RepRap 3.0

I know Mendeal is fairly new and here I am talking about the next iteration...but are their firm plans yet for the successor to mendel?

is there somewhere on the website\wiki that covers future design iterations?

It seems that a heated table would be included as well as an improved extruder...

Kevin

I suspect 3.0 will be highly influenced by that $80,000 prize - especially since the Reprap team had input on deciding what the requirements to win the prize are.

So I would think the main things for 3.0 would be
1) Reliabilty - hit print and walk away, come back a week later with all your parts printed out
2) Printing circuit boards

I think that prize is unwinnnable in the specified timeframe..but would love to be proven wrong...

maitri982 Wrote:
-------------------------------------------------------
> I think that prize is unwinnnable in the specified
> timeframe..but would love to be proven wrong...

Agreed on both points.

RepRap needs to decide what they want to focus on. Each goal of reprap could lead it in a different direction. It depends on what they decide to prioritize.

I think Mendel will be hard to improve on when it comes to the multi head multi functional RepRap machine. I think Mendel will still be made after RepRap 3 comes out. It will likely become the work horse "main" printer for a long time.

If you want easy to put to together, I don't think anything RepRap could come up with would be easier than a Makerbot. I could have put it together in 8 hours if I had known what I was doing. Again I think the Makerbot "style" printer will stick with us.

Darwin/Rapman 3 is a dead end. I don't think people will choose to "go back" to that design.

There are really only 2 printers left, one that would win the prize, and 1 that we need.

The one that would win the prize would be something along the line of the Sarrus Linkage Mark III linkage printer. It would take a LONG time to print, but does increase the printed portion a significant amount.

I HOPE RepRap 3 is something along the lines of a Micro-Mendel, that focuses on low cost, small print area, lots of hand saw-able/laser cutable parts. Basically anything that can't be cut with a hand saw would be RP. If the printer is small enough and hand tool friendly you could likely get RP parts down to 20 hours printing, and you would only need a drill and a hand saw to fab the rest of the components besides electronics and motors.

---

repraplogphase.blogspot.com

Quote

RepRap needs to decide what they want to focus on.

I think they're trying to specifically avoid this. Instead it's a "everybody works on the part that interests them".

I haven't seen mendel with multiple heads yet...but i guess its coming...

I personally would like to see much larger printers. I know this possibly decreases the chance of self replication, but it increases the number of practical parts that could be produced for everyday use...which I think is more important.

Kevin

I would like to see finer resolution, with colors.

What first interested me in Reprap was the idea of printing out custom D&D minitures and spaceships. But alas, this is harder than just making a smaller hole in the extruder

sorry for my naive question Not having had my hands on one the intricacies of printing resolution are lost on me. it seems to be lost on many as it has not been solved yet.

kitep- ironically, the extruder nozzle size cancels out of the math- this can be simply explained by the extrudate stretching when the head moves faster than the nozzle exit velocity, and blobbing when the head moves slower.

The hard part about printing models is the search for a suitable support material- something that the plastic adheres to strongly enough during printing to avoid distortion but is also easy to remove afterwards.

---

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

I have recently found that stretching to achieve better z resolution also makes it harder to print fine details in x and y since the filament gets pulled inside curves.

---

---

Darwin clone, Gen 2 electronics, Arduino Duemilanove w/ AtMega328, 5D Firmware, Pinchwheel extruder
[www.codeerrors.com]

I personally don't agree with most of the goals of the prize so that isn't really here nor there for me.

For 3.0 my personal goals are:

1) Much lower barrier to entry, the design should have a goal of 10,000 implementers worldwide within 24 months. I think necessitates the use of a single design which can be built with either standardized or printed parts.
2) Host software that is a single installable package on windows/linux/osx and that is not embarrassingly bad in terms of UI and usability. Enhanced and integrated support for 3d object warehouses (thingiverse, etc). Quality default profiles should be included and it needs to 'just work' out of the box.
3) A design which supports 2 scales, a smaller version (not so small it cannot print itself) without fanciness like a heated bed and a larger version which has a heated bed/area and 1 cubic foot / 30.5 cubic centimeter build volume.

