Prototyping
Mechanical Construction
After the final design was modeled and the specifications were determined, the construction of the prototype began with the reception of the Advanced Cerametrics Piezoelectric Kit. The reason the group waited until the kit arrived to start prototyping was to determine the actual size and weight of the strips and whether or not our final design needed to be modified. The kit contained six PZT (lead zirconate titanate) strips where three were shimmed with steel and three were not shimmed. Also provided in the kit, there were three PMC strips, but they were not used since they were not needed for this type of application. The PZT strips that the group received in the mail were measured as 5.39’’ or 137 mm.
There was some initial debate between the members of the group about what material should actually be used in the production of this prototype to make it easier to build and more manufacturable for the machine shop. The cover was made of transparent acrylic for prototype purposes so that other people could see what was actually happening during the capture of the wasted vibrations. The mount for the piezoelectric strips was a made of ABS from a previous mount the group received. The original mount was modified to accept the same number of strips, but is smaller and more compact that the initial design.
There was conflict on whether or not the base should be made of acrylic or some type of metal. The group decided on making the base out of aluminum instead of steel for weight savings and discarded the idea of the acrylic based on durability issues we expected to encounter. The stainless steel brackets were purchased from Ace Hardware and modified for our application. The necessary mounting hardware (screws, bolts, nuts, etc) were also purchased from Ace while a majority of the electrical components were sourced from RadioShack and the Electrical Engineering Department.
The size of the unit was mainly dependent on two constraints, which are the overall length of the strips and the amount of room on the vehicle’s control arm the prototype would be mounted on. As stated before, the size of the PZT strips received from ACI were 137 mm in length, but the constraints on the control arm only allowed us a total length including mounts and wiring space to about 152.4 mm or a 6 inch length. From our technical analysis, a change in the size of the strips was advised to conform to our new constraint. Approximately 10mm was removed from each of the strips to a final length of 127mm or about 5 inches. Even though the strips fit within the size constraint, they were still cut because some slack needed to be present for all the wiring and the back end of the strips. Although the strips would all be the same size, the masses attached to them would vary accordingly to accommodate different frequencies that could be captured while they would vibrate at its natural frequency. The mount for the strips had an initial thickness of .75” and a height of 1.25”, but the initial length was approximately 6”. For our compact application, the mount was cut to a length of 3.5” so that it would be able to hold all the strips and fit within unit. The size of the base needed to be large enough to house the piezoelectric strips, the mount holding them, the PC board, and all the wires running to and from the unit. The group had the machine shop use .25” aluminum to make a piece that was 6” x 4.5” x .25”. The brackets, purchased from the hardware store, were 2.5” long and a thickness of .5” and were carefully fitted to the base to ensure an optimal fit. The design of the brackets was changed from the final design to make them easier to produce and source the parts.
Initially, there was a bracket leg that would slide in and out of a notch machined into the base, but since the machine shop did not have any material thick enough to accommodate this plan, the idea had to be reconsidered and the design changed. The new proposition included two L-brackets and two flat brackets connected by a long bolt and a nut to secure them together. There was some speculation that there would be too much excess movement, but once the unit was mounted, all the doubts were extinguished because the design created a tight, snug fitting with no movement at all. The size of the transparent, acrylic cover needed to be large enough to house all the necessary components: strips, mount, PC board, and wiring. The cover has the same basic width and length as the base itself and used acrylic that is .25” thick. The specifications of the cover allow for maximum deflection of the strips without any obstruction from the base or the cover and it is measured at 6” x 4.5” x 2.75”
Assembling the unit was straightforward after test fitting all the components to be certain they would all fit accordingly. After receiving the finished base from the machine shop, the mount was set on the base using a single screw and nut mechanism that is hidden in the middle. The strips were then placed onto the mount and screwed down using the available hardware. The weights placed on the strip were attached using magnets and these steel weights ranged from about 14.00g-39.01g. To allow the strips to vibrate at its natural frequency at the chosen length, the weight on one strip was 30.31g and on the other strip 19.22g and 19.56g were attached. The pieces of the case were glued together and attached to the base using two 1” long hinges that were cemented onto the case and the base.
