Short, narrow bridges not printing consistently in PETG
Hi, I've been trying to print this digital sundial ( https://www.thingiverse.com/thing:1068443) with PETG. The model has very narrow (single perimeter) and thin (1-10 layers) bridges of about 6mm length. Quite a few of these bridges either did not print at all, or poorly with only full bridging from the 5th layer onwards. At the same time, some of them printed perfectly. Not sure what it causing this. Slicer did not recognize these as bridging, so print speed and cooling was unaffected. Most were printed in continuous fashion, that is, as one long perimeter so retraction seems not to be a factor either. See pictures for the results. Any suggestion what might cause this and how it could be resolved?
Welcome to the community! 🙂
Mojoptix's digital sundial is possibly one of the hardest prints you can attempt on an FDM printer ;). The original model was developed algorithmically by using the sun's rays to "cut" into the block so certain segments only light up when light is incident from certain angles. As such, there are a lot of extra bridges because in the real world the different times will overlap so you can optimise further by removing extra bridges (essentially, you only need one bridge to block light from a particular angle, so in cases where 2 or 3 bridges are in the light's path you can remove some of the extra ones).
So, here's an optimised version of the sundial which has been cleaned to remove the extra bridges - https://www.thingiverse.com/thing:1149469
And here's a further refinement where the four segments are seperate, to help you in troubleshooting the settings without failing the whole large print every time - https://www.thingiverse.com/thing:1628706
I suggest you do all your testing on thing 1628706 🙂
As for the bridging problem - can you attach your slicer profile? Are you using slic3r? If yes then I strongly suggest migrating to PrusaSlicer. It works with all brands of printers (not just Prusa) and has a lot of upgrades compared to slic3r which will improve your print quality. On PS I've found better results when I untick "detect bridging perimeters" but of course this will depend on which slicer you're using and your settings. I will give you a more complete list of suggestions when I've analysed your profile 🙂
Looking forward to hearing from you!
Thanks for the extensive answer and references to other versions of the design - I had not find the simplified model before so that will come in handy later. Interestingly, in the comments of the segmented model 1628706 I see others commenting on the same issue with the bridges ... I did make a reduced design to test the issue, though with limited luck. This is the version attached in the PrucaSlicer 3mf-file that I used with my Prusa MK3s. The detect bridging perimeters was unticked - not that it mattered though, also with other models I've not seen the slicer being able to detect bridging perimeters when they are obviously present.
There are some odd settings in that profile. Is there a reason why you're using only one wall (perimeter) and 9/7 top/bottom layers?
I notice the bridge flow ratio (under Print -> Advanced) is 0.7. I've had good bridging results setting that to 1.0, perhaps the extra material deposited by raising that will improve the level of support provided in the initial bridge layer and hence the overall quality of your bridges. Having said that this setting most likely won't be applied if "detect bridging perimeters" isn't ticked, though oddly the colours in the preview show it using normal perimeter settings (rather than bridging settings) regardless of whether that checkbox is ticked or not 🤨 . Further research reveals this may be a bug in the way PrusaSlicer handles bridges that are also thin perimeters, as in the case of this print (ref: https://github.com/prusa3d/PrusaSlicer/issues/3320 ). I don't believe that is fixed yet, so unfortunately you'll have to work around this limitation by using settings that apply to the whole print rather than being able to apply specific bridging settings/overrides.
Physically, what probably happening is that since you're printing very fine layers (0.1mm) it's very easy for the thin, fragile thread to break when making that first layer of the bridge. In fact from your picture, it seems the thread starts fine, breaks in the middle then deposits the droplet of plastic at the end, so you get a "vee" shape for the underside of the bridge (material in the start and end points that eventually grows closer every layer and eventually merges in the middle).
Other suggestions - lower the temperature a bit if possible (I print PETG at 220-230°C) and increase the max fan speed to 90 or 100%. Since PrusaSlicer isn't detecting bridges, you can't rely on a bridge speed override to give you high speed during bridging, so you'll have to increase it for the whole print to have it available when needed.
Reduce your extrusion widths to 0.4mm for all of the features (under Print Settings -> Advanced) to be better able to print the thin features of this print. Like so:
Consider printing with adaptive layers. On a test run I found that it seems to print the whole thing at 0.12-0.13mm which will be more resilient to breaking and give better bridging than the 0.1mm you're using presently (not to mention much faster printing!). Here are my adaptive settings:
Lower your accelerations to 500mm/s2 under Print Settings -> Speed, and consider also lowering infill speed to 50mm/s or 80% instead of the 80mm/s it's at now (if 80% was the intention, add the percentage sign in the box because I believe not including the unit makes PrusaSlicer interpret the value as mm/s).
Lastly, do note that you're missing a lot of the thin walls even in this small section because a 0.4mm nozzle is a little too big to print all of them. You might have significantly better results going for a smaller nozzle like 0.3 or 0.2mm (still with 0.1mm layer height) though keep in mind that a 0.2mm nozzle will print approximately four times slower than a 0.4mm nozzle (so a 2-hour print becomes an 8-hour print! 😣 ).
Hope this helps, let me know how it goes! 🙂
Thank you again for extensive explanation as what may be the causes of this. I feel the breaking of the thin thread is indeed what has happened. I will give this print another try shortly with the suggested settings, first with 0.4 nozzle and if not satisfactory with a 0.25 mm nozzle I got recently. I recall I chose single perimeter to significantly reduce print time of the full model as I did not feel the model needed double perimeters per se, and kept it that way. It's been months since I printed this and it was in my early days of 3D printing so I did not make a habit of keeping the 3mf files, so I can't tell for sure what settings I used on the full model that failed, assuming though most of them were default settings taken from Prusa PETG filament, except for the perimeter setting. However, I do not recall changing bottom/top layers settings at all; and upon checking now, it seems that I have in my PrucaSlicer default settings for horizonal shells that do not correspond with the supposed defaults, i.e. I have 0.6 mm top shell thickness and 0.5 mm for bottom as default, whereas they should be set at 0, according to the 'pop-up window' when hovering above the setting ... this is the case on 2 computers I've worked with, so not sure why this seems to be off. Nonentheless, it's not a factor for this problem.
I'll keep you posted when I have results. Happy New Year!
We’re a bunch of 3D printing and design nuts. We just want to make great 3D design available to everyone.
Good design takes time, great design takes a process. You can learn this with us, to build your skill set in this rapidly expanding market.