There is a lot of experimenting going on with circuit printing which I think is an outstanding goal but it still seems 2+ years away, at least until component placing is supported. I think that should be a goal for V4.

I would like to see experimentation start and continue in V3 with laser sintering. It seems to me to be the easiest technology which can address speed and quality issues. This is basically a process where a not-so-powerful laser (equivalent to your dvd-writer possibly) draws a 2D layer in a bed of powdered plastic. Once a layer is complete a thin layer of powder is deposited over the just printed layer and then "wash, rinse, repeat". Layers an features can be extremely small limited only by the granule size of the powder and the wavelength/dispersion of the laser The advantages are that you can draw a layer pretty much as fast as you can possibly move the laser and that virtually any shapes are supported because the object is resting in a self-supporting layer of the powder. Even creating objects that contain other objects, such as a small marble inside a sealed sphere are possible.

Laser sintering sounds like it would be a major advancement on todays extruder FDM approach...I hope they take up your reasoned suggestions, they all make sense to me as an outsider looking in.

Mendel could be improved by significantly redesigning it to cut the print time...

I'm scribbling ideas, I think theres quite a bit of room for optimisation.

And I think laser sintering would push the cost too high for my liking, plus isn't really practical for metals (as you need protective gas to prevent fires!)

... lasersintering with good absorbing plastic could be done without protective gas with a pigtailed diodelaser - i've bought a comercial one with 8Watt optical power (@975nm Wavelength through a glassfiber with 100micron core) for around 340 Euros, so weaker types should be cheaper.

Salvaged diodelasers from DVD-burners have mostly 220mW, so would only work with dark plastics and not really fast ...

For sintering metal-powder you need Argon as protective gas and optical powers above 20 Watts, depending of the laser-type, powder-size and material.

Note: - maybe you can get some plastic-coated metallic particles which can be presintered with a weak laser and then heated in a kiln to solid.


A 50Watt-fiber-laser with a spot of 20 microns diameter can melt/evaporate most powdered absorbing materials with powers from 5 Watts on - i tested with black rock-spheres (0.1mm in diameter) or silicon-coated iron-spheres (0.01 or 0.05mm) from 2 Watts to 50 Watts, but i have to play with parameters to get proper bonding to the solid surface - so this is work in progress ...

Viktor

Is anyone posting info on a blog with picks of doing laser sintering? I for one would like to see it...

I wasn't proposing laser sintering as the only option, I think the main point of mendel is that it opened the door to multi-head use. If there is a faster/cheaper/better way to ultra accuracy and speed than SLS I am all for it. I still see plenty of use in FDM. I don't think FDM gets much better than the stratasys machines and if you are making compound curves surfaces there is a lot of finish work that has to be done.

The shapeways.com service uses a large industrial SLS (Selective laser sintering) machine. They probably have a video somewhere.

My plan was to start investigating with a laser diode packages as are commercially available. Cost is an issue but I was promoting experimentation in V3 not a final standard. The commercially available SLS systems I have looked at are all north of $50K US.

I am not interested in metals from a printing standpoint. A machine like reprap makes it so easy to make casting forms/molds etc that it doesn't seem worth all the headaches and safety issues to deal with printing metals at this point (outside of circuitry which I think is a totally different case then printing all metal objects).

My calculations put 350mw as the target for reasonable behavior parameters working with existing commercially available plastic sintering polymer powders but that is just theory at this point.

If anyone is interested in the sintering concept in general google "candyfab" it is a sintering project that works in table sugar.

Did some research on the laser sintering (ls). The example i saw was rather cool because it laid down metal powder so the final product was a fully usable metal mold that was then used in injection molding...very nice. But as other poster stated, you need to keep that whole process in a nitrogen atmoshere to prevent explosion.

