Posted by Alex R English on August 13, 2012
The Printrbot Kickstarter was a heck of a ride. After almost 7 months of waiting, our Printrbot PLUS finally arrived. We've not had much time to work with it since the build, unfortunately, so we'll have to do a post about its performance another time. For now though, we have a timelapse video of the build and some notes about the build process.
In the spirit of Printrbot, we took this build to the living room coffee table.
We've worked with quite a few printers, and even as experienced builders this build was a beast at 7.5 hours (assembly only, not including calibration, configuration, or any test printing). This was one of the longer builds among the printers we've used (more than twice as long as the MakerGear Mosaic). The factors that contributed most to the long build time were the complexity of design, the finishing required on many of the parts, the organization of the kit, and the format of the documentation. I'll admit that some of the build time was me making mistakes that may have been avoidable, but we'll get to that.
We had several missing parts, as we'll discuss later, but for the purpose of the timelapse we plowed through the assembly as though those parts were there.
The design of the Printrbot PLUS does an excellent job of decreasing cost, which seems to have been a primary design focus. In our opinion, the appealing price tag has come at the cost of user experience. Our experience was very much echoed by the users that commented in this thread over in the Printrbot Talk Forum.
On the whole, this was one of the more difficult build processes among the printers we have experience with, primarily because of design features that we found unintuitive to assemble.
Some of the laser cut parts appear to be symmetrical, and seem like they could have been symmetrical, but are not quite. Using these parts backwards means disassembling larger assemblies and going back many steps, in some cases wasting zip ties that came with the kit. Most of these parts are marked with an arrow to indicate which side faces the front of the printer, but at these stages in the build it was difficult to tell which side of the printer was the front. Perhaps a better design feature would have been markings to show where parts mate with each other rather than showing their orientation relative to the completed unit.
We were disappointed in the number of tools required. The documentation didn't address the tools that would be needed, which included several screwdrivers, hex keys, several pliers, some vice grips (in our case, as a stand-in for larger pliers), snap-ring pliers (pliers can be used if you don't have them, but it's uncomfortable), drill bits to ream parts, blades to trim parts, clips or scissors to snip zip ties, wood glue, extra zip-ties (only if you make mistakes).
There is exactly one piece that needs to be glued in place, even though the same kind of joinery used in the rest of the printer should have been able to be employed here.
Many of the parts needed various forms of cleanup. This is something we've seen occasionally from RepRap kits, but never to this extent on the more commercial printers. For the most part the laser cut pieces fit together well, but there were a few that were tight enough that they required shaving down with a blade to mate properly, or needed holes drilled out to finish incomplete cuts (there was only one of these) or to widen holes to accept the hardware meant to fit through them. Likely this is a result of overestimating the kerf of the laser cutter or cutting on one that had a smaller kerf than the cutter used to design the printer.
Note the stringing between the teeth of the gear and the z-wobble in the teeth of the smaller gear at the top.
Almost all of the cast parts needed to be cleaned up with a blade, many needed to be reamed to accommodate the parts they mate with. Likewise, the printed parts needed similar cleanup, holes reamed out, stringing cut off, bits shaved off, etc. The stringing was worse than we expected, and occurred in ways we haven't seen before, but we also understand how many orders they had to get out and how little time they needed to devote to each print. Some of the printed parts also had a tendency to delaminate when screwing into or through them (even after some reaming, though apparently not enough), or when reaming them.
The kit itself was less well organized than other kits we've worked with. Most kits have tended to have parts separated by type (all of a single part bagged together, for example, separate bags for all of each length of screw), or separated by build stages (all hardware for each assembly bagged together, for example all of the hardware needed for the x-carriage).
This kit had most of the hardware (nuts, screws, springs, etc.) in one bag, which made it more difficult to find the parts as you need them, especially given how many different (but close) screw lengths there are in the kit (another aspect of the design that was frustrating). Grabbing the wrong part by mistake was a common occurrence because of how similar some of the screw lengths were, and because screw lengths are interchangeable in some places but not others.
At the time of our build (and at the time of writing), the only official documentation for the Printrbot PLUS was a part number diagram for the laser cut parts and a set of assembly videos for the Printrbot LC. The Printrbot PLUS is similar enough to the LC that these build videos are mostly compatible, but the subtle differences were a source of some confusion during the build. The videos also skipped a few critical steps in the interest of time that caused mistakes that took more time and effort (and zipties) to undo.
