Looks like early July for some new reviews on 3D printers, most specifically the Revolution from QU-BD.COM, the PiMaker and the 3D Systems Cube. The Revolution should be delivered to me before the end of June, the PiMaker kit is being delivered shortly, and I’ve had the 3D Systems Cube for 9 months (but no time to really go into a full article on it).
QU-BD Revolution
PiMaker (fully assembled units)
Wow, seems like an eternity since last September and I’m long overdue for an update. First of all, there will be some new machines reviewed here in the coming months – most notably the PiMaker kit from the Kickstarter campaign, and the Qu-Bd Revolution machine. Both are expected to arrive in March. Look for some musings about the Cube and Cube 2 to be posted as well….
As for my Solidoodle 2, I’ve made some significant modifications. So in no particular order, here they are:
Power Supply
I really got tired of the dangling brick power supply, so I replaced it with a 12V, 12A unit and placed it inside of the case. The original unit was 12V, 10A. Because I like to take machines to technical events, the original design is not good as the wires on the 12V end are just placed into a screw terminal block. So to accomplish this, I also made the next change…
X Endstop Location
In the as received design, the X endstop is a switch mounted on the right end of the X axis. The wires for this are routed through the chassis and back under the Z axis. This was in the way for the location of the new power supply. So, I removed the switch, installed a smaller microswitch underneath the extruder mount, and added a fixed endstop plug where the original microswitch was mounted. The wiring for the new location now is in the same bundle that goes to the extruder. The pictures show it well, so I’ve included those here for you to see. The result is that now all of the wires that come from the lower left chassis grommet (looking from the front) now stay on the left side of the chassis! The wood stop was made from some scrap MDF board. I’ve left the MDF unpainted in these photos – much easier to see, but I later painted it black to match the plastic.
Electronics Cover and Fans
Again for reasons of making transportation easier
(as well as keeping the electronics cool) I think this should be a standard part of the unit but then at the $499 price it is probably not possible. The design is straight from Thingiverse thing noted here by lawsy: http://www.thingiverse.com/thing:31325
About the only thing left to do at this time is to attach some acrylic covers to the sides and a door to the front. I did not order the door/cover combo. I am gluing magnets to the acrylic sheets that are thick enough to also be spacers so that there is a small amount of venting of the inside. As for the door, two common hinges and a knob with the acrylic sheet will do nicely. This will keep prying fingers out of the box, but the top will still be open.
After a long summer hiatus, I’m finally back working on the Solidoodle V2. Suffice to say it wasn’t ready to print right out of the box. Several issues were apparent – the bed was not level nor at the right gap from the tip, and there was a gouge out of the Kapton tape cover on the heated bed. Then I was bamboozled by an undocumented change to how the unit would get to the home position in addition to the reported temperature settings being 20C off of reality. These are all things one might not expect, but then again, getting an early unit from the start of production you have to be willing to work out some bugs.
After two firmware updates, bed leveling, replacing the Kapton sheet, and working on Skeinforge settings, I have my first successful print.
First Impressions
On the hardware side, the V2 is much better than the V1. The metal frame is well made, with crisp corners and solid construction. Actual threaded inserts are used instead of sheet metal screws into thin steel sheet. On the minus side, the electronics sit on the back side of the case with exposed wiring, and while this is low voltage, in the V1 the controller board was nicely isolated inside the case. I ordered mine with a heated bed which honestly could use a higher wattage heating element — it is slow to get to temperature, unlike the extruder.
Temperature Problems
OK basically if you set the temp of the extruder with Pronterface to 200C what you get at the tip is 218 – 220C! This is not good. Would somebody please save me the trouble of doing a thermistor calibration table for the firmware? If I set the extruder to 210C or whatever in the gcode file or from Pronterface, it should be 210C at the tip! (consider this a mild rant).
Parts not Sticking to the Build Platform
A huge issue. I spent a lot of time leveling, gap setting, and cleaning the Kapton. The build platform should be around 100C but mine never gets there. 85C is about all it can do, and then it has the issue of being wrong anyway with the thermistor calibration. I’ve not taken the time to measure the actual temperatures, but I did find an absolutely bullet-proof solution (pun intended). The 3D Systems Cube printer comes with “magic glue” to coat the aluminum build plate (there is no Kapton on it) and it washes off when done. Without a doubt this is some sort of PVA water-based glue [update: no, it is rather something like "Mucilage" of days gone by...] It is thick with a slightly amber color. It washes away with warm water in an instant, so it cleans off both the platform and part without a trace. Putting a thin layer on the Kapton surface, my prints have stuck with incredible success and are fairly easy to remove and clean!
Improvements
After hours of getting this to work for the first print, I’ll have to admit that at times I was ready to pack it up and send it back or unload it on ebay. I’m no stranger to these machines (as this blog should prove) but the lack of updated and machine-specific documentation, a serious temperature calibration issue (which could lead to hot-end meltdown), and some other quirks have proven to be frustrating. And, the first prints are still showing oozing and other issues that mean the Skeinforge profile needs tweaked. I’d like to see the electronics better protected. Some of the wiring cables appear to be too short and I’m worried that with more use they will break from fatigue.