It does seem the laser does not need to be too strong since you keep the powder heated to a temp near its melting point...so the laser just has to put it over the edge. It seems the big challenges here will be in the fine x/y control of the laser and laying down the powder in a uniform thickness across the printing area over and over again.

I guess it may be possible with a simple "squeegy" type mechanism to swipe the powder across the printing area from a powder source that raises as the printing bed lowers...like a piston pushing a cylinder upwards with the powder initially filling the cylindar and slowly being pushed up and then swept across for each layer of printing.

you probably knew all this but i write it for my own edification...

Kevin

Kevin, I've seen that technique used with sawdust and PVA glue extrudate for basically making 3d objects in an MDF-like result. Also solves support material issues, but hollow structures cannot be made closed or you can't get the support out

---

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

Your thought with the sawdust raises a really interesting idea with a micro-misting extruder head delivering catalyst and a power based of epoxy (or anything else) that would harden. I think that is definitely worth exploring.

Laser wise you can also use UV responsive liquid resin, you just need to move the laser really fast. The hardened material has a slightly higher density so sinks in the fluid as it is formed. I can't remember which machine uses that, works wonderfully but the resin is pricey if I remember correctly. Possibly the resin can be gotten through the dental industry in smaller quantities as I think it is quite similar/the same to what is used in tooth sealant.

goinreverse,

i think it would be fun to work on an open source laser sintering project. I checked out the candy maker (http://www.candyfab.org/article.php/printingplastic) and they did a good job, but they avoided some of what is hard. e.g. - automatic application of thin layer of material.

if We wanted to do laser sintering, it could not IMO use a reprap type device as the basis for that. I think a recommissioned printer structure for X direction attached to a movable structure with stepper in the Y direction and then replacing the print head with a laser would be the way to go.

The trick is the printing base. It needs to lower appropriate amount, sweep in material from some source in a very even application and repeat this process. The printing area needs to probably be pre-heated before each print run to get the material up to near melting point and to keep the laser requirements more modest.

this has the chance to solve many issues i see with reprap...there is no real scale issue as you can print big, resolution will be better and the material offers inherent support so no need to worry about 45 degree angles, multiple material printing for supports, etc.

I think if i was to work on something this would be what it would be. If anyone wants to consider starting a collaborative project on this let me know and we can see what we can do.

Best,
Kevin

Kevin,

Sounds good to me. I need to think on it for a few days but I think the best way I can contribute to open hardware 3d printing is with design and implementation of a new device separate from reprap. I have already put in over 300 hours of effort to the point where my mcwire is printing objects. Those will now take what looks to be almost another 200 hours work/print time to complete, It has been terrifically educational but I have some practical things I am actually trying to do with these devices for my work.

I think that 200 hours printing and babysitting time may be much better spent designing and executing a device using my aluminum t-channel concept and connecting more with the makerbot community who sees the self-replication goals, at least today, as more of a hindrance than a benefit which is how I see it. I think in a long term sense self-replication is a good goal to research and pursue. Right now today the project goes through so many contortions for it that I see it self-defeating the most effective adoption, design and implementation.

What do you think is a fair budget for a device which is NOT a bootstrap but a final device. Basically a larger format cupcake, a DIY device, perhaps a little larger than mendel. The cupcake price point is really $1000.00 all said and done. My existing design with aluminum t-channels come in at $650 excluding electronics I think. However if you have 2 sets of planes (really just X1,X3,Y1,Y2), you can move twice as fast without hitting the inertial limits (g force), though cost would be a little more.

I think that it needs to support FDM and I would really like to get that working well. But I would also like to use it to research laser sintering. Note that commercial base materials are pricey starting at $1.50 per cc and going up. Not sure if we can find something as a cheaper substitute (like finely ground abs, abs dust is a definite safety hazard though). I guess that sounds more expensive than it really is... I think it is crucial that a device support multiple print methods.

