Accelerated Melting via Viscous Heating

Abstract

The purpose of this project is to perform testing on a new additive manufacturing technology. Traditional fused filament fabrication (FFF) printers use heaters to melt the filament in the nozzle. Not only is this method inefficient, but the speed and force at which the filament is melted limits print speeds. It is theorized that heating the filament through viscous heating will not only be more efficient, but also produce faster melting. This report discusses the testing and results of preliminary experimentations involving the effectiveness of viscous heating to accelerate melting in additive manufacturing. It was found that not did the addition of viscous heating increase melt rate, it was the most significant factor in increase melting rates in our experiments. The addition of viscous heating to extrusion-based printers could lead to significant increases in print speeds in the future, leading to drastically shorter print times.

Introduction

FFF printing is widely available and used in the consumer space as well as the commercial space. Although popular, FFF printing has a major issue: limited print speed. The print speed of FFF is limited by two factors, the heater temperature, and the applied force to the filament. As heater temperature increases the printer requires more energy and becomes less efficient. As force is increased there is a critical force at which filament buckling occurs and causes printing errors. A driving system for the filament such as gears can also slip at high forces which again can cause printing errors. Viscous heating looks to sidestep both issues by introducing heat efficiently into the filament to promote more melting.

Basic testing done previously on this project yielded important results that are depicted in Figure 1 [1]. At low RPM where conduction dominates the heat produced in the nozzle, all filament diameters melt at similar speeds. As the RPM is increased, all diameters see increased material melting caused by viscous heating. Larger diameter filaments see a larger benefit than smaller diameters from using viscous heating. This basic testing shows the potential that viscous heating has on melt speeds, especially when used with larger diameter filaments.

              Figure 1: Material melt vs RPM[1]

The data being collected from this project is preliminary. It will help to aid in the design and use of a prototype viscous heating FFF printer. Further experiments will be run with said prototype once it is fully constructed.

Experimental Design

This experiment will be performed on a simple test setup pictured below in Figure 2 [1]. The components include a 3-level metal frame, a heated bed, filament guide block, CNC router motor, and the filament itself.

Figure 2: Computer model of experimental setup [1]

Not only will the experiment be performed to understand the effectiveness of viscous heating on filament melt speeds, but it will also provide insight as to how other parameters affect melt speeds. The parameters varied will be temperature, filament diameter, force, and the rotations per minute (RPM’s) of the filament. These parameters and the values at which testing will be conducted are shown in Table 1 below.

Table 1: Experimental parameters

Testing Procedure

This testing is preliminary, and as such a simple setup is used to conduct the experiments while a prototype is being built. To perform a test, first 10[cm] long PA samples are cut and weighed as shown in Figure 3.

Figure 3: Weighing the test samples

Next, the router motor must be calibrated to the correct speed. This is done by altering the voltage of the motor, while the RPMs are measured using a tachometer as pictured below in Figure 3.

Figure 4: Measuring Router Motor RPM’s

After the speed of the motor has been set, the heated bed can be set to the desired temperature and the guide block is moved into place. The sample is then attached to the router motor and lowered into the guide block. The motor is turned on and the sample is lowered until it reaches the heated bed, this begins the test. The test duration is 1 min, during which the RPM’s are monitored to ensure that they are within an acceptable range of the desired value as shown in Figure 5 below.

Figure 5: Monitoring RPM’s during testing

After 1 minute the sample is removed from the test apparatus and is weighed to determine the amount of material that was melted throughout the experiment. The heated bed is then cleaned of any melted material before beginning the next test.

Results

After the testing, a data analysis was performed, and the following graphs were produced. Figure 6 shows the relationship between the independent variables of RPM, force, and temperature for a diameter of 3.175[mm]. This graph shows that RPM has a significant impact on the material melted. It also shows that as the independent variables are increased individually, they all increase the material melted.

Figure 6: Diameter = 3.175[mm] [2]

Figure 7 below shows the same relationship as Figure 6 and with the same independent variable but for a diameter of 4.625[mm]. The trends in Figure 7 are similar and match the trends found in Figure 6. Once again as RPM, force and temperature are increased the material melt is also increased.

Figure 7: Diameter = 4.625[mm] [2]

Figure 8 below again shows the same relationship as Figure 6 and Figure 7 but for a diameter of 6.35[mm]. The trends in Figure 8 are similar to the trends found in both Figure 6 and Figure 7. As RPM, force and temperature are increased so is the amount of material melted.

