20100715

MakerBot : Caveat Emptor

I'd like to discuss my MakerBot. Before I start, I must declare that many people have had success with MakerBot, so this post should be taken only as the real-life facts from the experience of a quite possibly less than competent MakerBot operator. With that out of the way :

New owners of MakerBots may be surprised by the amount of additional time and money required to get full performance out of their MakerBot. While this kit is certainly easier than attempting a full blown rep-rap, it has surprise drawbacks which have left many customers disappointed.

The first thing odd you might notice is that the kind, linear assembly instructions trail off just around the time you think you might want to turn on your MakerBot. Don't panic, it is around this time that you should desperately start bothering the MakerBot google group, since if you proceed blindly you might destroy your bot. MakerBots as they currently ship have a number of design defects, which will require your time and money to correct before you can fully enjoy the printer.

1 : The Teflon™ (PTFE) component of the part that extrudes plastic is prone to melting "losing its structural properties and deforming to an extent that it no longer can serve its intended function". If you've calibrated your temperature sensor incorrectly, or done something odd in assembling the hot end of the extruder, you will quickly destroy your MakerBot by melting deforming this Teflon component. What you weren't told is that this part can be protected by being wrapped in a bit of copper pipe and or hose clamp, but the kit doesn't ship with these. If you destroy this part, you will want to upgrade to an equivalent compoent made of PEEK from MakerGear that is significantly more reliable. Indeed, many MakerBot owners wonder why MakerGear extruder parts are not shipping standard with the kit, since most owners quickly upgrade. I spent about $60 in upgrade parts before I even really got started. I don't actually need all that, but I got some spare pieces in case something else breaks.

2 : You may experience inexplicable behaviors of the printer. Some of these are problems with the engineering on the electronics. Essentially, the electronics are not as noise resistant as they need to be. By default, you will need to twist all cables and install R18 ( 180 ohm ) on the motherboard extruder controller. This resistor does not come with the kit. Many people have also had to install ferrite beads to filter out noise from the lines. These do not come with the kit.

3 : You may find that printing objects with a footprint larger than an inch is very very hard. Why is this ? when the MakerBot advertises a 4"x4"x6" working volume ? As the hot plastic cools, it deforms, and as it deforms, it pulls away from the printing platform. Only very small footprint objects can be printed with the default MakerBot. In order to use the full volume of MakerBot, you will need to purchase the heated build platform (HBP) for $50.

4 : The heated build platform kit is hard to build and incomplete. It requires surface mount soldering. I personally had no trouble with this, but many people who don't breath rosin fumes as part of their metabolism do. Furthermore, the current electronics on MakerBot can not handle the 3-4 amp current the heater draws. You will need to purchase and wire in some sort of relay, solid state or otherwise. Relays don't come from the kit, but a board can be purchased from MakerBot for $15.

5 : Calibrating the software is really, really, hard, and some parts just can't be printed for reasons related more to software than hardware. There is no standard MakerBot calibration file for the 3D printer driver software, but people are working on it. It takes a while to understand this monster and I'm still not done calibrating it for my bot.

6 : The extruder filament feed is under-engineered and hard to make reliable in its present incarnation. MakerBot ships with the MK4 extruder design, but you can buy more functional part from MakerBot for $10, which may or may not solve your troubles, I haven't tried it.

7 : The threaded rods that ship with MakerBot are of sufficiently poor quality that you might find yourself spending days trying to get the Z axis mechanism to work properly. I finally figured out that if I rotated the rods so that they were all bent in the same way, the tension on the driving belt would be closer to constant and my MakerBot would actually work. However, there were many, many days of cursing and manual adjustment, before I arrived at this solution. You can buy new rods at some cost that will be of more correct tolerance. You can also print out parts of a kit ( but not the whole kit ) for correcting "wobble" in your Z axis.

so, to recap :

So, thats totaling at least $100 in upgrades just to get the bot in decent shape ( not including tax and shipping ). Thats not including unexpected repairs, and all the time spent cursing and wondering what you did wrong.

I guess, now that I look at it, $100 isn't that much, but it is evidence that this product is still in the alpha stages of development and cuts several corners, making it failure prone or downright broken. I am confident that with more time spent engineering, and perhaps with a minor $100-$200 increase in the cost of the kit, to increase the quality of the parts, this could actually be workable.

As it stands, many operators have spent hundreds of hours and been plagued with incomprehensible problems, and have never gotten their bots to print a single thing. Other people have had enormous success. I think part of it is down to luck ( how bent are your particular Z axis rods ? ) and pre-existing skill.