I think new software is needed. The pleasant 3d guy seems to be experimenting. Skeinforge is a good proof of concept tool but I can't seem to wrap my head around why it has to be so complex. It seems a very brute force approach. Something which has some built in calibrating routines and takes as input from those results the acceptable range of extruded material output (in cc per s or m), a few movement characteristics of the planes and then considers a set of factors of a design object (STL) should be able to produce consistently good gcode for a particular machine. Skeinforge requires that you manually tweak a mind-numbing array of tiny factors that are, at least to me, very arcane.

I look forward to your thoughts and anyone elses for that matter.

answers below...

What do you think is a fair budget for a device which is NOT a bootstrap but a final device. Basically a larger format cupcake, a DIY device, perhaps a little larger than mendel....

[Kevin] I think it is hard to tell right now. this device will be inherently more expensive than MakerBot is today. the laser and the probably need to pre-heat the material will add to the cost. Frankly given the cost of the commercial versions and the problems this method solves over FDM justifies a higher price. If it could be done for $1,500 that would be great...perhaps with clever design it can be less.

I think that it needs to support FDM and I would really like to get that working well.
[Kevin] For me, if we could make an effective LS device, that would be good enough. Why do you think FDM is needed? It likely changes the design significantly to have to carry a extruder and its electronics instead of a laser.

But I would also like to use it to research laser sintering. Note that commercial base materials are pricey starting at $1.50 per cc and going up...

[Kevin] I started researching HDPE powders and it looks like it is $.75-$.95 per pound...assuming the source i found was correct. One of the main things to determine is the best price\performance materials. We could even decide that a hybrid would be the best solution since the mixing of properties may provide the best price performance. At the end of the day a higher priced material may be needed as a trade off to getting better quality end product with no real restrictions are the end products shape as there is today with FDM.

I think new software is needed...

[Kevin] based on what you and others have said this sounds like a requirement. Hopefully some re-use of existing code can get this half way there.


have you posted any pictures of your design?

Has anyone looked at using makerbeams for the structure?

[makerbeam.com]

For the laser, i wonder if a repurposed DVD laser might be powerful enough?

[www.instructables.com]

If it is hot enough to etch the surface of a DVD perhaps it is good enough for HDPE sintering...

As it happens, I have a DVD burner that i am not using. I may dig out the laser tonight.

and more info on making a burning laser:

[www.youtube.com]

***safety warning***
Be EXTRAORDINARILY careful with lasers, you can't necessarily see the output that will make you blind and it can happen in a fraction of a second. I would strongly urge you to get proper high end eye protection from someone like wicked lasers and make sure you validate that it is covering the correct wavelength of light. I see kids on youtube playing with those and the potential for permanent site loss or worse is huge.
***safety warning***

Ok, enough preaching of safety.

Makerbeams appear to the same as be aluminum slotted t-channel though scaled down. Industry standard extruded channel is 1"x1" or 1"x2" , Makerbeam seems tiny (maybe 1/4 inch width) but I think that is a good thing in how to mount the smaller pieces so thanks for pointing that out.

I don't see supporting multiple methods of deposition/construction as a real obstacle at least in a larger format device and I think long term it is a requirement. The powder/liquid bed just sits on the Y axis and the head gantry is not dissimilar to the FDM extruder. Also check out the bower extruder which shifts the heavy stuff off the gantry. In the not that distant future the issue of printing circuit boards (1 years maybe 2 years) and placing components will be worked out, that is significant and I want to be ready for it.

I think FDM is important because there are certain structures that it is better suited for. I also think we will need it to build the LS device and I don't want to have a need for any other bootstrap devices. There are definitely going to need to be some unusual parts I think and FDM can help keep costs down. Also getting FDM working allows the project to immediately do something useful and validate software and mechanics of axis movement, crucially I think getting to that point also allows recruitment of other participants.

In terms of a laser I get the sense that you are thinking we just mount one of these "balloon poppers" in place of the extruder and off the to races. I am not against trying that but I think there is a lot of complex R&D needed to really get a better result than FDM. The wavelength needs to be consistent with the material to be heated. With current DVD writer diodes there is some overlap wavelength/materials but maybe not that ideal. Power/performance is also a hugely complex issue. Realize that the optics define the beam radius, smaller radius more power/precision but also a lot more moving to fill. That means potentially a way of altering beam radius which itself is complex if not impossible in budget or possibly having two laser heads, a fill and an outline. Duty cycle is an issue, cooling, venting gases, eye safety and on an on.

