Why RepRap is failing

Am I really going to defend this? If you really don't see the success that RepRap already is, then that's your problem. Enough people out there do see it, and others will follow when certain milestones are reached.

The question is not why, but when. It will happen. You're mistaking RepRap's potential for failure.

---

Regards,

Erik de Bruijn
[Ultimaker.com] - [blog.erikdebruijn.nl]

Quote

If that's true, and hundreds of Repraps have been sold, why can't people buy the parts for $20+shipping instead of $360??? Especially if people are following the recommeneded each Reprap builds 2 for friends...

I assume it could happen, it's just not organized.

If people buy repraps, get them working and never tell anyone or offer to help (or sell their services), that isn't the fault of the project. This isn't communism, we can't *make* people share stuff.

Personally, I'd be very startled if there were hundreds of repraps in the wild making high quality parts. I don't think the project is at that level yet.

... actually most active reprappers are battling to optimize the fabbing process or their mechanics.

Then there is the time-cost-relation: - when you need some days continuous fabbing to make all the parts, then this would mean someone has to spend some weeks to months of his (mostly spare) time to make a second kit and every breakdown or failure would elongate this task ...

My guess is that this 'breeding'/reproducing phase would start, whem the reprap is optimized so you need lesser counts of parts and the process is much faster and more reliable.

As with my CNC as repstrap: - it's no problem to mill all the parts for an extruder, but i'm troubled with many other problems and actually can't spend the needed time for finishing my paste-experiments or even copying parts for friends - maybe next year ...

Viktor

VDX Wrote:
-------------------------------------------------------
>
> My guess is that this 'breeding'/reproducing phase
> would start, whem the reprap is optimized so you
> need lesser counts of parts and the process is
> much faster and more reliable.
>

Yes! Well said. We certainly aren't there yet.

Obviously you cant force anyone to do something but you can persuade them. Why not set up a parts farming system. Set up some way for people to make a specific part and then buy x amount off of them. They could sign up, decide what part they are going to make and then they get double the plastic it required to make the parts or give them credits on the reprap store for more plastic. Someone making the same part all the time be able to put out a higher quality then someone trying to make every part and I imagine they would be able to find ways to speed up the process.

You basically need a path to a Reprap that begins with a Repstrap that you can build from scratch without having to depend on other people to give you parts. That's just about what Tommelise is achieving now that I've got it to mill plastic stock.

... i can remember a discussion in last year about repstrapping repraps and parallelizing the fabbing capability for speedup.

The point was to build as many repraps, as you have place in your basement and then let the 'overcounted' run with specific tasks continuously.

So you can make all parts you wants for you and for replacing breakdowns in your 'reprap-farm' and in background 'grows' a stock of giveaway-repraps.

This would mean that every wanna-be 'distributor' has to start with a farm of 5 to 10 repraps, a big order of filament and an even bigger electricity-bill too ...

Viktor

VDX Wrote:
-------------------------------------------------------
>
> This would mean that every wanna-be 'distributor'
> has to start with a farm of 5 to 10 repraps, a big
> order of filament and an even bigger
> electricity-bill too ...
>
Tommelise draws about 25-30 watts. With properly designed stepper motors (which ARE available), I doubt that Darwin would draw much more than 50.

Edited 2 time(s). Last edit at 09/05/2008 06:28PM by Forrest Higgs.

Hi Forrest,

my CNC-mill draws about 1,2 kW

I think an optimized faster and milling-capable reprap would draw between 100 and 500 Watts, so a 'farm' of 10 or more repraps in continuous mode and with all supporting equipment would be in the medium kilowatt-ramge too.

AFAIK some machining workshops around with 5 to 10 machines spend some thousand Euros per month for electricity and such, so this would be a heavy chunk to carry ...

Viktor

You don't need a milling machine to make a Darwin you only need a Darwin, and that only takes in the region of 50W so the cost of electricity is negligible compared to the plastic, which is only about $20.

