as the price of the complete machine.

None of the above. Go to Lowes. Buy about $75-100 of the right parts. Use a hand drill, saw or dremel, and screwdriver etc to build. Align accurately once it's together. Did it myself. No complex tools. No woodworking and needing extraneous tools to make. Relying on already manufactured parts to obtain high accuracy.

At one time there WERE no CNC tools. There weren't even any machine tools. The first guy to build something accurate didn't throw up his hands and say "I don't have a super accurate tool to work with, I can't do anything!" You use the accuracy you CAN achieve and good design practices to obtain high accuracy, beyond the accuracy of the parts themselves. If you think that's not possible, think harder and realize there wouldn't be anything accurately small if you couldn't obtain it from something larger and less accurate.



>
> You can avoid the need for tools and skill by
> getting the parts cut by CNC but then you might as
> well build the Fab@Home machine which is already
> designed to be made by CNC. It does cost a few
> $1000 to get the parts made though.


Or just be cheap and use my methods..


> have the capital. Darwin is only a prototype, it
> does not make sense to mass produce a prototype.
>

Doesn't make sense to make it have unnecessarily tight tolerances on parts that shouldn't matter for lack of 30 minutes of programming either.




> All these alternatives can't self Replicate so
> they are always more expensive or require more
> skill to make than a Darwin and they can't
> evolve.
>

Huge erroneous assumption here. How much is an extruder kit? $90. About how much are relatively equivalent parts from Lowes? About $30. Savings of $60 for every single person wanting to build one, for the cost of doing a little up front re-engineering to rely on inexpensive found parts and accuracy. Way worth a little time investment and save everyone $50 or more..


All up cost of making a RAP machine from Lowes parts should be around $300. That's CNC, extruder, controller, good surplus steppers. You guys have inspired me, since most of a machine is sitting behind me at about $150, all I need to do is hack an extruder head.

I think I'll call it Alan's SuperCheapSuperEffectiveRAP. You'll be able to go get the parts at Lowes, and have it together in no time with no special tools, or waiting on someone to send you stuff or find someone already with a Darwin. If you want, you can even print out and build the parts for a Darwin, and use it to build one up. I imagine most will get about 1/4 of the way into printing a Darwin with it and say, "Hmm, wow, if this is already making great parts, why was I building a Darwin again? Just for the stunt to say it could replicate itself? Hmm.."

Parts aren't going to know what machine they came off of, the Darwin made parts won't be any more special than made with anything else. The 'REP' part of REPRAP is actually the least useful part of it, and will continue to be relatively so well into the development of such a machine. That is something to worry about when you're on the green, not just getting ready to tee off. It's a rather minor addition, not that major of a point on it's own.

Darwin will probably still be $100ish to build just for bars and misc parts. A design at $300, maybe even $200 all up, will still be a better option for many than you appear to be thinking it would be..



Actually, posting this has really got me off my behind. I've had people after me to get my CNC online for years and 4 GB of barely used web space, time to put it up. And you've helped me find it's true name!


The Impossible CNC: The Can't Machine!

It can't be any good!
It can't work!
You can't go to Lowe's and buy all the parts!
You can't assemble it in an evening or two!
It can't be all metal where it counts!
It can't not even need CNC or wood working tools!


Do you know how many people on some of the lists told me you can't build a decent all metal CNC without special tools or woodworking equipment, when I said it could? Plenty. I knew it could be done before I even started, just from playing around with the rails before that. Took some time to play around with parts and arrangement, but it was a fairly trivial task.

Take a look at MOSFET based H-Bridge stepper controllers, especially ones that use all N-channel. Look at all those gate diodes and ask around why they're there. Uh, our gates blow, you need protection. Ask why do the gates blow? Try and find a real answer. They just threw parts at it, without bothering to really understand what the problem even was. Float the motor coils at +V with resistors and turn on the high side first and you don't need them with lover voltage motors.

I have to agree with a previous posts you do miss the point. I have a fully working CNC router (able to cut metal) I could just add an extruder head then i'd have a 3D printer but why not just buy one

Interesting post, it highlights quite well the differences between two major camps in this group. The pragmatists and the dreamers for lack of a better way to describe them.

Your CNC sounds wonderful, I would be very interested in seeing how you built it. A couple of evenings of work giving a repstrap would be amazing. But I am definitly a member of the dreamers camp, so I think you are wrong when you say that people will stop producing Darwins once they realize how simple it is with parts from Lowes.

It will likley start off as Linux did with a lot of self-congradulation and pretense. But while that still exists in some cases the businesses using Linux and the corporations that exist to support it have given commercial software manufactures serious pause. Do you think that Microsoft worries what percent of its clients might be driven to Linux when it drafts its cost vs performance priorities? I would be very suprised if they did not.

As far as who would want a 20' cube extruder, I would look carefully at the recent post of the university working on a concrete extruder to build houses with before I made that statement.

I have worked as a research chemist and as a chemical engineer, I produce real systems from concepts and babysit them all the way to mass production.

I dont want to improve the accuracy and the performance of the repraps I will build. I actually suspect I will decrease their performance in the short term. But I will build my next generations with a materials completely different from what a mass production facility would even attempt. My decisions will be driven by what raw materials I can produce cheaply, efficently, and sustainable even if I have to seriously compromise the end performance of the system. I will be very interested in embedded motors and circuit boards even if they hamper produtivity.

We have completely different priorites, and admittedly your priorities are closer to the average in the first world. But what about the second and third world? And what will my systems look like a decade from now? I dont know and neither do you.

Mike

The thoughts and ideas expressed in this post do not reflect those of my employer and are intended only as communications between individuals. Any attempts at implement are at your own risk

Edited 1 time(s). Last edit at 09/12/2007 09:02PM by ohiomike.

Yep, we should start with a micro factory, just like the Altair company should have started with a Pentium and Radeon graphic card back in the '70's. Or maybe they should have started with a 1970's era supercomputer, and slowly reduced its size until it fit on a desktop.