An Engineer’s Perspective on High Speed 3D Printing, Part Strength, and Accuracy
Since the release of the Bambu Lab X1, there’s been a lot of discussion around high speed printing. Two common things we hear are 1) high speed printing results in poorer mechanical properties and 2) high speed printing results in poorer accuracy. As engineers, we wanted to do some research to see if these statements are really true. This blog post covers some of the research and findings that we’ve seen.
High Speed Printing = Poorer Mechanical Properties
Surprisingly there aren’t that many research papers out there talking about this. Furthermore, no papers talk about print speeds that are much higher than 100mm/s, which we are easily breaking into now with new printers on the market. Here is the TLDR of what we found, along with images and links of the research papers.
In this Nature paper using PLA - researchers did find an inverse relationship between printing speed and print strength (in most cases, tensile strength), however, these same researchers also concluded that the decrease in strength caused by print speeds were minor compared to other factors such as raster direction and/or XY vs ZX strength. For reference, an example datasheet for BASF PA is shown at the end comparing the massive difference between XY and ZX strength. This is known as anisotropy and is one of the major disadvantages of FDM/FFF in general.
This paper investigated the affect of various parameters on a new type of PLA composite. While the specific material type here isn’t of interest, the microscopy images and conclusions are. The takeaway here is that, again, an inverse relationship is observed, but it’s small. The interesting part is that high speed printing can introduce artifacts, gaps, and inconsistencies during extrusion which seem to be the major contributors to poor mechanical strength as is visible by the microscopy images. However, innovations like linear advance, automatic calibration for linear advance (eg. the Bambulab LIDAR), high speed filament, and higher flow hot ends may solve this issue entirely.
This paper was one of the few that tested with ABS. Unfortunately we could not access the entire paper, but it’s interesting to see that their results showed higher speeds resulted in better properties. It’s unclear why this is. Is it because ABS is a amorphous polymer and behaves differently? Or was there something else in their experimental setup that caused this? Take this with a grain of salt.
This last paper again confirms an inverse relationship between speed and strength (again only with PLA). However, yet again, this paper drives home the fact that part strength decreases mostly due to defects during printing caused by inadequate extrusion.
High Speed Printing = Poorer Accuracy
While we didn’t take the time to search papers for this specifically, we have two thoughts here:
Because high speed printing is fairly new, we don’t believe enough time has passed to fully optimize and characterize all the different impacts parameters can have on part strength and accuracy. For example, a challenge with high speed printing is that you need to have better cooling in order to solidify the deposited layers quickly enough that they are solid by the time the next layer occurs. However, cooling too much results in poor interlayer adhesion (as shown by CNC Kitchen). Cooling too little results in poorer feature resolution and accuracy due to “sliding processes between successive deposited layers” (quote taken from this paper). To us, this means that theoretically there is a sweet spot for cooling that maximizes accuracy and part strength.
With the innovation of resonance compensation in 3D printers, and the overall reduction of mass of moving components (eg. the extruder, hot end, bed), all help improve (and are necessary for) accuracy during high speed printing. Stiffer, more rigid machine designs should also help here, as they do with machine tools such as CNC milling machines. This is something that engineering can solve. While it may be true that accuracy on a poorly designed printer that wasn’t meant for high speed printing may be worse, this doesn’t mean that high speed printing in general means worse accuracy.
Final Thoughts
We wish that there were more papers testing non PLA filaments like ABS at speeds >100mm/s. Most of the literature is using PLA, which isn’t an engineering material. Furthermore, since high speed printers are fairly new, few papers push the speed limit. Most likely because the hardware they have present isn’t capable of printing at higher speeds reliably.
It would be VERY interesting to see these tests done with a machine that was designed to be high speed and high flow. One could make the assumption that any decreases in part strength would be much less because that machine would be able to handle the speeds and flows required.
As an engineer, there are HUGE advantages to high speed printing, even if it results in minor decreases in part strength and accuracy. The ability to iterate at 4X the speed or more is transformational. Furthermore, if part strength and isotropy are necessary, other prototyping and manufacturing methods should be considered. FDM/FFF does not solve all problems.
Anecdotally, we have not noticed any practical issues with part strength when printing at 4X speed on our Bambulab X1 vs our Raise3D E2. We have changed the slicer settings from the default ABS settings though because we found that the default Bambu Studio settings for ABS had too much cooling, resulting in poor interlayer adhesion. Once we dropped from 80% to 50% we achieved parts that we couldn’t break easily. Interestingly, we print with zero cooling on our Raise3D E2 and get excellent strength parts. But again, this wouldn’t work with high speed printing because you need cooling at high speeds…so this drives home the point about speed/cooling optimization.
The 3D printing industry has been so stagnant for so long, that the introduction of high speed printing into low cost printers is very welcome in our opinion, and we believe that with the continued innovation of the industry, any issues associated with it can be improved on!