Evaluating the Accuracy of a 3D Printed Part Against the CAD Model

Published on August 20, 2018. Written by: GoMeasure3D Team


This article shows how to gauge the performance of a 3D printer by evaluating its 3D printed part.

3D scanning technology empowers us to capture accurate surface measurements of a given part to perform the analysis. The scanned part can then be checked for measurement deviations from the original CAD file with computer-aided inspection software.

No 3D printer can print to the perfect specifications of a CAD model, the design blueprint of the part. In reality, every physically manufactured part, including 3D printed ones, will have slight variations from the CAD model.

Variations to the CAD model are also affected by the type of 3D filaments you use. You will have to accept some level of deviation and determine what type of tolerance you are comfortable with.

Steps for Evaluation

  1. Print desired part from CAD file on a 3D printer
  2. Scan printed part using a 3D scanner to acquire point cloud data about its surface geometry
  3. Use a deviation tool such as an inspection software to directly compare the scan file and the CAD file

3D Printing

Axiom 3d printed part

A filament tester plate was used as the demonstration part for this project. It is a 2 inch tall hexagon with various features that are challenging for 3D printers. If the printer fails in one area of the test part it should be visibly obvious and can quickly demonstrate what needs to be improved the next time. It is important to be able to accurately analyze the dimensions of this part in order to understand the printer behavior.

The material used for a filament is a woodfill PLA composite from ColorFabb. It has 20% wood shavings mixed into the plastic that provide an aesthetically pleasing look that can be stained and finished like regular wood. We chose this filament because it’s more challenging to work with. This filament prints like regular PLA with a little extra fine tuning.

Axiom Dual 3d printer
Airwolf Axiom Dual 3D printer

The printer used in this project was an Airwolf Axiom Dual with a bowden tube drive. Airwolf provides a flavor of Cura 3D printing software that has tailored preferences for each filament type. In our lab, we use Apex instead of any other slicing software in order to maintain consistency with the printer brand. Once the part was configured in Apex, the 3D printer took about a half an hour to print the part.

3D Scanning

3d scanning 3d printed part

Once the part is printed, we scanned the part in order to capture its surface measurements for evaluation. The part was scanned using the HDI Advance 3D scanning system. This setup captured accurate information about the printed part in very high detail. Using the 35mm lenses, we were able to get 14 microns of accuracy across the scanning volume.

Flexscan3D 3D scanning software that comes with the 3D scanner collected the points of information and constructed a physical representation of the 3D printed part. The file was exported to .stl file format to be used in the inspection software for comparison.

Comparing 3D Printed Part to CAD File

3d printed part in cad and scan model

Geomagic Control X 3D inspection software allows for deep analysis of physical objects when scanned into 3D digital form. In this project, we imported both the original CAD file as well as the .stl scan file into the software in order to compare the two. The left photo is the original CAD file (the design blueprint of the part) from SOLIDWORKS and the image on the right is the scan file we acquired from the 3D scanner (scanned 3D printed part).

3D deviation map

Once aligned in the same orientation, these 3D objects can be directly compared to each other for discrepancy in terms of surface area and volume. A 3D comparison tool in Geomagic Control X gives a color map of the measurement differences between the two. Red and orange colors indicate where the printer extruded too much plastic and anywhere that’s blue indicates too little.

In general it seems the part may be a bit over extruded. While at the same time, side-to-side there was an average shrinkage of 0.04%. In the future we could potentially turn down the extrusion rate to perhaps 95% as well as scaling up the print size by 0.05% in order to get a more dimensionally accurate part.

Running the Test Again

Updated 3D deviation map

The same tester plate was reprinted with a scaling factor to compensate for the shrinkage observed and a reduced extrusion rate to see if this analysis provided positive change in print quality. The scanning and evaluating process outlined above was repeated.

The results show most of the heat map is green, which indicates that the part is very close to the actual measurements in the original CAD file. Edge to edge the shrinkage was almost totally eliminated.

Takeaway from the Test Results

The results are promising when it comes to tailoring printer settings to the 3D printing filament you are using. This process can be repeated with any 3D printed shape.

Overall, the amount of shrinkage observed for woodfill PLA from ColorFabb has a 0.04% decrease in volume size. We suggest that if you want to print with this material, try scaling the part by 1.004 for more accurate and precise prints for the best print quality.