Off the shelf the lasers that do what we really want with a good duty cycle and adjustable beam optics are $5-7K. My hope is we can do something useful with a laser that is <$700 or even <$500.

You might take a look at these:

[blog.makezine.com]
[planetstephanie.net]
[www.wickedlasers.com]
[www.newwishlaser.com] (Chinese OEM)
[www.thinklasers.com]

The wicked lasers have potentially the strength and duty cycle but I think their beam radius is 1.5mm so we would need optics to focus that. The easiest way to get something going maybe UV resin and a UV laser (or other laser that leaks into the UV spectrum). I think I figured out which liquid resin is generally used, it is Bisphenol A Epoxy Acrylate or Diacrylate (yes the stuff they just figured gives us cancer and probably shouldn't be in baby/water bottles).

Ok, brain dump complete for now...

goinreverse Wrote:
-------------------------------------------------------
> ***safety warning***
> Be EXTRAORDINARILY careful with lasers, you can't

[Kevin] yep...today was the first day in my life where i read much about lasers. the first lasers i read up on were CO2 lasers...that type of laser is scary strong. I was going to say that at my age (44) I am wise enough now to not do stupid things, but i sliced a bog ole gash in my hand getting the laser diode out of my DVD burner tonight, so maybe I am not so wise after all
>
> I don't see supporting multiple methods of
> deposition/construction as a real obstacle at
> least in a larger format device and I think long
> term it is a requirement. The powder/liquid bed
> just sits on the Y axis and the head gantry is not
> dissimilar to the FDM extruder. Also check out the
> bower extruder which shifts the heavy stuff off
> the gantry. In the not that distant future the
> issue of printing circuit boards (1 years maybe 2
> years) and placing components will be worked out,
> that is significant and I want to be ready for
> it.

[Kevin] Well it can't be like Mendel. The X\Y axis control needs to go on the head and not the printing bed...ala the Darwin. Doesn't really seem logistically possible to handle all the material handling on the bed for LS and also be moving it in two axis. so the bed is the z-axis and the tool is x/y. There's also likely to be extraneous stuff for handling the material spreading for LS that is completely not needed for FDM or part picking, but hopefully it won't get in the way of those processes if designed right.



>
> I think FDM is important because there are certain
> structures that it is better suited for. I also
> think we will need it to build the LS device and I
> don't want to have a need for any other bootstrap
> devices. There are definitely going to need to be
> some unusual parts I think and FDM can help keep
> costs down. Also getting FDM working allows the
> project to immediately do something useful and
> validate software and mechanics of axis movement,
> crucially I think getting to that point also
> allows recruitment of other participants.
>

[Kevin] harder to do than pure LS, but i hear ya. there's a lot of FDM experience here for sure.

> In terms of a laser I get the sense that you are
> thinking we just mount one of these "balloon
> poppers" in place of the extruder and off the to
> races. I am not against trying that but I think
> there is a lot of complex R&D needed to really get
> a better result than FDM. The wavelength needs to
> be consistent with the material to be heated. With
> current DVD writer diodes there is some overlap
> wavelength/materials but maybe not that ideal.
> Power/performance is also a hugely complex issue.
> Realize that the optics define the beam radius,
> smaller radius more power/precision but also a lot
> more moving to fill. That means potentially a way
> of altering beam radius which itself is complex if
> not impossible in budget or possibly having two
> laser heads, a fill and an outline. Duty cycle is
> an issue, cooling, venting gases, eye safety and
> on an on.

[Kevin] i do not know what laser is needed. I considered wavelength and see that there are red/blue/green and i think yellow available for sure. My plan was to find just find out by some testing. so starting with a DVD diode based laser amp'ed up to 100mW pointing at some various powders and just seeing what happens. Clearly you have been researching this longer than me...but i learn quick.