---

[www.hydraraptor.blogspot.com]

VDX Wrote:
-------------------------------------------------------
> ... i can remember a discussion in last year about
> repstrapping repraps and parallelizing the fabbing
> capability for speedup.
>
> The point was to build as many repraps, as you
> have place in your basement and then let the
> 'overcounted' run with specific tasks
> continuously.
>
> So you can make all parts you wants for you and
> for replacing breakdowns in your 'reprap-farm' and
> in background 'grows' a stock of
> giveaway-repraps.
>
> This would mean that every wanna-be 'distributor'
> has to start with a farm of 5 to 10 repraps, a big
> order of filament and an even bigger
> electricity-bill too ...
>
> Viktor

Wouldn't it make more sense to have less RepRaps (maybe just 1) and have them produce patterns which can be used to create molds which can each cast many new parts (or RepRap the mold itself instead of the pattern)? I would think that for many parts, casting would be preferable to repeated prototyping.

I'm guessing it might be quicker and cheaper (with parallel casting - multiple parts at a time), but I don't have experience with any casting techniques, so if someone with such experience could chime in with specifics, that'd be nice.

It might help to alter certain minor parts of the design a bit to reduce the amount of cores and stuff (is it possible to RepRap the cores?).

Casting is labour intensive, which makes the parts expensive. The whole point of RepRap is the machine does all the work, making the parts cheap. It is slow however, so you need lots of machines. One person could run lots of machines, but the original scheme of make two sets of parts and pass them on, putting the onus on the receiver to do the same makes far more sense. There would be lots of machines making parts but owned by lots of people.

The problem at the moment is that everybody has bought their Darwin parts so there is no obligation to make parts for others, although I know several people at least are attempting to do just that.

Proof of concept is provide by Vik Olliver who RepRapped a set of parts for the price of the plastic and I have RepStrapped a full set of parts.

It is only a matter of time before I make some more parts for people. If there was enough demand I could build a farm and churn parts out but I would have to sell them at a cost to pay for the farm and my time to build it, which would again make them expensive.

And as Forrest points out there is nothing stopping people obtaining the parts by building a cheap RepStrap.

Edited 1 time(s). Last edit at 09/06/2008 03:20AM by nophead.

---

[www.hydraraptor.blogspot.com]

Brace yourselves!

nophead Wrote:
-------------------------------------------------------
> Casting is labour intensive, which makes the parts
> expensive. The whole point of RepRap is the
> machine does all the work, making the parts cheap.
> It is slow however, so you need lots of machines.
> One person could run lots of machines, but the
> original scheme of make two sets of parts and pass
> them on, putting the onus on the receiver to do
> the same makes far more sense. There would be lots
> of machines making parts but owned by lots of
> people.
Why can't we automate the casting process?

I mean we do have a Rapid Prototyping device, why can't we build a prototype automatic casting machine that accepts variably shaped molds and pours the casting material (someone help me out here and suggest one) into a standardized opening (or perhaps non-standard opening which is specified by the controls)? To make the molds from RepRapped patterns, I suppose you could RepRap a box, place the pattern in the box, and then cast the mold out of the mold material. Perhaps a standardized reusable mold system where the two halves are designed such that they can be quickly and easily locked together, placed on the molding system, poured into, and removed for cooling (can you remove it before it sets without messing it up? I would think it wouldn't matter with a locking mold, and this would allow more molds to be poured with the same machine without waiting for the previous to set). Maybe it could be equipped with simplified (one axis/motor per) arm systems which automate the removal and positioning processes (though this is probably a less important part which costs a lot more in development time; if the machine could automatically grab the RepRapped pattern or mold from the Darwin, this part could allow for an entirely automatic process; however, I do not think this is likely to happen in the near future, as currently a completed build requires human removal and evaluation).

