Imran's personal blog

October 24, 2011

Other peoples’ designs for better 3d printers.

Filed under: Uncategorized — ipeerbhai @ 10:35 pm

Hi All,

Work on my “Imran’s Better Bot” is going well. I’ve got the first design complete and a few test assemblies made to see if the various ideas I have will work. Earlier test assemblies found major issues — major enough to require some heavy re-designs. The current test assemblies have found some issues, but most are minor things so far. Along the way, I’ve learned a lot about the basic concepts to use. I’ve also found a lot of great bots out there that beat the RepRap and Makerbot hands down. I thought I’d list the important ones here, why they’re good, and what I think is wrong with them as well.

SumPod.

This is a riff on the MakerBot design, except done with MDF( Medium Density FiberBoard ). MDF is strong and lightweight, but thick. Stronger than the MakerBot or the RepRap, it can handle light-weight CNC work. The downside of this design is the closed-box makes accessing things hard, and space to mount tools inside it are limited. Still a very nice design, and one I really like.

An aluminum cased version is now also available.  You can buy these off IndiGogo.

Ultimaker.

This is another closed-cube design, using a cable extruder and NetFabb CAM software. The cable extruder allows extremely fast motion, and the NetFabb CAM software can do .04 mm layer height. This gives them the speed and quality title, at least until Junior’s 3DhomeMade project hits. It’s more expensive, and the small print head means few, if any, tools can be mounted to it.

ShapeOko.

This is the one to watch. It does most of the things my design does, and arguably, does them better. The core problem I see in the RepRap/Makerbot FDM world is that the devices are very weak CNC bots with poor accessibility and difficult builds. The ShapeOko is a fast build, large usable area, and high strength platform that overcomes most FDM wekanesses. It has two weaknesses — one weakness is that its Z height is only 4 inches — while the RepRap can do more. The other weakness is that it uses makerslide, which just isn’t available yet. I think in time, this will be one of the better platforms ( perhaps the best ). I would stop my project and switch to it if it were possible, but I want a bot before January of 2012, and I don’t think the ShapeOko will be ready and available until 2013. Not unless MakeSlide starts shipping soon — and so far, it looks like it won’t ship until summer or later next year.

Update — as of Today, Dec 19, 2011 — Makerslide and Shapeoko are taking orders for people building a ShapeOko! I’ve placed an order for the parts from makerslide and Shapeoko, and there’s a chance it’ll get here this year! I’m excited — the ShapeOko and Makerslide have had my attention for a while. It’ll be great if it’s as good as I expect. I’ll let you all know how awesome, or not, it turns out to be. Hopefully, I can drop work on my system and focus instead on toolheads and software, where I really want to be focusing my time…

Update — I bought and built a ShapeOko.  It’s nice little mill, but I no longer think it will make a good 3d printer.  Everything is on the gantry.  All motors, motion plates, etc…  This makes for a very heavy gantry.  This heavy gantry then moves, giving it a lot of inertial mass.  It may be able to print, but that mass means the prints will have lash/vibration problems.  You’ll need to modify the GCode to compensate, and so far, I haven’t met anyone who;s figured it out.  If you do use the ShapeOko as a printer, and get it working, please let me know.  I’ve been searching the web, looking at blogs, even emailed the creator.  It’s almost May, 2012 — so far, I’ve seen no evidence that it works as a printer, and since I have one, I don’t think it will.  I have designed a Wade’s mount for the ShapeOko.  I’m still testing the mount, but I’ll release that to Thingiverse.  If the ShapeOko can print, it will be at 24 volts, with an 18 ohm resistor added to the hot-end.  I don’t have a 24V power supply, and I’m focused on my own bot design right now, so it’ll be a bit before I come back and finish the work needed to try/test this out as a printer.

Zen ToolWorks.

