Keywords

1 Introduction

Rapid development of 3D printing industry leverages the affordable 3D printer market dramatically. People can easily access 3D printing technology without advanced knowledge. Nevertheless, it is difficult to sustain high quality by using affordable 3d printers, because the quality of printed outputs is not just derived by machine specifications but also various situated factors such as printing mechanism, material properties, slicing algorithm, individual know-how, and so on. Early 3D printer users have conducted the evaluation for the quality of their own 3d printer by printing test models. That has been accumulated useful information about quality factors and models for tests on popular 3D printing communities such as Thingiverse.com. Nonetheless, most data of printing quality available is subjective and experiential information [4], because their purpose was to share problems and solutions from the failure, not to achieve objective evaluation. Comparable data of printing quality is essential for end users, HW/SW developers, and engineers in the field of 3d printing. As it enables users to predict the quality of a physical output, users can set an effective design guide or strategies to secure the best quality in advance. Controlling the quality of printed outputs in design process can greatly reduce the cost and time of post-build finishing processes [3]. The quality database would be helpful for a user-to-be or a consumer of the printing service to make a correct decision as well. In the platform such as 3dhubs.com, a local printing service network, printing quality data could make it more reliable.

In this paper, we examined the related works on popular 3D printing communities in order to categorize common factors more simply and found the limitations and improvements of such experiential evaluation. Then we suggested a series of the advanced test models of the categories and quantitative quality indexes that can be collectively executed and delivered in the community.

2 Case Study

We studied to verify the availability of existing quality evaluation projects of Thingiverse.com, 3DBechy.com, and Make: magazine. We observed terminologies, shapes, and evaluation methods in used for;

  • 378 models categorized “Popular-3d printing-3d printing tests” by Thingiverse.com

  • a test model of 3dbenchy.com

  • 7 test models of the article “What is Print Quality?” in Make: magazine, vol.42 (Table 1).

    Table 1. Features and limitations of the cases
    Full size table

3 The Proposed Category

First, we reorganized quality evaluation items according to the quantitative measurement methods in order to overcome the limitations of existing cases and make new quality evaluation standards. We divided them into 2 groups: Dimensional accuracy and Surface finish. In addition to that, we included build time issues as an important quality factor to be measured, because the time taken for 3d printing indirectly represents various invisible settings of slicers for improving the print quality (Table 2).

Table 2. Grouping evaluation items with quantitative measurement methods
Full size table

3.1 Dimensional Accuracy

Dimensional accuracy of the output can be critical for applications where fitting and assemblage are important or when parts have very small feature sizes [5]. Test models belonged to the dimensional accuracy group are measured by using the digital calipers or self-judging. This is a typical way of quantitative measurement used in existing cases, and it measures how accurate the size of the output corresponds to the size of the digital 3d model. Dimensional errors normally occurred because of mechanical issues, extrusion width parameter, material properties and so on.

  • Accuracy (Backlash): An accuracy quality factor could be general dimensional errors and the backlash phenomenon of 3d printers. In mechanical engineering, backlash is clearance or lost motion in a mechanism caused by gaps between the parts. It can be seen when the direction of movement is reversed and the slack or lost motion is taken up before the reversal of motion is complete [1].

  • LOD (Level of Detail): Capability of the 3d printer to describe more detailed objects accurately is also related to Dimensional Accuracy evaluation. LOD factors could be morphological issues. In this research, each one is focused on: Positive fine, Negative space, and Thickness (Table 3).

    Table 3. Shapes of LOD quality factors
    Full size table

  • Bridging: Bridging is where an otherwise unsupported gap must be crossed by a layer to form the desired structure. Variations in travel speed, extrusion quantity, and cooling will affect the bridging capabilities of the manufacturing processes.

3.2 Surface Finish

Surface finish is an important quality factor that can be recognized intuitively. Low-quality surfaces lead to a direct impact on post-processing costs, aesthetics and functionality of final outputs in production process. Surface roughness is often measured for evaluating surface finish quality. It is quantified by the deviations in the direction of the normal vector of a real surface from its ideal form. The overhang test in existing cases is relevant to the surface finish of 3d printing. The angle between surface to be print and a build plane is a critical factor to determine surface quality. Additionally, we categorized orthogonal surfaces as surface finish group in order to cover surface quality in general forms such as side, top, bottom, and supported surfaces (Table 4).

Table 4. Types of surface
Full size table

4 Design and Implementation

Based on the proposed category, we designed 7 different test models and collected 17 samples from 8 manufacturers during this research. In this paper, we would verify the possibility to compare quality of 3 samples printed with Ultimaker Original, Ultimaker2, and Flash Forge Creator.

4.1 Test Models with Quantitative Quality Index

According to the evaluation items, we have developed new test model standards as below:

  • Challenging features do not interfere with each other. (designing different models for each evaluation item, not an integrated model)

  • Test models have unified size and design.

  • Test models have appropriate shapes for measuring with a digital calipers and a digital microscope (Tables 5 and 6).

    Table 5. Test models and indexes of dimensional accuracy evaluation
    Full size table
    Table 6. Test models and indexes of surface finish
    Full size table

4.2 Comparison of the Printing Quality Index

Conducting the proposed quality evaluation process for 3 outputs from different conditions (printer-materials-slicer-manufacturer), we have verified the possibility of getting comparable data. In the evaluation results, outputs from Flash Forge Creator and Ultimaker Original have higher level of accuracy. Flash Forge Creator also produced better surface finishes. However, as we have established the quality indexes of negative value such as dimensional errors or surface roughness, lower index value means higher printing quality. In order for users to figure out the quality index more clearly, it is necessary to collect enough data to get reference value and to redefine the index of positive value in future research.

5 Conclusion

We insist the importance of the quantitative data about printing quality and its constructive value within user community. Analyzing related works conducted by Thingiverse.com, 3dbenchy.com, and Make:, We found that the user community had autonomously contributed empirical evaluation of the printing quality yet it was insufficient to propose the quantitative index of evaluation for the test models. We reorganized the evaluation items and suggested new test models to measure dimensional accuracy and surface finish. We anticipate these test models evolve as community-cultivated standard. For future study, we are currently planning a printing quality database service that end users and individual manufacturers of affordable 3D printers can feed and refer collectively.