So remember, if you buy one of these, it is not a kit for amateur, or even perhaps intermediate hobbyists. If you are a novice at the time you purchase the kit, you will rapidly be required to upgrade your mechanical, electrical, and computer science skill set. It will take some time, and some people still don't have their bots printing. Budget extra time and money. The design cuts corners that can make it hard to work with, and if you buy now you are paying for design refinement that will finally bring the product out of alpha stage. If you do succeed, enjoy printing your calibration cubes. Maybe in a few months you'll work your way up to this.

What I am disappointed about is MakerBot is shipping these kits as if they worked almost right out of the box. They really should be advertising their kits as "parts only, functional MakerBot may require additional parts and design changes". I think that their advertising has been very effective, with many interviews with diverse news outlets. But, I also think that their advertising borders on false or deceptive, and might take some less experienced individuals by surprise.

update : opinions on the MakerBot forum echo these. There is agreement that the kit takes a lot of time and some upgrading may be required, but that this is generally expected for a DIY kit ( I didn't know that ). The argument is that this is still a fair buy for people with lots of time and patience, and that if they start shipping a huge number of kits they might have to put up a notice to adjust expectations closer to reality.

update : After I thoroughly determined that the Z rod defect was beyond my ability to correct, I asked for two new rods from MakerBot. They are shipping the replacement parts at no cost. So, if something really doesn't seem right, and suggestions from the forum aren't helping, and your rational mind determines that the fault is in the part, MakerBot is willing to replace the defect.

update : MakerBot will soon be upgrading their extruder design to Mk5, which should eliminate some of these problems.

So, issues 1,6,7 should be resolved shortly. Issue 2 of electronic noise has been solved on my bot by simply twisting all communications and power wires, and installing the 180 ohm terminating resistor. Issues 3,4 are inherent to the use of ABS : it will warp if not kept warm during the build. I am told that PLA does not suffer from this. Furthermore, issues 3,4 aren't really issues if you're good at soldering and buy all the HBP parts with your kit. Issue 5 isn't going to go away anytime soon since it requires quite a bit of software development. As the kit becomes more standardized and the variance between bots is reduced, a standard calibration may become available.

So I can now update my recommendation for the MakerBot : Research ahead of time, budget a couple hundred extra dollars, budget extra time, be patient, expect to have to improvise hacks, communicate with the forums and if all else fails the MakerBot people themselves. Learn to enjoy the process of building and debugging. If you're technically competent enough, in the end you will probably get a working 3D printer for a very reasonable price, thats easier to assemble and debug than a RepRap.


9 comments:

  1. Do you have any idea how the MakerBot compares to the RepRap in price/performance?

    ReplyDelete
  2. Yes, RepRap is theoretically cheaper, and also much much harder to build. MakerBot may lack a users manual and be in the alpha stages WRT engineering, but at least all the parts come in the same box and there are some websites that tell you more or less how to put it together. Its unclear if RepRap has such support. Additionally, the MakerBot software, and driving electronics firmware, is reputed to be superior to those of RepRap, though they do share some of the same parts. MakerBot software still has a beta open-source feel to it, but apparently its better.

    I'd still recommend MakerBot, its just very very do-it-yourself, and it sounds like the people that have had the most success have engineered their own solutions and had access to a machine shop, so they can laser cut their own upgrades, etc.

    I took a poll and the forum and it suggests that a realistic expectation for the maker-bot is 10 to 20 hours assembly, then approximately 6 weeks while you train yourself, debug the bot, and order maybe a couple hundred in upgrades. Some people are lucky and get things working well on day 1. Some people don't have the technical knowledge and that 6 weeks drags on for a year.

    So, this should be well within the capabilities of any partially employed techer, and the MakerBot was a good introduction for me, since I have very little experience with electronics, and no experience with robotics or polymers.

    If you are more mechanically inclined, and have more time, maybe the RepRap is the way to go ? I'd probably go with a DIY subtractive milling machine if I had the faintest idea how to build it. I'm told these get better resolutions on more materials, but can't create as complex of geometries.

    ReplyDelete
  3. Apparently, these are three guys in a garage, more or less. Approximately 1700 units have been sold, with constant upgrades all the time.

    The early batches we not always usable and needed experts to bring them in to shape. Now, the bots are basically usable, and maybe intermediate hackers can get them working. You still need upgrades to get decent quality prints, in my opinion.