I’m a big fan of competition in these sorts of things, and at the price it is a reasonable value if only the unit would live up to the results I got in the V1 right out of the box. BTW an excellent new blog on the V2 just went live in late August, so check out Ian Johnson’s Solidoodletips for lots of helpful material if you are looking to get your V2 set up.
Did I mention the Cube(tm) 3D printer? Yes, mine is working as I write this. A review will come next. How are the prints? Incredible right out of the box! 
UPDATE: After trying to make a known size calibration cube, it became painfully obvious that something was wrong. The X and Y dimensions were well within specifications, but the Z was crunched by about 15%. I thought of all sorts of things, including the firmware calibration factor for the Z screw but that was not the problem. The Z axis motor driver was way too hot and the idle voltage was well over the .500 V the troubleshooting page suggests. This is how it came from the factory, so I made sure all of the stepper driver boards were adjusted to the suggested values, or slightly less, and now things are fully functional. I’ll put up some prints of parts soon. One change from the V1 to V2 for the Z axis is that the V2 uses a 5/16 threaded rod and the V1 used a 1/4″ Acme screw (which is far more precise). But, such is the need for cost savings.
With the summer heat here in the East, staying inside in air conditioning was a more comfortable option. And, with the arrival of my Printrbot LC+ last week, building this long awaited kit was the perfect indoor activity.
The motors and other electrical parts come with pre-wired connectors so that part is plug and play. However, no cable management materials come with the kit, so I used my own spiral wrap to dress the runs.
Not So Good
Documentation for the build needs a lot of help. For the LC+ what you basically have is a 2D layout of the laser cut wood parts that are more or less in spatial relation, but no 3D views of any assembly other than the extruder, which is common for all the models. What you do get are (IMHO) tedious assembly videos that are specific to the LC+ without any printed documentation and mostly one (above view) camera angle. Mr. Drumm sometimes contradicts himself or gets chatty with really unnecessary explanations. There are times when I just gave up watching and fast forwarded to some frame with a sufficient view that helped to show the result. The folks at the H-1 project (www.seemecnc.com) have an incredible step-by-step build book that Mr. Drumm should emulate for future sales from his store. Could my kid build this? Yes, he’s an aerospace engineer. Could your 12 year old build this? Probably not without some serious help. While I understand the response from the Kickstarter campaign was over the top, one of the first persons I’d get under some contract for help would be a CAD draftsperson to craft detailed documentation.
The z-axis uses two motors – one to drive each threaded rod. I’m wondering just how well this will work, because if one motor skips due to friction, the tilt of the x-axis will be affected. As it is now, it appears that there is some warping from side to side, so that may be something to address when I try to run it.
To insure positive engagement of the x-axis endstop, I turned the screw around so that the head was facing the switch. Not sure this is in the video or not, but if you use the threaded end the alignment is off enough that it misses the edge of the switch lever. Since this should be a once and done adjustment, I don’t see this as a big deal.
The Bad
Yes, there is some bad to share. Most of the printed plastic parts left a lot to be desired. The biggest problem was they were so undersized they couldn’t be used. The other problem was that on the down-facing surface, there was a noticeable flare where the plastic melted, suggesting to me a too hot build surface. This was most evident in the drive gear for the extruder. The result is that unless you remove the flare, the mesh of the extruder gears is very poor.
The extruder body itself was also undersized. Thus you could not easily place the hot end into the hole without first reaming it out. In the large extruder gear, the head of the hobbed bolt could not be fit into the opening. I was able to use the bed mounts and the Z-axis motor to screw adapters, but I reprinted on other equipment all of the other plastic parts. Comparing the sizes didn’t need calipers, as it was readily evident to visual inspection.
The heated bed didn’t come with a glass cover (which was expected) but there were no clips to hold down the sandwich of glass and platform as shown in the web catalog shot. In my case, I’m having the glass made with the corners cut at 45 degree angles away from the mounting screws so that the cover will allow for a bed leveling design I’m adding. Although it is mentioned that you can cover the heated platform with painters tape or Kapton tape, the glass surface has the advantage in being very flat and uniform. A local shop made mine for $8.
Conclusions?
For now, I’m waiting to get the glass for the heated build platform, and complete the assembly and confirm my electronics hookup. Of course, you will end up with a PC power supply on the loose, but that is the nature of this unit. Is it worth $699 at the store? Well, stay tuned, as a review of a Solidoodle 2 is coming next week or so. Although it has a smaller build envelope (6″ cube) it comes fully assembled and tested.
These days the numerous offerings of 3D printers for home use is very reminiscent of the early days of personal computers when models were showing up everywhere — VIC20 anyone?
UPDATE: I’ve given it the “smoke test” and after powering up, it connected to pronterface and all axes work and the hot end works correctly! Still waiting for that custom glass piece for the bed! ;-( Build video coming soon!
