- Purpose of fixture
PARC collaborated with a leading manufacturer to develop a new process for printing pastes. The new process extruded metallic gridlines from a printhead, and when compared to screen-printing, resulted in improved performance. I designed a fixture to examine the micro-fluidic behavior inside the printhead during the process of extruding the gridlines.
The printhead had two inlet materials, A and B, which came together before leaving the printhead to form the metallic gridlines. A high-speed camera fastened to 2 one-axis stages was pointed at the printhead nozzle where materials A and B came together. The printhead was constructed of several optically-transparent layers, all clamped with bolts and two metal plates. To achieve desired material pressures inside the printhead, I designed dispensers with pneumatic actuators to amplify their input air pressures by about 6X. More information about this is in the Equipment Modules section.
Figures 1 and 2 below show a SolidWorks CAD screenshot and photo of the fixture, respectively.
- Equipment Modules: Dispenser Air Controller
To regulate the input air pressures to the dispensers, I designed a dispenser air controller (Figure 3 below) to have facilities compressed air as an input, and two regulated air pressure outputs for the material dispensers. For each dispenser, the controller housed a manually-adjustable pressure regulator with digital readout gauge, a solenoid valve, and plumbing (i.e. plastic tubing and fittings).
- Equipment Modules: Material Dispensers
Each material dispenser consisted of an off-the-shelf pneumatic cylinder (double acting, single end rod), a material "fuel gauge" sensor (see section below), and custom machined metal parts to push the material from a cartridge. The custom machined metal parts pushed on a machined plastic plunger inside the cartridge at a pressure equal to 6X the inlet air pressure. The cartridge assembly consisted of an off-the-shelf plastic cartridge, a machined plastic plunger to push the material out of the cartridge, and o-ring seals for the plunger. A PTFE smooth-bore hose was connected from the material cartridge outlet to the printhead.
See Figure 4 below to view the material dispensers.
Because the pneumatic cylinder was double acting, full retraction of the piston was only possible if its retraction inlet was supplied with air. I did not want full retraction of the piston because a cycle of full retraction and then full extension of the piston would involve repeated high force contact with the plastic cartridge's plunger. When a dispenser's solenoid valve turned off, the piston retracted minimally due to gravity.
- Equipment Modules: Fuel Gauge Sensors
As more material left the cartridge during an experiment, both the cartridge's plunger and the cylinder's piston moved closer to the cartridge outlet. The fuel gauge sensor detected the position of the cylinder's piston. I calibrated each fuel gauge sensor's 0-10 DC voltage output to be proportional to a percentage of material remaining in the cartridge. Before each experiment, the operator could view the material percentages remaining in the cartridges to determine how many cycles to run. I also wrote software that ended an experiment when either material percentage reached 2%, similar to a warning for a car's low fuel warning.
See Figure 5 below to view a fuel gauge sensor.
- Equipment Modules: Pressure Sensors
In the pair of PTFE smooth-bore hoses that connect each material dispenser cartridge to the printhead, there was an inline pressure sensor. The pressure data was measured to quantitatively characterize the fluid behavior from cycle to cycle. Each sensor was mounted in a machined block with welded flare fittings; the sensors are circled in Figure 6 below.
Each sensor required +5 V DC to operate, which came from an amplifier dedicated to each sensor; the amplifier required between +11 and +28 V DC Each sensor's positive and negative output signals went to its corresponding amplifier. These output signals went to both a DAQ module and the PIC.
The DAQ module's software was used for measuring the pressure response over an entire cycle to better understand the rise and fall of the pressure.
For the experiment, the operator could set the time each cycle when the PIC read the pressures; this data was recorded once per 100 microseconds, and then averaged for 10 readings. After averaging both materials' pressure data, the PIC printed these averaged pressures on the host computer's terminal emulator window. The terminal emulator program, Tera Term, logged the data for each experiment in a text file.
- Equipment Modules: High-Speed Camera
I implemented a high-speed camera and optics on a vertical optical post with 2 one-axis stages for viewing the outlet nozzle of the printhead. The high-speed camera was triggered on with a +5V DC TTL input signal from the PIC.
In the camera's software, there were many settings to configure for the operator. In particular, the most important settings were the trigger mode (set to "Random" mode), frame rate, bit shift (i.e. the 8 bits that were captured out of the 10 possible monochrome bits), and number of frames to capture at each trigger input signal. To be able to view particle motion well in the nozzle, the camera was set to record at 2000 frames per second. When the camera's memory buffer filled up, the video file was automatically downloaded to the host computer where it was later compressed.
See Figure 7 below for a photo of the camera and optics.
A 16-bit PIC micro-controller was used to do the following:
- turn on MOSFETS to send on/off commands to the solenoid valves
- turn on a MOSFET to send a +5V DC TTL signal to the high-speed camera for triggering the recording of a particular number of frames selected in the camera's software
- read analog fuel sensor data
- read pressure sensor data
- communicate with the host computer via RS232
Figure 8 below shows the electronics panel; see Figure 3 to view the PIC inside the plastic container.
The electronics panel in Figure 8 includes (clockwise from left to right): the pressure sensor amplifiers, +24 V DC power supply, voltage divider circuits (inputs: 0-10 V DC; outputs: 0-5 V DC) on breadboard for fuel gauge sensors, DIN rail terminal blocks, and DAQ module. The PIC handled +/- 5 V DC, and the fuel gauge sensors outputted 0-10 V DC, so that's why the voltage divider circuits were needed.
I used a Microchip IDE to develop a software package for the PIC, and then purchased a Microchip MPLAB ICD3 module for debugging and uploading code onto the PIC in conjunction with an off-the-shelf C compiler. I used Tera Term, a software terminal emulator, to display communication on the host computer.
The operator needed to only modify a header file in the software package to vary experimental parameters. These parameters included:
- when to turn on the solenoid valves for Materials A and B each cycle
- when to turn off the solenoid valves for Materials A and B each cycle
- when to trigger the high-speed camera each cycle
- when to acquire pressure data each cycle
- how long to wait between each cycle
I taught two interns how to operate the equipment for automating equipments. The fixture reliably let an operator begin running an experiment in as little as a few minutes. Many experiments were run over the course of 5 months.