Personally, I would make a prototype of such a system myself and try my ideas, but I'm still working getting my RepStrap up and running. As soon as I do, though, I will try it.

---

> The problem at the moment is that everybody has
> bought their Darwin parts so there is no
> obligation to make parts for others, although I
> know several people at least are attempting to do
> just that.
I think some sort of centralized incentives system would help. I don't believe in asking someone to work without obvious compensation, and most people (conciously or subconciously) won't work without some form of compensation (actually, I think anyone who works without compensation would be deemed by the majority of the medical community as insane - altruism doesn't exist in humans).

Perhaps some sort of buy-back program: either the RRRF or the original RP part supplier could buy back the original parts sold or preferably new parts created by the new machine (the problem with the new parts, I would imagine, would be gauging quality and determining a fair price; though, if they function properly, perhaps the price could simply be determined by the accuracy of their function); the purchasing entity would then re-sell the parts to new community members.

Though, I don't quite see the point in such decentralized production at such low scales (that is, I don't think this kind of system is the best option for the growth of the RepRap community yet).

---

> Proof of concept is provide by Vik Olliver who
> RepRapped a set of parts for the price of the
> plastic and I have RepStrapped a full set of
> parts.
>
> It is only a matter of time before I make some
> more parts for people. If there was enough demand
> I could build a farm and churn parts out but I
> would have to sell them at a cost to pay for the
> farm and my time to build it, which would again
> make them expensive.
Interesting point. Perhaps a system which takes the load off of individuals and spreads it out over time and over the community?

I'm imagining a system where someone pays a central entity (perhaps RRRF or something similar), indicating their desire for the RP parts of a Darwin machine, and the entity spreads the part production out over the community, requesting, say, a few parts per month per member. Each producing member, of course, would be paid per part, but by reducing the amount of time each member's RepRap has to be dedicated to replication, each member can focus their RepRap's time on other things, either personal or development (testing new stuff etc, which is actually a less obvious part of replication - mutation). The entity could actually stockpile a few parts in advance to mask the leadtime (so the new member could immediately get their RepRap shipped to them). Perhaps, after acquiring enough parts for an entire RepRap, and provided they don't have any unfilled orders, this entity could put the extra parts together (with the non-RP parts they happen to have) and start producing parts themselves. When they get an order, I suppose they could deconstruct the RepRap, perhaps giving a discount for a used RepRap (or perhaps not, seeing as it could have been properly adjusted and such, easing the load on the new member). Of course, this last part about constructing their own machines is not a necessary part of the idea; it's just a possibility.

Anyway, I think the term is "exponential growth", and I think this was the original intention. [en.wikipedia.org]

I do think the idea of each owner committing a single week of continuous production for the purposes of replication is a bad one (I don't think anyone actually suggested this, just used it as an example to show how much time it would take to self-replicate). Though, if you spread this week over a few months, it seems much more reasonable (especially if the machine is expected to continue to replicate throughout its lifetime). If you wanted each RepRap to dedicate enough of its operation for complete replication in, say, 12 weeks (about 3 months), then you could, for example, split this time up into 1 day of replication work in 12, or a little more than one day per two weeks.

The point of all this, of course, is that if we organize it properly, replication doesn't have to be such a burden on members (it takes a minimal amount of their time, and they are compensated for it; in turn, they also get a cheap rapid prototyper to do it with - what's not to like?).

