Hey everyone! 
I’m a small manufacturer of musical instrument accessories and have been using additive manufacturing from the start — first FDM, then later SLS. I’ve been running the Formlabs Fuse1 ecosystem for almost three years now, and it’s been a game changer for my business.
Most of you are probably familiar with the Formlabs SLS workflow, so I won’t dive into that here (maybe in a separate post). Instead, I wanted to share the results of a recent dimensional accuracy check I did on some of my parts printed with Nylon 12. Formlabs offered me Nylon 12 powder for the deeper analysis of the results of my recent dimensional accuracy check for the parts printed with this material, so here we are.
TL;DR: My real-world Fuse1 printed parts tested here actually exceed Formlabs’ stated accuracy specs.
My products don’t depend heavily on absolute dimensional accuracy (within reason — around ±0.25 mm for interlocking features), so I don’t usually check measurements across runs. In fact, I’ve never checked these particular parts before, since I’ve never had to reject one due to fitment issues!
This run included about 1,600 parts at 52% packing density — In the blue highlighted area there are 10 model variants total, each grouped with like variants, and printed in four layers per variant from the bottom of the chamber to the top:
For reference, here is Formlabs documentation on dimensional accuracy and print repeatability:
https://formlabs.com/white-papers/measuring-sls-dimensional-accuracy-and-print-repeatability
Here is a summary of my measurements (in mm), taking a random sample of 10 from each model variant.
Bottom of Chamber
Model1- nominal: 10.5 x 47.5 x 3.17 average across 10 samples: 10.47 x 47.62 x 3.14
Model2- nominal: 10.5 x 47.5 x 3.69 average across 10 samples: 10.46 x 47.62 x 3.7 (one 3.55 outlier)
Model3- nominal: 10.5 x 47.5 x 4.16 average across 10 samples: 10.49 x 47.6 x 4.17
Model4- nominal: 10.5 x 47.5 x 4.66 average across 10 samples: 10.47 x 47.59 x 4.67
Model5- nominal: 10.5 x 47.5 x 5.16 average across 10 samples: 10.47 x 47.62 x 5.11
Model6- nominal: 10.5 x 47.5 x 5.66 average across 10 samples: 10.47 x 47.57 x 5.65
Model7- nominal: 10.5 x 47.5 x 6.17 average across 10 samples: 10.49 x 47.56 x 6.18
Model8- nominal: 10.5 x 47.5 x 6.66 average across 10 samples: 10.47 x 47.56 x 6.63
Model9- nominal: 10.5 x 47.5 x 7.17 average across 10 samples: 10.48 x 47.55 x 7.18
Model10- nominal 10.5 x 47.5 x 7.66 average across 10 samples: 10.47 x 47.55 x 7.66
Top of Chamber
Although my testing was a bit crude compared to the Formlabs white paper, I was happy to see that my results were well within the quoted tolerance of ±0.5% or ±0.3 mm, and generally uniform throughout the build chamber. For reference, I’m utilizing the entire build volume. Formlabs mentions the Z axis having the most trouble but in my sample this does not seem to be a serious issue overall with these models (not sure what’s up with model 5 though!). It looks like I could improve my X/Y results by running a calibration? They are close enough that this has not mattered in practice.
I found these results super interesting! Many of you are seasoned pros, so this might not be groundbreaking, but for me, I’m still in awe of what this technology can do. I’m a genuinely satisfied user of the Formlabs ecosystem and thought others who are considering it might find this info useful.
If anyone here is lurking and has questions about running this equipment from a real-world user’s perspective, I’m happy to help where I can. I’ll also be posting about a few other perspectives on the ecosystem soon.