Figure 8: Diameter = 6.35[mm] [2]

To better understand the effect of each parameter being varied in the testing, a statistical analysis was performed on the data. In the pareto chart below, the magnitude of the standardized effects displays which testing factors influence the melt rate more. The chart also shows which combinations of parameters carry the most significance. In this analysis, any value greater than the red line at a value of 2.01 implies the parameter is statistically significant.

Figure 9: Pareto Chart [2]

From Figure 9 above, the most influential input is found to be the RPM of the test. This is a promising result, as it shows that the addition of viscous heating in our experiments not only has a significant effect on the melt weight of the test, but it is the most significant factor influencing the melt weight of the test.

To further explain why this is an important result, Figure 10 below displays how equal melting rates may be achieved at varying temperatures through the addition of viscous heating.

Figure 10: Correlations between temperature and RPMs [2]

Following the horizontal red line in Figure 10 the melting rate of 225 mg/min achieved by a test at 295 C and 0 RPM can be equivalently reached at a temperature of 280 C with 1550 RPM. Similarly, a rate of 225 mg/min can also be achieved at 265 C with 3000 RPM. The addition of viscous heating to 3D printers should allow for much greater melting speeds, leading to shorter print times. At the same time viscous heating could be introduced to reduce the necessary temperature and force for a specified melt rate.

To further predict how to achieve a specific melt rate based upon a set of parameters, contour plots can be used as a prediction tool. After the statistical analysis, the following contour plots were made to predict the melt weight for different combinations of independent variables. Figures 11-13 below show the melt weight for a combination of diameter and temperature at a force of 36.62[N] over several RPMs. The dark blue contour is the lowest melt while the dark green contour is the greatest melt.

Figure 11: Melt Weight Contours, 0 RPM [2]

Figure 12: Melt Weight Contours, 5000 RPM [2]

Figure 13: Melt Weight Contours, 10,000 RPM [2]

It is important to note that there were not enough tests run to create data for 10,000 RPM but the contour plot, Figure 13 can be used to predict the material melt for 10,000 RPM tests. It can also be seen by comparing the figures that as the RPMs are increased, significantly more melting occurs, as noted by the increase in green contour areas.

Conclusions and Future Work

The preliminary experimental results provide support for the use of viscous heating in extrusion-based 3D printing processes. It was shown that the melt rates achieved at high temperatures can also be obtained at lower temperatures through the addition of viscous heating. As shown, the standardized effect of the RPMs on the material melt rate was the single highest impact value when compared to other parameters such as temperature and filament diameter. As more sophisticated prototype printers are developed to make better use of viscous heating, significant increases in print speeds could occur. In an ideal case, these print speeds could be 2-10x traditional FFF speeds which may make FFF processes more applicable to mass production environments like in the commercial enterprises.

The results of the viscous heating testing are promising, but the testing was performed with a rudimentary experimental test setup. To address this, a more sophisticated prototype model is currently being developed. Upon completion, additional testing can be performed to validate the preliminary testing results as well as test values which were unobtainable with the initial testing set-up. Specifically, RPM values of 15,000 – 20,000+ can be tested to examine their effectiveness. Rather than rotate the filament to induce viscous heating as was done in this project, the prototype will feature a rotating nozzle capable of reaching high RPM values. The cross section of the prototype model is shown in Figure 14 below.

Figure 14: Cross-section of prototype [2]

Lastly, comparing the experimental data to a theoretical model would be ideal to not only understand how viscous heating affects melt rates, but also to predict the melt rates of untested parameter combinations.

Acknowledgments

We would like to thank Allen Román, our advisor for this project, for both his help in the lab and for sharing his analysis that resulted from the testing.

Bibliography

[1]  T. Osswald, A. Roman, ‘Accelerating FFF Printing via Viscous Heating’, University of Wisconsin-Madison, 2022. Available: Accelerating FFF Printing via Viscous Heating

[2]  T. Osswald, A. Roman, ‘Accelerating FFF Printing via Viscous Heating – Experimental Findings and Predictive Algorithms’, University of Wisconsin-Madison, 2022. Available: FFF Printing via Viscous Heating – Experimental Findings and Predictive Algorithms

[3]  J. Quintana, “Understanding Melting in Fused Filament Fabrication Process” Ph.D. dissertation, Dept. Mechanical Eng. , Univ., Madison., WI, US, 2021. Available: Understanding Melting in Fused Filament Fabrication Process