I think the main burden is on the central entity which has the responsibility of taking purchase orders, schedueling part production between members, grading parts, compensating members in either store credit or cash, receiving and storing parts, and finally consolidating and shipping parts. I think the first steps this entity would have to take would be
[0) Finding and marking down the names of potential candidates for the described members in 2).]
1) Setting aside funds to compensate the producers.
2) Forming some light, good-natured, win-win contracts with members of the community who are willing to dedicate a certain amount of time during each production period to production of parts of a certain grade (i.e. within the Darwin standards, which are pretty light at the moment) in return for store credit or cash (or perhaps some other form of compensation I haven't thought of; I think filament falls under store credit). They would of course be able to organize the time within each period however they like (either doing it all at the beginning of the period or spreading it out or whatever). The contracts are simply to state the terms of payment and such, and I think the members should be allowed to opt out at any time (though I think that once the central entity orders a part from the producer, it would be in the best interests of the community if they were not allowed to reverse out, unless production had not yet begun or the producing member agrees). Anyway, the point is to keep everyone happy and to avoid any scandals that might occur. The producing member should be allowed to choose approximately how much he/she will be asked to produce (so if you want to produce more, you can, and if you can't, you don't have to).
3) Schedueling production of parts enough for one Darwin, and constructing it as a good proof of concept (and a testing of the workings of the whole production system, like the grading process, etc.). I think producers should probably be paid after parts are received and graded. Perhaps a good way to test parts would be inserting into this original proof of concept and testing function (at minimal cost, perhaps just positioning etc, without actual extrusion etc); of course, there are some parts which can be evaluated without actually testing (like by measurement, etc).
4) Schedueling production of parts enough for a few Darwins, and at the same time beginning to market (well... maybe just announcing that they are ready to sell) and sell them (lead time would of course be stated).
5) Once a few sets are bought, and feedback is received from the new members (who may then enter into such light contracts; maybe you could give them a price discount at the beginning by having them sign such contracts - they could back out, as stated above, but it would still encourage them emotionally), you would then ask for more members to enter into such contracts and thus ramp up production.

That's fairly straightforward. So the real question now is: who are the candidates described in 0) and 2)? I would be one, but I'm not up and running yet.

Edit: Also, who would be this central entity? Is RRRF up to the task?

---

> And as Forrest points out there is nothing
> stopping people obtaining the parts by building a
> cheap RepStrap.
True enough, but it would save a lot of people a lot of time if we had the RP parts of the official, standardized, developed (at least more than most RepStraps) Darwin design readily available at low prices.

I think some (perhaps many) developers with a bit less time are waiting for such a situation, and, while they may not be able to contribute as much to the project (as they have less time), they would still play a role in the development. This is mostly speculation, of course, but it is logical.

[Now that I look back on my answer to your question, the answer is that time is what is stopping people from building cheap RepStraps. Motivation is also another factor, but it is partially dependent upon time.]

Edited 1 time(s). Last edit at 09/06/2008 05:59AM by Joshua Merchant.

Quote

Now that I look back on my answer to your question, the answer is that time is what is stopping people from building cheap RepStraps. Motivation is also another factor, but it is partially dependent upon time.

Even things like the McWire, intended to be a cheap and simpl way to build a cartesian bot are not entirely cheap or simple. If a kit were to appear that was half the cost of, say, the bitsfrombytes darwin kit, people might be keener to get building. It might be a challenge to make such a thing though, and I don't image many people are keen enough to sort it out

As it stands, it is a *hassle*, which is quite enough to discourage people who mostly want to extrude stuff, not tinker with a cartesian bot.

Alright, I'm a little confused. I see suggestions to use a repstrap to build a reprap. Once you have a repstrap, what do you need a reprap for? IOW, what can a reprap do that a repstrap can't? From what I've read, they seem to be pretty much the same thing.

Quote

what can a reprap do that a repstrap can't? From what I've read, they seem to be pretty much the same thing.

The reprap design can, and will, evolve. This makes it possible to print out a next gen reprap when teh designs are available. Sure, you can improve your repstrap design to your heart's content, but when someone else has already made the designs for a better extruder or a toolhead swapper or whatever, and all you need to do is to print the bits off and buy a few bucks worth of metal parts, surely you'd want to take that opportunity?

In the meantime; there is no particular benefit it making a darwin over some other cartesian platform with an extruder attached. But perhaps your friends and associates might be interested in the reprap project, and you're in an ideal position to help them get set up by printing off a set of parts for them.

