I’m printing a bracket which swings around in 3 dimensions such that no matter what the orientation is there’s one arm that ends you stressed in the Z direction. I’m new to this 3D printing stuff so I’m wondering what the options are… Here’s a view of the bracket (piece in red) with the failure location circled in green and the force applied in the green arrow:
Is it being caused by the cross bar? Just delete that, test print it, break it, and see. I wonder if printing the crossbar is allowing the material to cool just a bit more while it’s busy doing that.
I don’t think it’s the cross-bar, it breaks where there are the fewest wall loops and the smallest vertical cross-section and where the biggest lever forces are applied.
If you look at the first image, the horizontal part is screwed down (just to the top-right of the green circle), and then the force on the bottom of the loop causes the bottom of the circled area to the compressed and the top to be stretched. What fails is the stretching of that top part, i.e. Z-layer adhesion or tensile strength.
It’s possible that if I lay the whole thing on its side such that the cross bar is vertical that then the Z direction is not too stressed 'cause it would mainly be in the center of the loop and while the force is applied there, there would not be a lever arm causing tensile strength. Printing in that orientation would require a ton of support, though…
I did a couple of tests… First I reprinted at 265C instead of the std 255 for PETG-CF. I believe that made it a tad stronger, but I’m not sure. Then I added an internal “I-beam” by cutting out a 0.1mm vertical slice causing the slicer to add walls. That significantly increased strength: I couldn’t brake it by applying pressure with my thumb, I had to use the palm of my hand. With a bit more modeling I could make that two internal I-beams…
This photo shows the failed original with just cubic infill at the top and the same part after failure with the I-beam in there. Both parts are only half of the original U shaped bracket to save filament and time.
I also made a model where a 5mm slice is cut away, printed separately in a different orientation for strength, and then glued back with cyanoacrylate. But I didn’t do a good job and ended up with too small a glue interface so it didn’t help.
How would you fix that? I’m using the std profile except that I’ve boosted temperature by 10 degrees.
I was trying to get some extra stiffness. What’s the issue with the CF? Bending strength and tensile strength in the Z direction are essentially the same for PETG and PETG-CF according to the spec sheets. I’ll give PETG a shot just to see. I don’t have any PC…
Geeze. 100% infill, please. If you’re printing for decoration, you want to use as little filament as possible, you use sub-100% infills. You want to maximize strength, you use 100% infill.
But also, you’re under extruding, or something’s not set right in the slicing settings. There should be absolutely no gaps here, where you can see there are gaps almost as wide as the extrusion itself…
I’m using the std PETG-CF filament profile on my X1C except that I did a flow rate calibration with orca slicer and I boosted the nozzle temp by 10 degrees.
Agreeing with @RocketSled here. Your layers look off. You can clearly see the lines on the wall and the layers on the infill. Have you done a flow calibration? Are you using default processing profiles or custom? I think you need to “squish” the layers a little more, and increase the temp too. With layers looking like that they are loosely attached to each other.
This is the surface that broke. what does the surface that was in contact with the build plate look like up close, and also the top-most surfaces?
In the upper left corner of the print, in the image I snipped in my previous post, where there are big gaps as the filament goes around the corner - that shouldn’t happen. I said “under extrude” except the exterior appears to be pretty good and I’d expect under extrusion to show up there, too.
Did you make any changes at all to the slicer settings? They turn orange if they’re not set to their defaults. Check for that…
If you did a flow cal, it may be bad. But… you could just try bumping up the flow a little higher than you calibrated and see if that makes a difference. Go up by like 0.01 or 0.02…
Thanks for the replies with suggestions! I’ve done some more tests… I printed 4 versions of a small I-beam profile that is just a couple of layers thick to see how the wall layers come out. I set top-layers to 0 in the slicer so we get to see the walls (the sample is dimensioned so there’s no infill). I have 4 samples:
in common: PETG-CF, 0.4 nozzle, 0.2 “strength” profile tweaked to eliminate the top shell layers, textured PEI plate
#1 std filament settings, auto-flow-calibration
#2 filament settings adjusted based on orca flow calibration, auto-flow-cal off
#5 same as #2 except flow increased from 0.95 to 1.0 and temp from 255 to 256
What I notice (prob hard to see in the pics) is that they’re all pretty OK in the long center beam, but there are gaps when going around corners. Samples 1-2-5 are pretty much the same, sample 6 is significantly better. So upping temp 10 degrees and increasing flow by 0.05 doesn’t help.