UPDATE #2: Glass piece came but now a big problem – the heated bed is defective! Somehow it is a dead short, and every time I turn it on from pronterface the next thing to happen is the power supply goes down from overload. I have a very sensitive ohmmeter and it says “dead short.” So, thankfully I had another 8×8 mendel prusa heat bed in the shop. I used this instead for the one that came with the kit and all is well! I contacted printrbot by email asking for a replacement but no response so far… The next thing is to actually try a print!
UPDATE #3: Someone from printrbot finally responded to my inquiry and at their request, I sent the heated bed plate back to them for testing on July 19. Well, two weeks later (8/4) still nothing heard… I did just send an email asking for an update on my board. Compared to Makerbot (where I find the response is almost immediate), the customer service here so far leaves something to be desired….
UPDATE #4: After some fun with the glass company and a replacement heated bed, the unit is done! But due to my work obligations, vacations and such, I’ve shelved this project because my 3D Systems Cube(tm) is now on the way. So, maybe I’ll get the time to actually fire this up again and test a real part on it. I’ll get a report out then!
After a long wait (the Kickstarter project funded last December), I received a note from the shipping agent that my LC+ was on the way. So, sometime next week it should appear at my door and as I have done before, I’ll be posting assembly shots and commentary on the kit.
I received one of the first batches of machines sent out into the wild, being delivered in late March. I’ve run quite a number of jobs with it and am generally very pleased with this new model. I have the dual extruder version. Here is some early experience feedback from my one month of use.
First, the machine came incredibly well packed, with custom cardboard inserts and clever use of materials to cushion the bot. No damage was found from shipping. Setting it up was pretty simple, although I found the leveling of the build plate to be fussy (more on that below). My first test print from the SD card went fine. As I already had a supply of several color ABS spools, I opted for the “standard” order which comes with two spools of ABS: natural and black.
The good: it works! Part build quality is very nice and compares favorably to a uPrint ™ machine from Dimension that costs $15,000 that I have in my lab at work. In fact, it would be very hard for someone to tell the difference, with the notable exception that the uPrint machine only will accept natural ABS and they do not sell tinted material for it (for that you have to purchase the next model up at $19,000). MakerBot ABS works out to about 65 cents per cubic inch while the “professional” materials for the uPrint ™ are $4.66 a cubic inch.
The bad: well, a few annoying bugs cropped up the first month. Getting a part to stick to the build surface can be a real problem. I’ve followed the cleaning and leveling instructions and still have had issues with builds that have long dimensions on the platform. Raising the build plate temperature to 105C helped. And the 4 screws used to level the build plate are just barely long enough to do the job and retain some tension on the springs that are meant to keep things tight. One of the thumbscrews fell off during a build from vibration. I also hear that they have increased the thickness of the aluminum build plate in models shipping now -mine is slightly warped so that the edges bow up from the center which is the lowest point.
The x-axis limit switch is not mechanically fixed to the PCB it is soldered to, so with time it became angled up off the board and didn’t trip at the end of travel when homing the x axis. A spot of glue puts it in place permanently.
I also ran into a bug when switching from dual to single extrusion. Moving from a just finished dual extrusion job to a single extruder, the head not being used is still pretty hot. This seems to fool the software and so the machine actually times out with a head error and shuts down after a bit thinking that the unused extruder should not be hot. The solution I found was to let the machine cool down completely and then start the single extruder job. But this certainly slows things down if you are looking to build a number of parts in sequence.
And, a bug in early hardware (mine included) is that you can’t load or remove an SD card when powered up or bad things can happen. I don’t know what the exact hardware problem is. So far I’ve not encountered it, but do plan on powering it off before exchanging SD cards.
Also, just a small gripe – at this price, you’d think they could provide the machine with a coat of polyurethane on the plywood, but none of the wood surfaces are protected.
Even so, all in all this is light years ahead of the CupCakeCNC machine they started with (and my model still makes parts!) and higher quality than the ThingOMatic. Now if they would offer a good 3D scanner!
I finally got a chance to write some things that need posting! And in looking at my last post, it has been two months of silence at this end… So, here is what is in store for upcoming material:
1. An update showing the final, complete, and yes — working Sumpod. This includes the Version 2 hot end in place, and some wiring amendments. The machine is running the Sumpod version of Marlin (thanks to Stohn) and I’m using pronterface to send files to it. I’ll also include a sort of “wrap up” lessons learned file that completes my build instruction file set.
2. A review of the Solidoodle after numerous hours of printing. But for a few tuneup adjustments here and there, it has been problem free. If only there was a heated build platform the prints would be totally awesome (instead, as you might expect with ABS on a cold piece of acrylic, the first few layers are not even).
3. The H-1 (www.seemecnc.com) is awaiting final wiring – using a RAMPS board instead of the electronics supplied with the kit.
4. And, as I’m expecting a Makerbot Replicator(tm) sometime in the next week, there will be an unpacking and first impressions post. I’ve scoured all of the published documentation including the electronics, and one very interesting feature is that the new MightyBoard electronics package uses digital potentiometers to trim the stepper motor V-Ref lines.
So, stay tuned!