    In a matter of months, I'm sure they will incorporate fixes for essentially all of the problems I listed above.

    The assumption is that these early flaws happen because they are simultaneously designing and shipping their product.

    so.. MakerBot : yes, but wait if you can, see if they fix these issues.

    things to look for :
    -- full re-design of the extruder controller hardware, possibly with higher current MOSFETs to remove the need for relays and other such nonsense for high current loads
    -- Moving their extruder to something called the "paxtruder" and the MK5 "drive gear"
    -- See if they figure out a surface coating for the heated build platforms that actually attaches to the plastic
    -- updating the hot end of the extruder to MakerGear parts : PEEK for the insulator, not PTFE, and a modular ceramic heating element instead of hand-wrapped wire
    -- updating the Z axis mechanism to something less terrible.

    ReplyDelete
  4. Here is a link to the a current forum discussion thread that I hijacked to solicit user stories. These range from "worked great, only took us a day to get up and running" to "it's been two months and I'm so depressed".

    ReplyDelete
  5. update : I finally contacted makerbot about the problems with my Z axis mechanism, after concluding that it really was a part defect and not operator error. They responded within 24 hours and tell me they are going to ship a replacement part. So, officially, if you get a Makerbot and the Z mechanism really isn't meshing, they will help you. Now, MakerBot Z axis will still have a slight, ~0.1-0.5mm wobble, but thats almost expected for the lowest of low end printers, and can be ignored if at least the bot is moving.

    Also, MakerBot is soon going to release an updated design of their extruder, which should eliminate all problems I mentioned with the extruder. Hopefully the new design won't add any new bugs.

    ReplyDelete
  6. hmm.. this link presents a viable counterargument for why it doesn't matter that MakerBot is flawed and you should buy it anyway.

    ReplyDelete
  7. I have to agree on NOT TO BUY the MakerBot or The Thing-O-Matic. Has any checked out the Purple Platypus BFB3000 ? It looks like Bit From Bytes is selling it for 2000 british pounds. I bought the MakerBot 3 or 4 months ago and it has blown a hole in my pocket about $1200 worth, and I have only been able to print one model, it breaks down every time. AVOID MAKERBOT AT ALL COST!. Most of these makerbot owners seem to be a bit dweebish in accepting a shotty product, and saying "well, oh it does need a little work, since its not an appliance"... basicly, that the machine should break down occasionally because it costs less than others. But face the facts. MakertBot sold you a $1200 product, A PRODUCT is suppose to work. period. I have basicly paid MakerBot in brooklyn $1200 to give me about $10000 worth of headaches and they have lost me about $20000 worth of business from clients who wanted quick proof of concept prototypes right away. I am in the process of seeing what Legal action I can take against this Brooklyn Firm.

    ReplyDelete
  8. A few thoughts on your troubles :

    I hear that the Up! printer is also relatively reliable, though it doesn't handle support material like this Panther 3D printer from Purple Platypus seems to.

    I would never rely on a MakerBot for business. If people are buying these machines for actual production, MakerBot needs to refine their advertising. If they aren't doing this on their own, perhaps a gentle push ( threat of legal action ? ) in the right direction would help.

    If all else fails you may be able to recover some of your money by selling the kit on e-bay.

    Also, I've finally got my makerbot operating reliably. I can try to fix yours if you're in the Boston area, of at least try to help with common problems. Presumably the forums are approximately helpful, but it still takes tons of _your_ time to get the thing up and running.

    The other thing is that most of the design flaws in MakerBot have been found and solved by the user base. But, rather than releasing an updated kit that resolved these flaws, MakerBot released the thing-o-matic, which solves a few of the makerbot flaws and introduces that conveyor belt which, from the sound of it, is also flawed. Theres this whole business on the side of correcting MakerBot design errors ( MakerGear, I mean ).

    ReplyDelete
  9. I've never seen anything on the makerbot site that says the thingomatic is appropriate for production in a business setting. I also happen to think that it's pitiful that anybody would resort to legal action about something like the makerbot. I have, on the other hand, seen plenty of places on the site where they discuss the fact that it's largely experimental and is for hobbyists. I have a thingomatic, it's still not set up perfectly but whenever I get a step closer I'm stoked. I have been in the construction business and the restoration business since I graduated from high school and I knew from the very beginning that this wasn't going to be a money maker. I don't know what gave anybody that idea.

    ReplyDelete