I reviewed the Polar3D printer. This is not a typical 3Dprinter. It has a unique hardware design (radial) and has incredibly robust web-based control software. Worth a look for sure at $799 (even less for educators)
Tks for the great write up : )
Thanks for this excellent review of this intriguing printer.
I am a sucker for all-metal machines, and this one fits the bill just like my metal Printrbot. This could well be my next printer.
Just one question: You wrote about the hardware that it has a Raspberry Pi micro-controller, but that is a small contradiction - the raspberry pi is not a micro-controller, so did you mean that it has both a Pi AND a micro-controller as well?
How well does it print near the center of rotation? It looks like most of the pictures show printing away from that area. It seems to me that for some features the table would have to rotate very rapidly while printing in that region to maintain linear velocity.
@Jason_Doege , I talked with this guy at Midwest RepRap fest, and he has logic built in that dynamically adjust speed for just that circumstance.
He would have to. But I think there are geometries that mathematically require (effectively) infinite rotational speed. Imagine for a minute that you put the apex of a 90 degree corner right on the center of rotation, for instance. Of course you wouldn’t need actual infinite rotation, but it would have to rotate very very fast to maintain print quality in that circumstance. Fast enough that there might be issues with centripetal force on the model if any of it was near the edges. Plus the motor response would change quite a bit as mass was added to the platter, especially near the edge.
It is one of the reasons I like and chose the #DeltaMaker . Its moving mass remains constant and the forces on the motors have a constant and predictable range.
Jason, there will always be artifacts at the very center, but most of the time you would just adjust the models position to place it so there is a solid print area at the center or nothing at all if you wanted no artifacts.
We suggest moving the parts to the edge for speed as well, since the plate doesn’t have to rotate 180 degrees to get to the other side. Those are the only downsides of this design… however, the benefits are much larger… for example, we have a 4" travel, but can build on a 48" sq area… compared (directly) to the MakerBot Mini’s same 4" travel only giving you 16" sq area.
It’s also significantly more stable in that we also have no requirement for bed leveling, since we only print along the radius, which is exactly between the two linear rods the “shuttle” rides on… even if it is off, it will only rotate around that radius… which is always in line with the travel motion. This leads to a very very simple system and high reliable.
As for the speeds… we do dynamically adjust the speed… but there are limits to how fast we can go… but we also adjust the flow rate to compensate for when we can’t change the speed any further… that allows us to maintain the quality there.
Yes, polar printers have a “singularity” at the center of the build table. The early versions of this printer would actually crash if you tried to use the center of the build table. (Divide by zero error I believe.) That was fixed with custom firmware code to handle the center a little more intelligently. Short version is, you have to split gcode motion segments that cross the center, and insert a special “rotate only” motion to the queue to get through the singularity. This is likely to cause some minor blobbing due to extruder flow lag, but I doubt most people would notice.
Aside from the center singularity, there are two meaningful issues with polar printers:
- Rotary motion stages tend to have more backlash and lower resolution than linear motion stages (although this is fixable with good hardware design)
- The print resolution varies greatly with radius, so you end up with high step rates near the center and low precision near the periphery.
I’m sure it works fine, and it’s certainly very cool. Not how I’d personally build a printer though.
@Paul_Gross , It has both a Printrboard (Rev F) and a Raspberry Pi II on board. It runs Marlin (stock) firmware… and all of our custom code is on the Raspberry Pi, including the polar coordinate conversion routines, which allow you to use any 3rd party software, like @Simplify3D or any of the open packages as well.
@William_Steele If it’s stock Marlin, does that mean the acceleration calcs are all done assuming Cartesian kinematics?
@Ryan_Carlyle , actually, we’ve never had an issue with it crashing at the center… it’s always been handled since day one with our software… also we don’t insert a “rotate only” command… Marlin’s logic simply sees it as a move command. With small segments (think about G2, G3 arc commands) we already move with small enough motions that the printer doesn’t know it’s crossing the singularity. Divide by Zero points can simply be skipped, since at most you will miss a .05 mm point, which will be filled in quite nicely by the next segment.
As for the print resolutions, we also don’t vary that… the mechanical system was designed for 4/1000s of an inch at a 16" radius… which is much higher than most Cartesian based printers can achieve at our 4" radius. We then “down sample” the points towards the center to match the vertical resolution so the print is quality is consistent across the entire object.
@Ryan_Carlyle , yes, but it turns that the motor acceleration capabilities are what’s important… and they don’t know they are in a polar printer… they just know how fast they can accel or deccel… which is just a simple setting in Configuration.h.
@William_Steele I must be thinking of a different polar printer from a while back then. The similar name made me think it was the same project. Interesting implementation, thanks for taking the time to explain.
Yeah, now that you say that I seem to remember a polar printer that couldn’t print in the center 1" or so… was it this one? http://forums.reprap.org/read.php?185,178379
@William_Steele Wasn’t thinking about that one, I meant the Pi maker… that… I guess you made as well? https://groups.google.com/d/msg/3dprintertipstricksreviews/VuMOdtvNmmY/zfSvNgIM3swJ Glad to see the technology has improved since then, anyway
Oh yeah… that was my first polar based printer. The only reason it couldn’t print in the center is because I hadn’t written that portion of the code yet! 
(In fact, it could only print in one quadrant.)
I pulled my original prototype out the other day and used the modern software with it and it ran flawlessly… it actually amazed me!
@Paul_Gross Looks like Bill Steele answered this correctly (which I was unable to do 
@Jason_Doege great question - i answered this on the blog too - but the Polar3D guys recommend putting smaller prints mid way between the center of the bed and the perimeter to avoid that center spin spot having build up. I experienced that in a very minor way on one part and wouldn’t expect to see too much on a larger part - but i need ro experiment more.