I wonder if there are adjustable optics for lasers? Electronically adjustable of course. Since cameras have this feature now and cheaply, i would imagine it exists. better to avoid having two lasers if possible.

>
> Off the shelf the lasers that do what we really
> want with a good duty cycle and adjustable beam
> optics are $5-7K. My hope is we can do something
> useful with a laser that is <$700 or even <$500.
>

[Kevin] if we could do it cheaper that would be better. i was hoping to avoid pre-heating the material which means less laser powder needed, but perhaps pre-heating will be required and perhaps could also serve double duty for FDM as doen by nophead.

> You might take a look at these:
>
> [blog.makezine.com]
> ke_a_dvd_burner.html
> [planetstephanie.net]
> pie-2/
> [www.wickedlasers.com]
> [www.newwishlaser.com] (Chinese OEM)
> [www.thinklasers.com]
>
> The wicked lasers have potentially the strength
> and duty cycle but I think their beam radius is
> 1.5mm so we would need optics to focus that. The
> easiest way to get something going maybe UV resin
> and a UV laser (or other laser that leaks into the
> UV spectrum). I think I figured out which liquid
> resin is generally used, it is Bisphenol A Epoxy
> Acrylate or Diacrylate (yes the stuff they just
> figured gives us cancer and probably shouldn't be
> in baby/water bottles).
>

[Kevin] we don't want to use stuff that is harmful if at all possible. do you want to be in the room while LS of BPA is occurring????

> Ok, brain dump complete for now...

[Kevin] is we want to get serious about this than a dedicated wiki and perhaps blog are in order...then its time to start talking rig design...this non-trivial for laser sintering for sure as laying the material for each layer at a uniform thickness will be very challenging...along with the other issues you mentioned.

here is a laser i saw burn wood:

[www.budgetgadgets.com]

its only $45. probably needs better optics...but that price is good. not sure the temp it achieves versus what we need, but easy enough to test.

On another note...i have been using Google Wave at my company for collaborative efforts. it lends itself to this type of back and forth. This is not to replace a wiki or blog, but for maintaining a conversation better than forums or email. If you want i can setup a wave and invite you and someone else i know for thsi project..than anyone else who epxresses interest can be added too.

I should get wave setup but I am jumping between computers and OS's all day long so standard web works better for me at the moment. Etherpad might be a good option.

In terms of safety I should point out that pretty everything about FDM and laser sintering is bad, those ABS fumes from the extruder, bad, PLA, slightly better but still bad. Check out the MSDS sheets for these plastics. Pretty much all melted plastic offgasses and prolonged exposure is not a good thing. I think the laser takes it another step because I am guessing is has the tendency to vaporize some amount of material.

I don't think UV liquid resin is really any more dangerous than the other options. It does have the issue that leaks/spills can make for a bad day... more from a cleanup standpoint than safety.

I think that camera optics are not high enough quality glass for lasers but I could be mistaken. Laser glass is ridiculously expensive so probably finding some standardized to repurpose would be a good thing.

Some laser physics are coming back to me, I think what we want is a called an electrically actuated laser collimator. Thinking about it now it may also allow us to get away with a weaker/cheaper laser that has a larger beam radius like the wicked lasers low-mid range.
Is there any semi-dangerous idea that has not been documented on instructables? [www.instructables.com]

I made an etherpad, anyone feel free to contribute, note the license.

[etherpad.com]

Looks live is quite a bit better than etherpad...if you want to send your email all I need to do is invite you. there is no real setup other than making sure you use chrome or have the chrome frame installed in IE\Firefox....which it will do for you automatically.

My email is kcramer at sapiosciences dot-com

The problem for me is installing chrome/chrome frame on tons of different computers and OS's, just not something I have time/energy to deal with right now. I have used wave, seems like a lot of people say the opposite, that etherpad has some things they are waiting to be added to wave... YMMV.