Attaching the unit on the control arm required the use of the two L-brackets and the two flat brackets. Each L-bracket was bolted into their respective place on the base and the unit was positioned onto the control arm. Once the prototype was in position, four 4” long bolts were inserted and secured by tightening the flat bracket on the opposite side of the control arm using the appropriate nuts. After the bolts were tightened, the group tested the unit to see if any excess movement not supplied by the input of the suspension was present. The bracket design actually proved effective because there was no movement and the unit was secured in place. Another concern from the group, was to see if the weights would be held in place by the magnets or if they would move or fall from the vibrations of the vehicle. The magnets proved to be effective as the weights did not shift or move positions during the test of the prototype.
Electrical Construction
The construction of the electrical portion of the harvester unit was a 2 stage process. First, the circuit was built on a breadboard to ensure that the circuit worked and harvested energy as expected. The second stage of construction was the transfer of the board to a printed circuit board that would be installed on the final prototype.
Stage 1
To construct this initial test circuit, a breadboard, a 470μF capacitor, jumper cables, and a 100V rated diode rectifier were used. The figure below shows how the circuit layout. In order to ensure that the circuit was working, a voltmeter was used to measure the voltage of the piezoelectric input, across the capacitor and across the rectifier at its DC output nodes. The next step after the circuit was built was to ensure that the circuit was going to result in stored charge in the capacitor. The mounted piezoelectric strips were connected to the breadboard and into the AC inputs of the rectifier and a voltmeter connected across the capacitor. The strips were flicked and the change in electrical potential indicated that the circuit was operational
Stage 2
The construction of the prototype circuit shares some of the same components as the initial test circuit. However, a capacitor has been added to increase the charge capacity of the system as well as a secondary circuit that was to serve as the demonstration unit. The prototype circuit would have a toggle switch added that shifted the charged capacitor and connect it to an LED. The capacitor would then discharge through the LED which would light up. The circuit was built on a printed circuit board. This was in order to ensure that the circuit board took up as little space as possible in the final prototype. To ensure that the voltage across the capacitor was easily logged, leads were attached across the capacitor. This leads would be connected to a virtual instrumentation panel that would provide operation data for easy analysis. The secondary circuit was built on a perforated board with jumper cables connecting the capacitor and switch to the LED. This would serve as the visual demonstration that the capacitor had stored energy and that the circuit was working as expected. Below is also a picture of the final electrical circuit.
Preliminary measures were needed prior to the finalization of the design. Empirical data was necessary to supplement the theoretical data of the expected frequencies output from the control arm of the vehicle. This is because it is critical that the natural frequency of the PZT strips correspond to that of the vehicle’s suspension to obtain resonance. Resonance is the tendency of a system to oscillate at maximum amplitude at its natural frequency. By empirically determining the frequencies at the suspension arm of a vehicle, we are then able to tune the PZT strips to obtain resonance and therefore create the maximum amount of energy from said strips even if the driving vibrations are small and periodic.
In order to acquire this data, a National Instruments Data Acquisition Card was utilized in conjunction with an Analog Devices accelerometer. Labview 7.0 served as the intermediate software to translate the output from the accelerometer into recognizable data or further analysis.
The test rig consists of:
1. National Instruments Data Acquisition Card – Model USB-6009
2. Analog Devices Accelerometer – Model ADXL203EB
3. National Instruments Labview 7.0
4. Support Bracket
5. 22ga wire
It was necessary to use LabView in order to effectively acquire data. This took some time through trial and error, but a front end was created to monitor and log the data (dual axis acceleration) that was being output from the accelerometer.
The LabView virtual instrument utilized the input/output of the NI DAQ card to power the accelerometer and to record to disk the data output. Each VI was modified for the objective of frequency and amplitude testing. Vinod Challa was able to assist in the LabView setup and here is a screenshot of the inner workings.
Screenshot of the backend
Here is a screenshot of the front end.
A link to the final product specifications is provided.