This thing is awesome. I’ve had a chance to play with one at a hacker-space, and I admit, I lusted after it. You can build it in 4 hours and have a working CNC mill and printer in a single day. It’s also only 360 bucks, without electronics. Expect to spend another 150 in electronics. The firmware doesn’t exist, so you will have to roll your own. No open-source firmware currently supports both CNC mill and 3d printing at the same time. So, you have to modify the Hydra-MMM firmware with pieces from Sprinter/Marlin to get a working bot. It’s also only 7″ square bed area, which is smaller than RepRap. It’s much slower than either RepRap or MakerBot, since it has to handle CNC milling — meaning you need to get faster lead-screws to make it work as a printer. For these reasons, it’s in the “Very Close” bucket. Given how much I’ve already spent on my RepRap, I can’t bring myself to part with the $500 and the time to get one of these fully working. However, anyone who hasn’t yet built a RepRap and has the required skills should go this route, at least until the ShapeOko or my better bot is commonly available.  All Zen mills are moving table designs — these lend themselves best to conversion to 3D printers.

Update — Zen now has an f8 screw that they designed, and the $700 machine was at maker faire in Silicon valley, printing on dual heated beds!  Truly awesome to watch it work.  They’re slower than most printers, but they do function.  They also have a lathe attachment for their new designs.  The $700 machine has no electronics, beds, etc… It would really cost about $1500-$1800 to get a Zen working as a printer.  I think most of the printer/mill combo machines will end up in this price class, since there’s so many parts needed to do both.

MendelMax

MendelMax is RepRap’s version of an extruded aluminum bot. It can be built in only a few hours, and it’s all the rage right now in the RepRap community. If you really want a single-purpose machine, and insist on a RepRap, then the MendelMax is the choice for you. However, I’m not a fan of any RepRap based designs. I’ve posted before on why — the design doesn’t lend itself to multiple uses. I think a CNC bot is the right answer — just add an extruder head to a cnc device.

EasyMaker

EasyMaker is my own design for a 3d printer/mill.  I know the most about this one, as I’ve designed it.  I had a RepRap, and didn’t like it very much.  It sort of sucked.  So, I got a mill.  But I never got it to work as well as I wanted.  So, I decided to design my own robot, that works as well as I like.  It’s fast, powerful, and expensive.  Its fully open source, unlike the Zen, and the design is released on Thingiverse.  It uses a mixture of extruded aluminum, steel rods, and makerslide.  It is both a moving table anda  moving gantry — the only robot I know of that can do both.  The makerslide will probably be replaced with OpenRail, and there may likely be some openbeam added as well.  EasyMaker can be reconfigured for printing or for milling, can be driven by belt or screw, can run with 3 or 4 motors, etc… It’s a very flexible a design.  The downside of EasyMaker is that it is very flexible a design.  You have to figure out the right configuration for what you want to do, then configure it appropriately.  Still, this concept makes it a faster printer than a Zen or a ShapeOko, and it makes it a better mill as well.

The game changer.

The two principles people don’t seem to understand in 3d printing is that the current 3d printers are based off CNC mills. They use CNC’s language — G-Code — to control their motion/extrusion. Any CNC mill can be made into a 3d printer, provided that you are willing to print slowly. While mills can be made into printers, the inverse is not true. Printers often don’t have linear strength, and would make terrible mills. Also, the 3d printers tend to use moving tables in very tightly enclosed spaces. This makes build times long, and accessibility poor. It also drives up cost, as many small parts ( screws, nuts, plates, etc.. ) are needed with lots of machining ( holes for screws, rods, etc… )

If you could get a strong, lightweight material and use a high precision moving gantry CNC mill, you would have a great printer. But that’s not possible with existing designs.

The big game changer here is Extruded Aluminum. This material solves the major issues that the RepRap and Makerbot have in common — getting a strong platform that’s good at linear motion for a low cost. Everyone is using long steel rods right now, and though steel is strong, ( it has a Young’s Modulus constant of 200 Giga-Pascals! MDF has a rating of 3, and diamond has 1,200 ), it’s very heavy( 7 times the density of water! ). So, people use thin rods ( say 8mm or so ) to compensate for the weight of steel. But thin rods bend easily in shipping, so you end up with bent rods that weigh a ton, and your Z stage wobbles. Extruded Aluminum solves this.

