Multi-Metal Bedding Technology (Senior Design Project)

“3D-Printing” metals have always been a challenge in the additive manufacturing industry; traditional techniques used such as Selective Laser Sintering (SLS) inhibits the use of more than one metal powder during every powder bed recoating process without cross-contamination. Asides from that, jetting means have been implemented such as material jetting/binder jetting where metal particles are mixed with a for of photocurable polymer fluid/ binder that helps a multi-material model to be 3D printed in its desired shape; the drawback using this method is that lengthy post-processing is required such as inserting the final part into a furnace to fuse the metal particles and incinerate the polymer particles that make up the composite. This causes dramatic shrinkage and a final part that has a low metal density (80%). The proposed solution and alternatives to these methods is to create a metal powderdispensing mechanism/nozzle that would accurately and selectively deposit two different metal powders in a preconfigured pattern on a single voxel (printed layer), while preserving high metal density that is achieved through the SLS process. Further work would entail the implementation of a CO2 highintensity laser to fuse the metal powders on each layer before another layer is recoated using the proposed selective metal-bedding technology.

In our project, 80/20 T-slot aluminum bars were used to provide a solid foundation for the printer and for each attachments of fixtures for the X and Y linear stage mechanisms that were frequently reiterated and had to be easily removed. The final design is shown in figure 1 below:

Figure 1: Final Assembly of the Multi-Metal Selective Bedding Apparatus

As shown above, all linear rod fixtures, bearing holders, and Y-axis gantry were additively manufactured using a Fused Deposition Modelling (FDM) 3D printer for the ability to rapid prototype with low cost. This introduced several accuracy constraints in the final powder-printed pattern in both powder line width resolution, and layer height accuracy. The large tolerances introduced from manufacturing these components (as compared to using subtractive manufacturing means) created room for vibrations in the carriage to occur, creating small amounts of unwanted depositions in certain areas of the printed pattern.

Figure 2: FDM-Printed Roller and Nozzle Assembly (0.4, 0.3, and 0.2mm print resolution)

The next constraint also came from the high tolerance of the roller to the powder opening slot in the powder funnel as shown above in figure 2. Despite the resolution of the powder spot size increasing when decreasing the nozzle diameter on the FDM 3D printer, poor diametrical accuracy on the 3D printed nozzle outer and inner diameter was still observed which caused unwanted powder deposition at certain regions of the printed pattern as seen in figure 3 below.

Figure 3: 3D Printed Pattern using FDM-Printed Roller, Funnel, and Gears

As a result, the team resorted to a more accurate means of manufacturing our components – Stereolithography Apparatus (SLA) as shown in figure 4.

Figure 4: SLA 3D Printed Bevel Gears

Figure 5: SLA 3D Printed Powder-Depositing Rollers

This manufacturing method alternative has a high print resolution of 100 micrometers in the X and Y direction as opposed to 0.4 millimeters on the FDM printer, and 10 micrometers in the Z direction as opposed to 0.2 millimeters respectively. This significantly improved metal deposition accuracy and eliminated the majority of unwanted depostition (as shown in Figure 6).

Figure 6: Improved Powder Deposition Resolution using SLA Printed Components

Lastly, the restriction of space in the X and Y axis linear range of motion made it challenging to fit two different nozzles/dispensers on the Y carriage without the risk of collision, unnecessary distance of the center of gravity each extruder from the center of axis of the linear rails which would significantly increase vibrations. Hence, the stepper motors, powder-dispensing nozzles, and translational gears had to be repositioned and optimized.

Creative Synthesis of Multiple Design Solutions

Firstly, we explored various drive mechanisms to be able to translate the rotation from the stepper motor axis to the powder roller axis.

Figure 7: Gear and Pinion Powder Deposition Mechanism

Initially, a step-up gear and pinion was used to increase the speed of the roller with relatively low stepper motor speed to reduce chances of skipping steps that come from high step speeds (design shown in Figure 7). This method proved to be ineffective due to the inefficiency in space for the larger gear with the positioning of the stepper motor, as well as the poor control in the alignment of the gear and pinion that propagated into the roller which caused a “pulsing” behavior in the powder deposition.

As a result, we picked a desired location for the stepper motor closest to the center of the Y-axis and selected a gear that would allow for the configuration. The final drive iteration was selected to be a bevel gear configuration that would allow for minimal space usage and low drive deviation (shown in figure 8).

Figure 8: Bevel Gear Drive Mechanism Setup

Furthermore, powder-dispensing mechanisms were also mocked and tested. The first conceptual iteration was an auger-style design as shown in figure 9 below.

Figure 10: Auger-Style Powder Dispensing Mechanism

Through testing, we discovered that this design had poor ability to stop flow due to gravity and vibrations during non-printing movements (such as homing) without the use of a mechanical valve. Additionally, the implementation of a driving mechanism was also infeasible without the contamination of metal powder in the gearbox.

The design progressed into a roller-style mechanism where a teethed roller would be used to excavate powder through an opening slot in a funnel. The first iteration of a 4-roller design (as shown in Figure 11) was created and tested.

Figure 11: Four-Teeth Roller and Funnel Design

The results were seen to improve with minimal unwanted powder deposition. However, one of the drawbacks was that it had a choppy federate and would only deposit powder every quarter turn of the roller. We later solved this by increasing the number of teeth on the roller to a 20 teeth configuration as shown in figure 12.

Figure 12: Twenty-Teeth Roller Configuration

Evaluation and Analysis of Proposed Solutions Leading to a High-Quality Design

The link to the video of the apparatus operating at 5mm nozzle width and at distance from bed of 2.23mm: https://www.youtube.com/watch?v=eno-bxTHY9A

During the testing stage, the main challenge weighed between the aperture (diameter) of the powder dispenser, powder feed rate, and height of nozzle from the print bed. The results of the experimental runs were conducted, and results were tabulated and plotted as shown in Plot 1 and 2 below.

Plot 1: Correlation of Nozzle Size and Printed Line Width

As shown in Table 1, a decrease in nozzle width would increase the print resolution, with the actual printed line being less than half of the selected nozzle diameter. A contrast between a 7mm diameter nozzle and a 5mm nozzle is shown in Figure 13 and Figure 14.

                   Figure 13: 7 mm Aperture                                     Figure 14: 5 mm Aperture

However, a nozzle diameter of 2 mm was tested at the same feed rate and was seen to induce powder build-up at the convergence end of the nozzle. Further test would entail optimizing this process by reducing the feed rate and decreasing the traveling speed of the extruders.

Plot 2: Nozzle Height from Bed vs Part Dimensional Accuracy

From Plot 2, a minimal nozzle height from bed produced the square geometry with the highest accuracy.

In summary, we achieved a X and Y axis line width of 1.92mm ± 0.25mm which is a deviation of 0.48mm in width when compared to the line width planned in the firmware. Furthermore, a layer height of 0.25mm ± 0.15mm, which is shy of the benchmark layer height requirement for Selective Laser Sintering (SLS) 3D Printing (0.060mm to 0.150mm).

Future work would entail the reduction of vibrations by using subtractive manufacturing to provide a more stable device that would not interfere with test results. Moreover, stepper motors with encoders would be used to ensure no steps were missed during rapid acceleration/deceleration of the gantry to maximize accuracy of final geometry. Lastly, more iterations of test will be conducted to minimize nozzle diameter and optimize the feed rate and travel speed to accommodate for a smaller opening.