My journey to choose a high end 3D printer (Raise3D E2) — Part 2
Ok, so life went wonky and I didn’t get a support request out to Raise3D yet, but I’m going to share the outcome of my first print on the Raise3D E2.
My goals at this point were :
- Will it print?
- How well would it print?
- Single extrusion
- Does it ring?
- How does the sliding Y platform effect the print?
As mentioned in the previous part, the test part was from the following
I printed the top part which was really something great in my opinion. I need to test the same print on other printers to compare, but let’s look closely at the outcome of the top half of Howl’s Moving Castle.
In the beginning we started printing. This is after approximately 1 day of printing taken with my iPhone from up top and through the cover. I didn’t want to interfere with the print.
The print was coming out absolutely amazing.
How it was printed
Let’s talk a little about the print configuration.
The filament
The filament used was a fresh roll of OEM BASF PLA but without a label as it was “for evaluation only”. From my experience, it’s been some of the highest quality PLA filament I’ve encountered, but BASF really doesn’t mess around. This was a sample sent to a vendor who would eventually put their own label on it and they went through a process of testing MANY different PLAs before falling in love with this one. I simply prints perfect every time. I can safely say from testing on multiple printers that if there are flaws in a print from this PLA, it’s either due to the printer itself or how the filament was stored.
The settings
I would love to go into a lot of details on this one, but all I did was to choose [Raise3D] PLA and change the layer height to 0.05mm.
Wall separation (under extrusion?)
So, this is a common occurrence on most printers when you start reaching that 0.05mm per layer condition. To handle really really thin layers and not create bubbles, a gigantic nozzle like 0.4mm has problems dealing with thin layers. You’re pressing a 1.75mm filament through a 0.4mm nozzle which is round and from a 3D perspective, the extrusion is more of an elliptical cylinder as it comes out. There is a limit of how far you can squeeze the round nozzle shape to an elliptical shape before the proportions don’t allow it to extrude widely enough to compensate.
I believe slicer software should recognize this condition and rather than attempting to extrude more, when it’s slicing, it should consider the 3D physical properties of filaments and move parallel segments of filament a little closer together to compensate. So, at 0.05, without doing any math, rather than distances between parallel lines being the 0.4mm difference, a “feel good distance” might be 0.3mm or even 0.25mm. This is something I believe should be easily fixable with an update to the slicing algorithm.
Minor annoyance at start
So, this is something I both love and hate. By default before starting the print, the printer performed the full bed mesh leveling. I believe the thermal conditions were similar or identical between this print and the previous test print made by the employees of the shop evaluating whether to sell Raise3D products. As such, it should have been able to use the saved settings. I hope this was just due to the very thin layers I was using rather than an “every time” kind of thing. After all, a proper mesh levelling takes time and Raise3D seems to be extremely thorough. I didn’t count, but I believe it sampled at least 20 points.
The Finished Print
A raft?
I hadn’t really thought about it, but given the absolutely perfect adhesion to the bed I encountered and what I believe was a nearly perfectly flat bed + mesh levelling before the print, I think it was a little silly that the default profiles in IdeaMaker include a raft when performing a single extruder print.
Rafts on IDEX platforms are REALLY useful when printing in duplication or mirror mode. This is because the Z-offset between nozzles can be corrected for by spewing some very thick layers at the start. For single extruder prints, it’s meaningless and generally just leaves artifacts on the base of the print. While the raft separation for the print seemed to be quite uniform and nice, it’s just an unnecessary annoyance.
The good new is that I’m aware of this now and will simply print without the Raft when it isn’t needed in the future.
Print time
I am very happy that the printer did such an amazing job. Compared to a printer which moves the Z position of the platform rather than the constant motion of the Y, it seems to have had a major impact on print time. When configuring for a near exact print on Cura with my extremely modified Ender 5 Plus, the print time was more than twice as long.
This is something to be expected though. Even though I’ve been AMAZED at the precision of the Y-axis of the E2, it’s still basically sliding the platform back and forth and back and forth. The platform is light weight, but if you combine the weight of the platform and the weight of a large print, it is extremely likely that even a great Y-axis like this would experience some noticeable layer shift if it moved quicker. This is a case where the Ultimaker S5 or maybe the Raise3D Pro2 would be significantly superior.
But since this would be my “make it perfect” printer… well next best thing to perfect when compared to my Form2, I don’t have an issue waiting a bit longer for excellent results… and as you’ll see as we continue, the results were absolutely superb.
Overhangs
It’s difficult to see in this picture, but the absolutely flat platform which was printed… dangling from mid-air… with nothing but a narrow rod for support came out superbly. If I were to measure the quality of a slicer/printer combination, this one picture would be enough to convince me to spend the big bucks on a single extruder printer. It was nothing short of a thing of beauty.