I did a max flow rate calibration to see whether that’s the issue I’m hitting and I don’t know how to interpret the results. I went from 2mm^3/s to 20mm^3/s and this is how it looks:
The odd thing is that the long straight sections are perfect all the way to the top but the shorter end section has these waves over 14-16mm^3/s. I rotated the part 90 degrees on the plate and the waves remain on that short section, so it’s not a X axis vs Y axis issue.
I did a belt tensioning and full printer calibration and the max flow rate test looks the same.
I recommend printing it with 100% infill and 1% magnification from regular PETG. Once it’s printed, put some baking soda in, cover it, and put it in the oven at 90 degrees for an hour, and it should all melt together. But the 1% increase is because the plastic shrinks when melted. There are many videos of this on Youtube. It is not possible to print such details strongly, that is the problem of 3d printing.
You can also press the printed part into ceramic clay, make an impression, and then press molten plastic into it, and you have a one-off part.
When looking for strength when using FDM printing you might have to print in multiple parts. I would look to do the loop separate from the other parts. This will allow you to place them on the bed to have the layers going along the longest parts to give the extra strength. Printing it in all in one piece is easier but you then reduce the strength properties. You want the layer lines to be as long as possible, short will result in failures when a lot of force is placed upon the part.
I would start using stock profiles / settings and only look to calibrate if those were not working.
I’ve been trying to address this underextrusion issue with not a whole lot of success. As far as I can tell it’s a slicer or bambu hw/fw issue when going around corners.
I created a simple I-beam model which I’m printing without top layers so I can easily see the wall layers. In the “long” straight sections (the whole thing is ~35mm long) the extrusion is good, but when going around corners there are gaps similar to those seem in the photo above.
Here is what 3 samples look like (ignore the black dots, they’re an attempt at getting the camera to focus on the samples instead of on the background):
left sample: ASA, 0.6 nozzle, “0.3mm strength” preset
middle sample: same except that one half has a modifier with “flow ratio: 1.2”, I can’t tell which it was, i.e., not effective
right sample has a modifier for the lower half with all speeds reduced to 10mm/s and it looks significantly better to me
The 3mf file for the right two samples is available for download
It seems to me that there’s a real problem when the walls go around corners and it seems to me that the print head needs to briefly slow down to allow the filament to actually fill the corner. Looking at it more precisely, these walls are laid down going counter-clockwise around the shape and it looks to me like the gaps are just before the actual corner. I suspect that the movement after the corner pulls the filament that was extruded before the corner across the corner in a diagonal. Dunno whether this is a slicer or firmware issue.
This is really blobby. Do you own a filament dryer and have you used it for this filament?
All your slicer settings are at their defaults, including the filament profile?
Why does the right most print have that center “feature” where the other two do not? Just lowering the speeds shouldn’t change the shape of the extruded lines. That third print is different somehow.
The printer does slow down for corners.
Your description would suggest that Pressure Advance needs to be tweaked. I think I probably agree with that. But it’s not the only issue here.
The center feature was due to the modifier cube in the slicer. Dunno what caused it. Moving the modifier around (after I printed) eliminated it.
I did dry the filament for ~12 hours, but I can dry some more. It’s been in the AMS since (a couple of days). This is PolyLite ASA using the generic ASA preset with pressure advance set to 0.02 (I can’t tell from the slicer UI whether that’s the default or not).
That must be a firmware feature 'cause I’m not seeing it in the g-code. I suppose there’s no setting to tell the firmware to slow down more for sharp corners… There are a bunch of posts of other people having trouble with rounded corners.
The printer will maintain constant velocity around a radiused curve. Don’t remember how that works, I think it’s handled by the MCU and not by the g-code itself. But a sharp corner requires the print head to decelerate to change direction. Speed must change for that to happen.
Filament drying is more an art than a science. The tools most of us have for this, a heated box, aren’t really all that optimal. Performance is heavily dependent on the ambient environment. I’ve had Nylon in the dryer for a full 24 hours, only to find I had to put it back in the dryer for another 24 hours. It was during a period a few weeks ago when it was really hot and humid out. Hotter air supports more moisture, so unless you’re using a vacuum dryer, the results are highly influenced by what the ambient environment allows. When it’s more hot and humid outside, you need longer drying time.
A key indication that filament needs to be dried more is if you get a long drooling extrusion from a heated-up hot-end even though the extruder motor isn’t pushing. But I agree with your observation that the issue with the gaps is present mostly at the corners. That suggests Pressure Advance tuning is wrong. Which can also happen if the filament is wet (wet filament will defy your attempts to tune it for PA).
