FUNCTIONAL REQUIREMENTS
Maintain oxygen level near printing area below 1%, or 10,000 PPM
Oxygen sensor implemented to accurately represent oxygen levels present in the fluidics loop with an accuracy of ±0.1%
Avoid powder disturbance in the print bed by introducing argon at velocities below 2 m/s
Gantry that moves in X,Y, and Z axis directions
FINAL DESIGN
The final design of the XLence AM 3D Inertion Printer consisted of a three-directional gantry system, a fluidics loop driven by a GDSTIME DC blower fan, a ZrO₂ Zirconia oxygen sensor and a custom nozzle that delivers fresh argon to the printing area of the metal printer. The gantry components are driven by an Arduino Mega 2560 using a Ramps board and Marlin position control software and the fluidics loop is controlled by a separate Arduino Mega 2560. The final design and placement of the components can be seen in the figure above.
GANTRY SYSTEM
The Gantry System was created using 3D printed parts, electrical conduit railings, timing belts, and five stepper motors. The design was adopted and created using the guidelines of the MPCNC, an open-source project from V1 Engineering.
The system is capable of motion in the X, Y, and Z directions and is able to print in a 0.0283 m³ (1 ft³) print bed meeting the functional requirements.
FLUIDICS LOOP
In the Fluidics Loop, the main purpose was to extract argon fumes from the printing area, filter unwanted particulates, and recirculate the remaining argon back into the inlet of the nozzle. This process was driven using a GDSTIME DC brushless blower, and a ZrO₂ Zirconia oxygen sensor placed in the fume extraction path which collected data regarding oxygen levels in the print bed.
The fluidics loop was able to maintain oxygen concentration near the print area below 1% or 10,000 PPM, and the O2 sensor was able to read oxygen levels with an accuracy of ±0.1%
NOZZLE
The purpose of the nozzle was to act as an aid in the argon recirculation cycle and to deliver fresh argon onto the printing area. The fresh argon was supplied using an argon tank which was led to the inlet of the manifold, into the nozzle, and then directly onto the printing area in a halo shape. The flow of argon through the nozzle was able to be delivered at a rate lower than 2 m/s so as to not disturbed thin layers of metal powder used in metal printing.
Ultimately, the final design was capable of aiding the argon recirculation cycle and deliver fresh argon to the printing area keeping the oxygen concentration below 1%.
BLOWER
The blower was essential in the design of the fluidics loop as it drove the extraction of argon fumes, aided the argon recirculation cycle, and delivered fresh argon to the inlet of the nozzle leading to the printing bed. The blower used was a GDSTIME DC blower that was controlled using an Arduino Mega 2560. The blower consisted of four pins, power, ground, PWM, and tachometer, with the tachometer being responsible for measuring the RPM in bits.
KEY COMPONENTS
MANIFOLD
The purpose of the manifold was to intake argon at its inlet located at the top of the part, separate it into ten-outlet tubes, and then deliver the argon to the tubes of the nozzle which lead to the halo diverging ducts. The manifold was capable of delivering a uniform distribution of argon creating a barrier from ambient oxygen in the atmosphere.
OXYGEN SENSOR
The oxygen sensor was a key component in the project as the ultimate goal was to maintain an oxygen concentration of less than 1% on the print bed. The oxygen sensor was mounted in the fume extractor and was able to accurately read oxygen concentration with an accuracy of ±0.1% .
RESULTS
OBJECTIVES
Establish that the oxygen concentration could be lowered by inerting the printing environment with argon
Characterize the oxygen concentration response to various blower fan speeds
Determine the optimal blower speed and input argon tank flow rate configuration to achieve 1% oxygen concentration while minimizing argon use
VARIED ARGON INPUTS AND BLOWER POWER
The figure on the left shows argon outputs from the tank at 5, 10, and 15 CFH while the blower's PWM input varied. The target of the oxygen concentration was 1% which shows that as the CFH is increased, so does the need for more power from the blower in order to reach the desired 1% steady-state oxygen concentration.
The figure on the left is a sample test of where the blower's power was kept at constant at 15.6% power throughout the test. For the first 60 seconds, the argon tank flow rate was kept at 20 CFH, however, after the 60-second mark, the flow rate was decreased to 2 CFH showing that the oxygen concentration was still kept at the same steady-state below 1%. It wasn't until the supply of argon to the system was turned off that the oxygen concentration level spiked. This shows that the blower running at 15.6% power is able to maintain the concentration of argon below 1% regardless of a 20 CFH, or lower input flow rate.
GANTRY STATIONARY/MOTION
Gantry Stationary
Gantry in Motion
One of the most compelling questions for this project was whether the gantry in motion against it being stationary would result in different oxygen concentrations in the print area, so tests were run. The graphs above show that the oxygen concentration was ultimately the same regardless of the gantry being in motion or not. This further supports that the design of the fluidics loop and nozzle assemble were well designed.
OPTIMAL CONFIGURATIONS
This surface plot is a three-dimensional plot that shows the oxygen concentration in relation to the argon flow rate (CFH) and the PWM input % which in this case is the blower power. Shades that are green in the plot are CFH and PWM input % combinations that will result in the desired oxygen concentration of 1%. From the plot, it is evident that as the PWM input % increases, so does the concentration of oxygen on the print bed.
The surface plot on the left is a two dimensional plot that shows the configurations of PWM inputs and argon tank flow rates that can achieve the desired 1% oxygen concentration on the print bed or lower. Several tests were run with such said configurations and the data shows that lower PWM inputs from either 5 CFH up to 20 CFH can achieve the goal.