I am looking very closely at the print from above, so while the print looks like it’s practically mountains and valleys here, it’s really nearly perfectly smooth. If I were to consider painting this, I might decide just to use two thin layers of airbrushed primer rather than sanding first. The results were in fact superior to every other printer I’ve used in the past.
This overhang was amazing. Not only did it handle a very complicated overhang and bridge nicely, but there were very few artifacts under the roof as it was printed. This meant that IdeaMaker performed incredibly well when choosing when to print the outside wall of each layer. Cura does well in this case too (Simply3D flopped) but as will Cura printing on an Ultimaker and knowing how really make best use of the printer to stretch the laws of physics, IdeaMaker and the E2 were certainly in the same class.
I wanted to start looking at the bridging and the overhangs a bit closer, so I broke out my digital microscope to get a better look.
You can see that the bridges were in fact bubbly and dangly and a bit of a mess, but if you consider that each of the absolutely picture perfect layers that adhered as expected were 0.05mm thick, you can tell that it did very well.
I think if I were to make any change to this image, I would look for a coasting option and a possible minimal preextrusion option in IdeaMaker to start and stop a little better.
The bridging and overhangs are really a problem here, but I intentionally printed this model with no supports to see precisely how it would deal with this circumstance.
There were no dangling filament strands, it actually corrected for the problems of gravity pulling down on soft melted plastic quite well.
I think a possible great fix here is either to move the nozzle just a hair faster so as to avoid letting the filament droop and to apply a little more tension, or maybe to increase cooling substantially when making these bridges… which unfortunately will still be a big problem on the next layer. I think that the folks at Raise3D could probably ace this problem with some experimentation and some profile tweaks. It probably wouldn’t require any changes to the printer or the algorithms.
To print this… the slicer was smart enough to build the supports and print concentric rings to avoid problems common in other slicers. IdeaMaker really did well here. And while we’re here.. look at two things
So close to perfect
Looking at the image marked “Another angle” above, we see that handling of arcs went VERY VERY well. Handling of corners… that could use a little work. The pattern around the window frame should have been cleaner or more consistent at least. It really seemed to struggle with this. But for reference, consider that these are VERY small details.
Does it need better acceleration control?
I think this printer could either benefit greatly from an accelerometer on the Y and X axis, or it may be possible just through some impressive g-code voodoo to predict sharp turns and tune the acceleration around corners. But I think for sharp angles and the tiny tiny amount of ringing on Y-axis, the printer is falling short in acceleration control.
Super-mega-hard small features
I don’t think this feature seen here is more than about 1.5mm wide. And it printed… flawlessly. I mean this was just beautiful. I want to test this on the Ultimaker S5 for a comparison… just this feature and perform a stringing test, but as you can see, it not only printed the layers perfect, but it also printed the tiny little indentations which mostly were actually part of the design perfectly.
Conclusions for this part
I had my concerns about ringing from the Y-axis… and compared to the Ultimaker and the Ender 5 Plus Plus Plus, it is noticeable under a microscope. This is nothing a simple accelerometer on the Y-axis wouldn’t fix, or they can add acceleration control to the software or the profiles. Either way, there was a little itty bit of ringing.
I didn’t make it a requirement, but I have a thing about how my 3D prints should be nearly water tight if not air tight. With the exception of the wall gapping, I think this definitely passed the test. The seems per picture perfect.
The speed was slower than I care for, but I’m not buying this printer for speed, I’m buying it because I want a printer with this kind of precision that will actually fit in the place I have reserved for it (not entirely fair since I built the table using the measurements for this printer).
I feel the thermal control seems extremely good, it’s certainly pro/prosumer grade but the software could do a little more to deal with complex bridges.
I told my wife this morning that if it passes the dual extrusion test for water soluble support, I’m bringing this baby home… but only if the store agrees to support the printer directly. Thanks to the absolutely “piss-poor” handling of dual extrusion by the FlashForge Creator 3 and and incredibly brilliant handling of dual extrusion by the Ultimaker S5, I have both extremely high and extremely low expectations for dual extrusion from an IDEX printer. This is just something I believe a raised nozzle printer will do better at every time. No one cares about IDEX for mirror and duplication, if you need to print single extruder twice as fast, just buy two cheap printers. It’s cheaper and probably better. The only reason for a pricey dual extrusion printer is to have two nozzles at different temperatures.
A little last note
I don’t currently have either a proper barometer or humidity sensor that will fit nicely within the spool storage area of the printer. Once I get them, I’ll attempt to measure relative humidity inside and outside of the printer. Since I intend to use PVA and nylon, both hygrophobic materials, and because the printer is really meant to hold to spools itself, I want a good solution that will support leaving an expensive roll of PolyDissolve in the printer for prolonged periods if possible.