The lack of a written set of build instructions or a still photo walkthrough was frustrating because it meant skipping around in the video to find the information needed. Unofficial assembly instructions and photos on the printrbottalk wiki helped, but were incomplete (they were also not linked to from the official site, which means a user would have to know to look for them).
The electronics connections were not documented well (which heater where, which thermistor where), and not all of the connectors on the board have helpful silkscreen labels. Electronics connections took more research, some educated guessing, and some trial-and-error, which shouldn't have to happen. I imagine (and hope) we've just overlooked some helpful resource.
Many of the kits shipped had missing or broken parts, as Brook has mentioned in his update videos. Ours was no exception, but we were taken care of quickly.
We received two extra 8mm bearings in place of a 12mm bearing. One of the 8mm bearings was defective (not Printrbot's fault), causing some of the balls to spill out upon trying to put it on a rod. Luckily the extras made this a non-issue. We were also short several nuts and bolts. Printrbot was very responsive to these problems and replied to our email within the hour; the parts went out the next day.
There was one issue with what we thought were missing parts that turned out to have not been a kitting mistake, but a design versioning issue. The printer seems to have been designed for square nuts (and the assembly video shows square nuts in use), but the kit we received used hex nuts. These worked interchangeably everywhere in the printer except in the assembly that mounts the extruder to the X carriage.
You can see the problem in these photos. The plate with the square holes is supposed to slide down over the hex nuts as they rest in another assembly. The problem comes from the corners of the hex nuts sticking out further than a square nut's edges would, colliding with the edges of the square holes in the wood plate. Hex peg in a square hole.
How the square nuts sit in the unfinished assembly:
The hex nuts in place, the corners extend out further past the edges:
The plate in place, locking the square nuts in place; a very comfortable fit:
The hex nuts in position with the plate resting on top. The plate will not press down without forcing it, risking bending the rails or breaking wooden pieces:
The holes in the plate that slip down over the nuts are square and hex nuts do not fit them. Others have reported this problem too, so we knew we weren't just missing something obvious. Upon reporting the missing square nuts, we were told that the hex nuts were compatible there, a point we disagree on. You can force them, but this made us nervous. We sourced the square nuts ourselves, as the support person from Printrbot recommended.
We reported the problem to Brook, along with the photos above. He indicated that forcing it was the solution at the moment, but that he would be updating the design to work properly for future kits. We're glad to hear it. Perhaps our plate was cut with slightly tighter holes than others. We went ahead and used our square nuts.
While we haven't had much time to work with the printer since its assembly to evaluate its performance, we do have some thoughts on its design as they apply to the end-user assembly experience.
The process for leveling the bed (documented in the Getting Started Guide) involves shims between the bed and the parts that mount the bed to the rails. It is very difficult and time-consuming to get a precise level dialed in as compared to other printers. With four attachment points and no springs involved, we also worry that the bed (or a y-axis rod) can be warped (possibly permanently, but definitely temporarily) if a corner is tightened down when the corners aren't properly shimmed. A warped bed means you don't have a flat surface to print on and a warped rod means your bed won't travel in a straight, level line. When leveling the bed, we recommend leaving the corner opposite your reference corner slightly looser than the others to try to avoid this potential problem.
Precise and easy bed leveling matters most when printing at high resolutions, when getting going with new materials, or when printing in unfavorable conditions (like varying humidity, which can cause wood printer frames to swell and warp slightly), but it does have an impact on first layer adhesion and print quality all the time.
The lengths of cables in our kit weren't well suited to good cable management, which made it difficult to get the cables out of the way where they won't get snagged and where they look nice. The power solution is very hacked together, using a PC power supply with adapters and jumpers, but it does work and does probably reduce costs.
We were actually surprised by the hot-end. It came pre-assembled and seems well designed, was easy to mount, and looks decent. We're hopeful for its future in the community.
While our comments here on the build process don't shed the most positive light on the printer, the Printrbot PLUS is probably a viable option for those most sensitive to price who are willing to put some time in. For the user who wants a trouble-free build, this may not be the best option, but if you're willing to tinker you might do ok.
Since we're still yet to have logged much time using the printer, we've tried to speak only to the build process and not to operation. If you want more opinions, the printrbottalk forums are a great place to ask questions and find answers.