Aluminum is strong ( Young’s Modulus is 69 GPa — 20X stronger than wood! ), but low density( 2.7 g/cm3 — much less than steel ). You can make a very large-area beam for the same weight as a low-area steel rod. Thus, you get a much stronger part, since Area is a square function. Extruded Aluminum also has embedded T-Slots, meaning you need no machining, as you can directly attach nuts right into the T-Slot. You can make your design nothing more than plates that mount the beams, giving you few parts. Makerslide takes this idea really far, and if it were available, I’d use it. Since it’s not, I’ve based my design around existing extrusions. Prototype deisgns tell me it will work. Calculations I’ve done using the laws of physics tell me that the material is more than strong enough to provide a 6″ CNC height with .01mm accuracy. But three major problems remain. I’ll talk about those some other time.

Advertisements

4 Comments »

  1. Hey, read through your blog a fair bit and looked into the other options you suggested. I’m a college student interested in building my own 3D printer as a side project, and I’ve been eyeing the Prusa for a couple months now as a means to that end. First, disclaimer: I’m not experienced in this area. I’ve been researching in my spare time, but I’m still quite a novice.

    Anyway, you recommended Zen ToolWorks especially, but I’m not sure that’s what I’m looking for right now – based on their website and other sources, it seems that it’s not so much making your own 3D printer as simply purchasing one.
    Also, I’m not intending on building a printer to use as a serious tool, but rather for my own enjoyment. With all that in mind, what is your advice? Should I forsake the Prusa, and go with something like Zen? I’m all ears.

    Thanks for your time.

    Comment by Matt — November 4, 2011 @ 9:22 pm | Reply

    • The 3D printing world is not like that. Nothing you can build or buy right now is a “It just works out of the box” setup where you learn very little about how it works. In any device, you’ll be building and fixing things for a while, and all of these devices are very complicated. Many people I know who’ve gone down the 3d printer route have ended up hating their bot — even if that first bot is a Makerbot. Almost always for the same reasons — they’re too much work for too little reward. The only way to fix this is the either reduce the work, or increase the reward ( or both ). Hence, I’m firmly in the “buy Kits, not parts” camp.

      Once you have a 3d printer, you’ll see how limited they are, and kick yourself for a bot that can’t also do milling. A good CNC bot lets you put a machine shop in your dorm room/apartment — especially if you understand how it works. Too many people don’t — and either don’t know that their RepRap/Makerbot printer is a CNC machine, or think that CNC only works with a router. They can’t see that CNC works with any tool you as human can work with — 3d Print head, Jig Saw, Drill, rotating saw, oscillating tool, laser, etc… And once you understand how CNC works, you can build printers that can print other materials. Want a powder printer? There’s no plans for a powder printer, and no-one will make them, since Z-Corp will sue them just for publishing plans! But if you understand CNC and can think of the first principles underneath it — you know how to make a powder printer.

      So, I’m strongly in the “build a CNC bot, really learn the principles underneath it, then use them to FDM print from your bot” camp.

      The ZEN is not for a novice. As I mentioned — it’s slow and no firmware exists for using it as a printer. Those are complex problems to solve — you’d have to change the screw design, you’d have to run at 30V instead of 12V, and you’d have to create firmware that works with those changes. And you’d be doing it on your own — most people who buy a Zen are machinists, and think of it as a router. Hence, they don’t bother to think of it as a printer, or a laser cutter, or anything else — and it shows in the electronics kits/firmwares in the wild right now. And embeddd programming is nothing like desktop programming — you have to learn a new way of thinking. Going this route would be much harder than a Prusa.

      If you can somehow get your hands on one, the ShapeOko is my recommendation for a novice. The problem with ShapeOko is that the makerslide system just isn’t available to the general public yet. If you can wait until makerslide is commonly available, then you could make a ShapeOko and you’d be pretty happy. You’d also have a “base” bot — something you could use to build many different tools with — not just a mill or an FDM printer.