Etc.

Kitep Wrote:
-------------------------------------------------------
> Alright, I'm a little confused. I see suggestions
> to use a repstrap to build a reprap. Once you
> have a repstrap, what do you need a reprap for?
> IOW, what can a reprap do that a repstrap can't?
> From what I've read, they seem to be pretty much
> the same thing.

Two thoughts.

The reprap is designed to build replacement/spare parts for itself when it breaks. A repstrap may use parts that it cannot build itself and are expensive to buy (ie hydraraptor's xy table).

You could build a repstrap that is high quality and long lasting or you could build a repstrap just good enough to get you to a darwin. In the later case you would move to the higher quality/upgradeable darwin as soon as possible.

The reprap is designed to build replacement/spare parts for itself when it breaks. A repstrap may use parts that it cannot build itself and are expensive to buy (ie hydraraptor's xy table).

But that's just it. A repstrap can also build replacement/spare parts for itself. And a reprap also uses parts it cannot build itself. I must be missing something, because I still don't see the difference.

You could build a repstrap that is high quality and long lasting or you could build a repstrap just good enough to get you to a darwin. In the later case you would move to the higher quality/upgradeable darwin as soon as possible.

Now this may be the answer. A reprap is just a higher quality repstrap?

The reprap design can, and will, evolve. This makes it possible to print out a next gen reprap when teh designs are available. Sure, you can improve your repstrap design to your heart's content, but when someone else has already made the designs for a better extruder or a toolhead swapper or whatever, and all you need to do is to print the bits off and buy a few bucks worth of metal parts, surely you'd want to take that opportunity?

But you can do the same with the repstrap. If someone makes a better extruder, print one out on your repstrap, and then use it on your repstrap.

The one difference I do see, is the repstrap that's a milling machine, not an extruder.

Kitep Wrote:
-------------------------------------------------------
>> The reprap design can, and will, evolve. This
>> makes it possible to print out a next gen reprap
>> when teh designs are available. Sure, you can
>> improve your repstrap design to your heart's
>> content, but when someone else has already made
>> the designs for a better extruder or a toolhead
>> swapper or whatever, and all you need to do is to
>> print the bits off and buy a few bucks worth of
>> metal parts, surely you'd want to take that
>> opportunity?
>
> But you can do the same with the repstrap. If
> someone makes a better extruder, print one out on
> your repstrap, and then use it on your repstrap.
>
> The one difference I do see, is the repstrap
> that's a milling machine, not an extruder.


Sure, you could hack up a RepStrap to make use of the inevitable improvements to the RepRap design. But at some point, the amount of effort it takes to adapt others' changes (which assume you have standard hardware) to your design will probably surpass the effort it would have taken to print/assemble a standard darwin and apply the stock improvements. At that point, unless your RepStrap has other advantages over a Darwin, you would have been better off to just make the transition in the first place.

The idea is to be able to leverage innovation from a wide network of users - that is what a standard design is able to facilitate. Of course, there is no requirement to build a Darwin... and I'm sure that many people will be perfectly happy with whatever RepStrap they can get running. That is just fine.

Quote

But that's just it. A repstrap can also build replacement/spare parts for itself. And a reprap also uses parts it cannot build itself. I must be missing something, because I still don't see the difference.

Can it? Many repstrap designs will not be capable of that. They are constructed to be cheap and easy, not to be universally capable.

What you seem to be saying is 'if I make a really good repstrap that can definitely be used to manufacture all of its component parts, then it is just as good a darwin, and I can mill stuff too'.

