Low temp with ABS causes the delamination you are getting. ABS should be printed between 240 and 250. For some reason it is widely reported at lower temps such as what you are using (230’s) but that is not high enough to guarantee good layer fusion. Not sure where the 230’s number came from or how it became so prevalent, but it is dead wrong.
Most of the older hotends out there contained PEEK heat gap isolators, that would melt if they got above 245. Earlier firmware revisions and controllers were kind of iffy about real temperature vs reported temperature, so the general rule was230 as a safety margin to prevent hot end melts. Some of the older makerbot ones were particularly subject to early core meltdown. I think it was the type of plastic they were using. At any rate, 230 became the default temperature because that was the highest you could safely print reliably at that point. Its one of the reasons people think ABS can’t adhere well under any circumstances too.
There would be no need to over constrain. You could leave the front rod free to move in Y, but restrain it in X and Z and you won’t have a binding problem. It will require a creative mount, but it could be done. And you don’t care about Y since the other two rods will adequately restrain the extruder in that direction. The forward rod would reduce Z deflection, and prevent X rotation.
Thanks for the temp info for ABS. Based on what I’ve read on the forum I thought is was the lack of an enclosure. The room temperature where the TAZ resides varies between 59F and 67F (max); don’t know what it was when the Benchy prints were made.
Is the flex in the Y direction the Z mount itself or the X axis bars? If it is the bars, that flex should be reduced greatly by larger diameter rods without need of a third rod.
Probably true. But I’ve always hated the cantilever design. Especially when using a dual extruder. It’s not just about deflection, it’s also about stiffness, and resonant frequencies, which will greatly impact acceleration capability.
One thing about raising the ABS temp, that could create other problems printing 3Dbenchy or any small parts. If you print it again with higher temps, you’ll probably need to print two or three benchy’s at a time, in order to let each layer cool down so you don’t get other problems.
Lack of an enclosure won’t cause delamination, just warping. yes that’s a really cold environment for ABS though and will cause some issues with warping.
billyd wrote:
If you print it again with higher temps, you’ll probably need to print two or three benchy’s at a time, in order to let each layer cool down so you don’t get other problems.
Been a lot of years since I looked at gcodes, and just checked the RepRap/Marlin list for the dwell code. The G04 dwell code is implemented, and it would cover this problem instead of printing multiple copies. Do the slicers’ advanced features not make use of this? I would expect a slicer idiom available something like “filament retract - dwell - filament forward to extrude - resume”. The value for dwell time might be somewhat of a bygosh & bygolly value, but might be able to be estimated based on where the layer ends on the boundary and the next layer starts. If not in slicers, it should not be too hard a task to write a post processor to insert the code into the gcode file with the user supplying dwell time as an input.
If the rods are of sufficient diameter, I would think the stiffness and resonant frequency problems would be cleared. No machine on which I worked ever used a third “rod”, and the cantilevered mass was far greater than anything my little toy TAZ** will have. The “rods” were made larger to accommodate the increased mass of the Z axis; they do have an advantage of being supported full length rather than just at the ends as they are rectangular so it would be a factor. The relative strength/stiffness of the rods will go up very fast with diameter compared to what mass on the Z axis will have even with dual extruders. I have not looked deeply at the openrail done by piercet, but that really handles the vertical stiffness. It looks like it side loads the bearings and wheels rather than placing bearings top/bottom front/rear to handle the cantilevered forces; that rail is likely adequate to handle them with such bearings if the side loading does not.
** “little toy TAZ” is not a statement the TAZ is a toy - it is about the TAZ relative to machines I worked on. The TAZ is a machine tool, and can cause serious injury with burns and pinch points if not treated with proper respect.
Well there is the extruder assembly itself, which is composed of many 3D printed parts, and so the stock unit is fairly flexible in comparison to the steel rods. The third rod would help raised the effective stiffness of the extruder itself.
That said I agree it may be overkill. I tend to lean in that direction, which is definitely one of my shortcomings. I get a little ocd with this stuff.
The dwell could be used, but we are always fighting ooze during a retract. So if you used that, you would have to use a prime pillar, so good plastic was at the ready when the next layer is ready to be printed. Rather than deal with all that I usually just print two. But certainly not the most efficient use of plastic.
ps the alternative would be to slow the print speed down so it takes longer for each layer to complete. But I find with small parts this isn’t really that effective at preventing overheating, since the nozzle is always near by, never really giving the tiny part time to cool off.
The openbuilds extrusions are designed to take horizontal loads as well. The bearings in the wheels are ball bearings, not roller bearings, with two bearing sets per wheel. The wheels are trapazoidal in profile, and ride in a matching socket. You could add additional berings front and back as desired, however current tests indicate that is not a configuration that aids print quality. It basically just adds mass and cost for zero gain in print quality.
Thank you for the info about prime pillar. It occurred to me that it is likely that one will often have multiple small parts to print which can be done at the same time to alleviate the cooling time dwell needed. The pillar or duplicate unneeded parts might arise only occasionally.
For the moment, the work I’m doing is corrected just using software to correct the Z axis sag, but I suspect I will need to beef up the rail as parts become more demanding since the software only (partially) corrects for Z axis error, and does nothing for rigidity.
Great info on your mod and the openbuilds products - it will help me with a project down the pike. Took a brief look at the openbuilds web site and it looks like a lot of the problems I was anticipating will be quickly solved by using that “off the shelf” material! Are there adjustments to load the opposing bearings against each other? Not heavily, just enough to prevent rocking.
Each of the bearing wheels has a small groove in between the two insert bearings that make up the glide surface, When installing the wheel, the inner spacer and outer bolt head can adjust the amount of preloadd on both bearings, while the side with the cammed bearing handles the preload in the other plane. It seems to work pretty well.