      If you can’t wait, then maybe the SumPod or MendelMax is a good bet. I think the square-box design of SumPod makes it a poor general tool — you couldn’t fit a jigsaw inside one, so you’de be forced to work with small tools that fit inside the box — like a print head, drag knife, or cable-fed router — but it’s available now, and more importantly, the designer thinks of it as a “base” robot, making it a good starting point for understanding the first principles involved. MendelMax is strong, and so, it could also handle some level of CNC routing. The design is too close to Prusa, and room for large tools doesn’t exist on it. The designer is also stuck in “this is a printer” thinking. But Extruded aluminum is a game changer. because once you have some extruded aluminum in your hands, you suddenly can think of more ways to construct things.

      Hope this helps!

      Comment by ipeerbhai — November 5, 2011 @ 12:41 am | Reply

  2. Hi Irman, Thank you for the rundown. There was a lot of good ideas in there!
    I think the one thing you may not have mentioned is that the reason the prusa / ultimaker are so much faster at printing than traditional cnc machine is because they don’t use leadscrews. That kind of precision is not needed for fdm, and the speed gain is incredible for switching to belts. Unless you start driving really powerful motors or servos and super expensive 4 start threaded rods you are never going to achieve equal speed for both machines.

    One thought I just had was: how about you start with a traditional cnc machine. Use the z axis from that machine and design a x-y platform that uses belts for printing. The threaded rods on the z axis wont slow anything down. Swap the x-y table in when you want to print, and use your traditional x-y axis for cnc engraving, etc. Admittedly, you are building almost two machines at that point.

    The other comment is: of course a laser would be faster for you when you are designing a flat design like the one you illustrated above. You need to pick the right tool for the job. 3D printers excel at printing organic shapes. For instance; building a rabbit from a laser cutter would either end up with a very blocky looking object or would consist of hundreds of very thin layers. The latter wouldn’t be very fast. 3 axis cnc machines have very similar limitations. It’s only a matter of time before you start adding degrees of rotation (and cost) to get the part you really want. A 3 axis cnc machine can’t make a rabbit, and it certainly can’t make parts to build a 3d printer.

    I’d love to have all three machines! And a lathe, and a cnc router, etc…

    Comment by tinyenormous — December 17, 2011 @ 3:39 pm | Reply

    • Starting with a CNC design and having the ability to switch between belts and leadscrews is exactly what I’m doing in my design. Though you can move fast enough with leadscrews to print — Hydra-MMM proves this, getting 200 IPM from their design. Hydra-MMM is the bot I’m drawing my inspiration from. The problem with Hydra-MMM is that no designs exist — just firmware. What I have been able to figure out about it from their pictures and views is that I need a strong CNC machine to create the parts, and I don’t already have access to such a CNC machine — and having a machine shop do the work is very expensive.

      The idea is “base-bot” — something that can do a lot of different things, based on the tool-head attached. The tool head can be a plastic extruder, or an end-mill, or a jig-saw, or an oscillating tool, or an ink-jet head, a paint brush, laser, etc…

      The problem that I have in design is a cheap 2-dimensional linear motion system for an X-Z carriage. The simplest design for this is to use Hiwin carriages and rails, bolted on to the 80/20 beam, connected together by a single plastic plate. This design would also cost $100-$150, just for the 2 carriages and rails — at an overstock place that may or may not have inventory! So, I’ve been designing a cheap 2-dimension linear motion system using extruded 1/8″ aluminum bars bolted to 80/20 beam. The same length using the prototype would cost about $25.00 — much cheaper. But this has its own problems, and the defining problem is lack of an existing CNC machine.

      I’m thinking of trying sc12uu carriages and 12mm rails. I think those can be found straight and true in long lengths, since they’re so thick. I’d have to design a scraper for the rails to keep dust from clogging the bearings in the carriage. The problem is that my chinese suppliers aren’t reliable, and I have to hunt a few different suppliers before I find one that can deliver on quality, cost, and time as I want.

      Comment by ipeerbhai — December 18, 2011 @ 1:50 am | Reply


RSS feed for comments on this post. TrackBack URI

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Create a free website or blog at WordPress.com.

%d bloggers like this: