An honorary Duggar

Nophead’s Mendel1 has produced 15 sets of Mendel parts, and is hard at work on it’s 16th!!!  My understanding is that it would take about 60 hours to print a full set of Mendel parts.  I have to admire anyone who prints up Mendel/Mini-Mendel parts because of the amount of dedication it would take to do so.  After spending 90 hours2 printing something, I don’t know that I could part with it.

My point is that we really have guys like Nophead, Spacexula, and Cyrozap3 to thank for cranking out parts and pushing replication forward.

I hereby bestow upon you gentlemen the Duggar medal of continuous replication.

  1. I don’t know if he’s named it yet – so I’ll call it Anna until he corrects me. []
  2. I’m just assuming it would take me 50% longer since I don’t know what I’m doing… []
  3. Dude, I realized yesterday I’ve been misspelling your name.  I’ll go back through and fix that – sorry! []

New affordable 3D printer – the Ultimaker



Update 3/28/2011:  The Ultimaker is available for pre-order!

Erik de Bruijn along with a newly formed Dutch RepRap group have put together this new design for a low cost RepRap alternative.  The stated goal of their blog is “designing/developing an easy to build low cost 3D printer with a small form factor but large build envelope.”  Erik has invited everyone to comment on this beta design.  It’s interesting to see their progression of prototypes in their second blog post.

The interesting thing about this model is that it combines some of the best parts of Darwin and MakerBot and what appears to be a Bowden extruder.

The very nature of FDM1 means that the robot will only move the Z axis a little bit every once in a while.  One of the design drawbacks to the Darwin was that the heavy extruder head (motors, gears, heating elements) were so heavy that the entire robot would rock or vibrate with the lateral XY movements.  MakerBot got around the heavy extruder head problem by moving the platform in the XY and moving the extruder head only up and down.  This design decision isn’t without it’s tradeoffs, however.  One downside is that their build area is much smaller than a Darwin.  Another downside of the MakeBot design is that once the object being built reaches a sufficiently large volume or height, moving the object around quickly on the XY platform causes it to vibrate, shake, and become somewhat unstable.

One of the main improvements inherent in the Bowden extruder is that it allows you to take the heaviest parts of the extruder head, separate them from the rest of the extruder head, and move those heavy pieces to a different location.  Using a Bowden extruder makes a Darwin style robot much more feasible – the small moveable print head won’t have the mass to cause the robot to become unstable.

It seems that combining either of the X or Y movements with the Z probably won’t matter all that much, since the Z axis will only move about 0.33mm or so per layer and the Z axis typically doesn’t operate at the same time as either the X or Y axis.  Combining the as the new RepRap version II, Mendel, design shows us that combining the Z axis along with either the X or Y axis, but not both, can lead to a very stable configuration.

What I like about the Ultimaker design is that it would appear to incorporate some of the best parts of the MakerBot and Darwin designs. It appears to have a bolt/nut/T-slot MakerBot style assembly structure using thin lasercut wood pieces for the body.  I found these parts to bolt together very quickly.  Contrast this to the Darwin/Mendel structure using lots of nuts and threaded rod and printed plastic parts to hold it together.  At the same time, by making use of a Bowden extruder and the Darwin body shape, it appears to be able to use most of the interior volume for printing.

I suspect it probably uses fewer parts than a typical Darwin, but I can’t be sure.  I also have to wonder about the cost of lasercut wood versus the cost of nuts and threaded rod.

Nice find RepRap Log Phase!

  1. Fused deposition modeling – basically creating a layer and then fusing a new layer on top of that layer in succession to build up an object. []

Who wants a mini-Mendel?

I like the mini-Mendel, don’t get me wrong, but there isn’t nearly the kind of documentation for it as you would find for the full fledged Mendel.  Sure, it’s a little cheaper, but a Mendel gives you nearly four times the build area.  Since the price barrier to entry into the RepRap project just isn’t that high, the bigger issues are probably going to be quality of documentation, support, skill level required, and interest.  Is there a detailed mini-Mendel construction guide somewhere?

A complete-ish RepRap sold

A few days ago I posted about a RepRap Mendel for sale on eBay with all electronics, motors, belts, etc fully assembled1  The auction (for a set sold in Ireland) settled at 760.00 Euros or about $1,017.18 USD.

Frankly, I would have thought it would have sold for more.  Printed parts are still selling in the $300-$450 range, electronics clocking in at $250 or so, bearings at $50, and all the other bits probably costing another $100 for belts, rods, nuts, bolts.  This means just the materials would cost $700 – $800.  Then there’s the delay and extra expense of sourcing all of the parts for yourself.

Interestingly, I had thought the bottom had fallen out of the RepRap printed parts eBay market.  However, a set of printed Mendel parts just sold on eBay for $455 on April 5.  I suspect the difference is that these parts are being generated and sold in the US which makes for an easier and slightly cheaper transaction

  1. To be fair, the seller mentioned it would require some adjustment to start printing. []

How MakerBot Industries can help RepRap even more

Some people have suggested MakerBot is somehow stealing thunder from the RepRap project. 1  Here’s a super simple way every MakerBot sold could, in a very small way, help the RepRap project.

Why not put a copy of every Mendel part as an STL on the SD card that comes with the MakerBot?  Actually, why not put Spacexula’s set of Mendel production STL’s?

This is a cheap and fast way to disseminate plans for RepRap files to people who are actually capable of making them.

  1. I totally disagree, but there you go. []

Have we reached the bottom already?

Having reached the bottom of a market can actually be a very good thing.  It means anyone who wants a set of RepRap parts can have them for as little as the market will bear.  Right now there are lots of options for someone who wants to get involved in building a RepRap/RepStrap.  You can get a MakerBot, RapMan, ShaperCube, Profound Devices, Isaac Mendel, or pick up a large selection of parts on eBay.

As an owner of a MakerBot, I’m far more likely to want to print up my own RepRap parts than buy them.  However, each of Spacexula‘s Mendel print sets would probably take me 2-3 hours of print time plus about 15 minutes of human intervention/monitoring.1  I would of course also print up parts for any friends/family who wanted parts. 2  I’d be willing to do this to create my own Mendel, but the idea of spending three weeks3 , say $30 in plastic4 , and about six broken up over that three weeks fiddling with stuff5 makes me wonder if there’s a better way to make $300.00.

Setting that aside, I wonder where the bottom of the RepRap parts market is heading?  Spacexula has suggested around $250 or so based upon the price for lasercut parts.  I suspect the price of printed parts will always be higher than the price for lasercut or molded parts – because of the time involved.  Lasercut parts can be cranked out as quickly  as a lasercan cut. 6  Molded parts can be churned out as quickly as the poured material can be dried.  Plus, no matter how nice lasercut/molded parts are – you still have a RepStrap, not a true RepRap.  Assuming identical quality, I’m probably always going to be more interested in printed parts over alternatives.

  1. Such as setting up, warming up, untangling plastic, checking, peeling off, etc. []
  2. At this point, this is purely theoretical/hypothetical since none of my family/friends are interested in their own RepRap.  Hmm…  Maybe I should talk about RepRap MORE??? []
  3. A little over one sheet a day []
  4. I haven’t weighed a sheet of Mendel parts, so this is pure conjecture. []
  5. Assuming no extruder clogs, blocks, PTFE bulging, oozing down the threads, and problem free printing… []
  6. How many phasers could a laser slice if a laser could slice phasers? []

I actually felt sorry for that bastard

Edit: Jeff – Just in case you read this – I do not really think you are a bastard.  I follow your blog and am very appreciative of your in depth posts and analysis on all things MakerBot/RepRap.  The title is more about me feeling like a sucker than a commentary about you.

Jeff posted about totally abandoning his dreams of building a RepRap Mendel.  He said what pushed him over the edge was the constant question, “So, what are you going to build with it?”  Of course, the guy had to post this on 4/1.

I actually felt sad at the thought of him turning his parts into a “melted sculpture of failure.”

I suppose it’s not his fault.  I traded my MakerBot for some magic beans yesterday.

Full RepRap for sale

A complete1 Mendel for sale?  That’s incredible!

There has been a truly amazing progression2 in RepRap parts lately.  While I’m not crazy about eBay as a way of selling3 , it’s a very democratic4 way of disseminating RepRap parts.  The first few parts and sets of parts were all printed, then molded, now MOLDS are for sale?!

Yes, the Platonic ideal of RepRap is that a machine makes the components of the next machine.  But, is it not also part of the RepRap ideal that these machines be disseminated as far and as widely as possible?  It’s really great you can use a RepRap to build another – but that doesn’t mean it’s the best/most economical way.

These RepRap mold are advertised to be good for roughly 50 pourings.  It is advertised to create 9 vertexes at once, but they all appear to be 1/2 vertexes.  The posting also suggests it takes 12 vertexes for a full RepRap.  So:

  • 12 vertexes per RepRap / 0.5 vertexes halves = 24 vertex halves required
  • 24 vertex halves required /9 vertex halves per sheet = 2.67 sheet uses per RepRap
  • 50 uses per sheet / 2.67 uses per sheet = 18.75 sets of RepRap vertexes

Admittedly, this is only a set of molds for the frame vertexes, not the entire set of parts.  However, it’s really only a matter of time before a set is up for grabs.

This makes me wonder – is there a different market value to printed parts rather than molded parts?

  1. ish []
  2. transmission?? []
  3. having sold things on eBay myself []
  4. fair? capitalistic? []

Seems like only yesterday

While searching up a link to the eD’s first sketch for the Mendel I realized that post is about a year and a half old.  That made me wonder how long I’ve been interested in the RepRap projected.  I first read about it in a SlashDot article back in April 2008.  I remember thinking that a DIY level 3d printer was just a fantasy.

Now, I’ve got one sitting in the next room and objected printed with it all throughout my home.  (Several window latches, a sugar packet holder, various little toys, etc).

What’s even funnier is realizing that I’ve been boring my friends and family with RepRap talk for more than a year and a half as I enthusiastically gushed about the possibilities of printing ANYTHING on a homemade printer than can make copies of itself.