The answer to that is, 'yes, of course'. That is the whole point of the robloks project ( [www.lumenlab.com] ) for example... it will be stronger and faster and more capable than a darwin, and can make big metal parts. It even markets itself as a reprap

Important difference? Metal mills are *expensive*. Big, heavy, noisy, dirty, extremely power hungry, to name but a few other flaws. The beauty of the darwin design is that is is quiet simple to set up and tune, not least because it does not need to have a hugely stiff frame and rigid, backlash resistant drivetrains. Furthermore the components are all dead cheap too... look at the cost of a big roll of ABS filament or a bunch of steel rods, compared to a stack of aluminium profiles or a few lumps of metal billet. Look at the cost of the X or Y axis assembly with toothed belts compared to a ballscrew and some slide rails. Oh, and you'll need bigger motors to push all that stuff around, and higher current motor drivers and a bigger power supply.

And there are ongoing costs in the forms of new milling bits and coolant and maybe dust extractors and and and...

I'm sure you get my point.

Thanks for the replies explaining the difference between Repstrap & Reprap. I actually found an old thread that talked about the same thing.

I completely agree with reprapper, reprap is seriously failing. Amongst what reprapper said,I have been asking the same question for help for a long time, each time re-phrasing it. No-one has been able to succesfully help me. I still cannot get the reprap software to connect to the Arduino board. This is pathetic.

jajo Wrote:
-------------------------------------------------------
> I completely agree with reprapper, reprap is
> seriously failing. Amongst what reprapper said,I
> have been asking the same question for help for a
> long time, each time re-phrasing it. No-one has
> been able to succesfully help me. I still cannot
> get the reprap software to connect to the Arduino
> board. This is pathetic.
>
Naah, not really. Reprap is just having a problem common to virtually every open source out there. It's being run by dyed-in-the-wool geeks, and I don't mean that in a bad way as I'm a confirmed geek as well, just not a Linux/Java geek. As a result it reflects a geek skill-set and geek fixations which are very difficult for less geeky people to take on-board. Right now, Reprap is a geek-centric project which is very difficult for people who aren't willing to become pretty good at least Linux and probably Java, too.

I got off that train year before last. It meant that I had to write all my own software in Basic and do all my own electronics design. While that's been a big job, I at least don't have to worry too much about my software and electronics doing some strange thing that I haven't got a clue how to remedy.

Stick with it, there are those here now who will walk you through getting your system running, if you are patient and polite. Alternatively, you could write your own software. That's what I did, but then I'm terminally deranged.

Don't kid yourself, though, Reprap works brilliantly. If the main blog doesn't convince you, take a look at Nophead's personal blog.

[www.hydraraptor.blogspot.com]

forest-

Does you program use the arduino? And who do you think might be able to help me with the reprap program?

I am following this thread from the beginning and in my opinion it is at a point something like

this should not just be said because some things do not work for you.

It's about 3 months since your last post now. It also sounds like the answer could be found by

reading the instructions or searching the forums.

If not, push your thread and tell the people that you have a problem to get your machine work.

To buid a reprap seems to need a lot of time at the moment and a lot of people are spending this

time for building or making things easier.

My first reprap contact is about 3.5 months ago and I still can't get the grin out of my face.

This is the most fun with something technical I ever had, and there was some. So in my opinion the

chance for reprap to fail is already gone.

jajo Wrote:
-------------------------------------------------------
> forest-
>
> Does you program use the arduino?
>
Nope. I stayed with Pic microprocessors when the rest of the team went off to Arduino and AtMel. I had a cheap, stable BASIC compiler by then and saw no need to migrate to C and Arduino. For me it would have been a waste of time.
>
> And who do you
> think might be able to help me with the reprap
> program?
>

Oh wow! Just about anybody on the core team except me should be able to help you. Zach seems to be at the bleeding edge of the Arduino initiative. I'd think that Zach or somebody he'd recommend could mentor you. A lot of people around here are using Arduino and the Linux-based Java PC-side software.

Jajo

Start a new thread in the software section.

Try to make the subject relevant (ie NOT "my system does not work")

List all the information on your set up (Windows/Linux)
State which Host software version.
State which firmware version.

State clearly your problem.

Post any error messages you see.

---

Regards

Ian