Kate Gleason College of Engineering

Multidisciplinary Senior Design

Kate Gleason College of Engineering

Rochester Institute of Technology

Rochester, New York 14623

Multidisciplinary Senior Design Team P16203

Team Members:

Andre Pelletreau – Electrical Engineer

Radoslaw (Kai) Maslanka – Electrical Engineer

Vincent Stowbunenko – Electrical Engineer

Jeremy Willman – Mechanical Engineer

Kerry Oliveira – Electrical Engineer

Abstract

The Smart Power Supply Test Bench is an automated test fixture used to reduce the time spent testing and the total cost of quality assurance on a custom power supply and control board assembly. The unit provides consistent power to an oil drill over thousands of feet.

The current testing procedure takes about an hour to complete. With many units being assembled and tested per year at Bear Power Supplies, there was a necessity to reduce the total time testing. The Smart Power Supply Test Fixture reduces the testing time to under ten minutes. It calibrates and tests the functionality of the controller board and power supply board serially and in parallel. Original constant current and constant voltage load designs, an Arduino micro-controller, and relays switching between active loads for testing aid in the automation. The production of this fixture will increase the speed, efficiency, and safety of testing.

Introduction

Bear Power Supplies in Phelps, NY designed a custom power supply for an oil drilling company. The power supply itself stays above the surface providing consistent power for the drill head over thousands of feet, with the distance varying depending on the location of drilling. The supply adjusts for drilling force when encountering various earth types. It also receives all AC input voltage, making it a universal product. The various design requirements created a tedious and laborious testing process. With the testing rate being 60 min/unit, monetary cost increases with each unit tested creating a necessity to lessen the time spent testing, including an easier testing procedure.

Considering the capabilities of the power supply, the customer wanted a semi-automated system utilizing a microcontroller to run each test for the unit, monitoring the AC voltage, and providing necessary feedback to the operator through the test. The fall semester of the 2015-2016 academic year was utilized to design three Printed Circuit Boards (PCBs), while the product was assembled, tested, and displayed at Imagine RIT during the spring semester.

Design Process

The Smart Power Supply Test Bench is required to be semi-automated, to protect the operator from high voltage present during testing and any other possible safety issues, easy to use, and ergonomically functional. The total desired time for testing is ≤ 5 minutes/test with a total maximum time limit of 15 minutes, which includes setup and connection of cables, testing, and disconnecting. Within the 15 minutes, the control board and power supply must be tested separately (serially), then together as a unit (parallel), hence 5 min/test. Assembly of the unit is not an included step in the testing process therefore not included in the total time.

To be semi-automated, the test bench should automatically calibrate the control board and test the power supply board at different load voltages, currents, and resistances. Feedback should be provided to the operator, displaying the test occurring and the step of the procedure also indicating if the unit is on. The feedback system should also display if the unit passed/failed, step in procedure where failure occurred, and reason for failure. AC voltage and heat dissipation should be monitored and controlled to protect the electrical equipment. A barcode scanner and an SD are needed to document and log the serial number of the unit being tested and the testing data. The recorded data needs to include passing/failure of a unit, reasons for failure if it occurs, results from each test, and date/time of each test. The UUT needs to be visible to the operator at all times and the enclosure must protect the operator from sudden, dangerous failures and from AC voltage. Because an operator will be testing power supply units throughout the day, the test setup should be quick and simple, and the testing procedure should not cause muscle strain. In the event the software fails to shut down the enclosure upon an emergency, an emergency stop button should be placed on top of the enclosure to shut down all operation. The system should shut down as well when the enclosure is opened. The enclosure housing all materials must be of dimensions 1.5’ x 1.5’ x 1.0’ (L x W x H). Lastly, the budget for the project was $1,500. The power supply unit used for testing was provided by the customer Bear Power Supplies.

Various techniques were approached to narrow down the design choices during MSD I including morphological charts, Pugh charts, and benchmarking. The morphological chart was the least useful process for us as it didn’t offer help with choosing a design path. The Pugh chart shown in Figure 1 offered some help with design choices.

Figure 1: Pugh Chart Concept Selection

Although the Pugh chart was meant to limit design selections, a hybrid of concepts 2, 3, and 4 were chosen because they all offered a better options for a few of the requirements. The hybrid concept selected was a scatter shield and a base enclosure to protect the operator from the UUT, heat sinks to cool the electronics along with fans, a switch to initiate the testing procedure, software to control the test boards as well as software calibration, local storage of the test data on an SD card, and an LCD screen along with LED indicators to provide feedback to the operator.

Microcontroller

Figure 2 shows the bench marking to determine the best microcontroller to use for the design.

Figure 2: Microcontroller Benchmarking

The Arduino Mega was chosen because of the RAM and storage size, the GPIO capabilities, as well as the nature of the programming language being more universal and easily portable. There are 54 Digital I/O pins, 15 provide Pulse Width Modulation (PWM) outputs and 16 provide analog inputs. This allowed for the ease of allocation of the pins to various software functions needed to enable control signals, to take real time measurements and to communicate with the various serial ports on the controller board. The choice of this microcontroller also coincides with the customers’ typical design choice for microcontrollers.

The microcontroller helps make the system semi-automated which was a customer requirement. It controls the system entirely by sending signals to the load board dictating the proper testing voltage and current, and directs the relay on switching to the proper Autotransformer. The Arduino lowers the total time for testing and also controls the data to store during a test which is required for verification of the proper functionality for the UUT. The data is stored on an external SD card. Additionally, the open source library from Arduino allowed for the team to quickly learn how to write software to provide communicate with the various pieces of hardware, like the SD card reader or the RS-232 serial communication protocol used to send various enable signals to the UUT power supply.

Software

The software was developed by the four electrical engineers on the team with the purpose being to automate the calibration and test protocols specified by Bear. A UML diagram was created from the calibration and testing requirements laid out by the customer. This enabled the software portion to be broken up into various functions assigned to members of the team, where each function has an assigned importance. A debug mode was developed using Visual Studio Software which allowed for the implementation of break points making the software testing process more efficient. Global variables were defined to be used in necessary functions, allowing for each function to be developed individually then brought together to form the finished software product.

Enclosure

Figure 3 shows the bench marking tool was used to determine the best material selection for the enclosure. Aluminum 6061 T6 was chosen as the final enclosure material because it is meets budget requirements while also being durable and strong enough for our testing needs.

Figure 3: Enclosure Material Selection Criteria and Benchmarking

Lexan, or polycarbonate, was the material used for the scatter shield because of its durability and transparency, meeting the engineering and customer requirements.

The enclosure is comprised of two main parts: the transparent Lexan cover (scatter shield) and the aluminum metal enclosure. The Lexan cover acts as protection for the operator from high voltage danger and a rare, but possible dangerous failure in the UUT such as an unlikely explosion. The metal base enclosure houses all of the electrical components such as the PCBs, the power resistors, and the autotransformers. The enclosure contains fans that help dissipate high temperatures caused by the power resistors and MOSFETs on the load board. Temperature analysis was done on the internal components of the enclosure in order to select the fans needed to dissipate the amount of heat produced by the boards. Figure 4 shows the SolidWorks thermal analysis.

Figure 4: SolidWorks Thermal Analysis of Enclosure

The analysis completed shows the heat within the enclosure with no acting fans on the boards, thus proving that the resistors and the power MOSFETs will be the hottest components contained in the enclosure.

From this analysis, it was determined that 40 CFM fans would be needed to provide enough airflow and heat wicking to dissipate the energy produced by the components.

The top of the enclosure has the emergency stop button, the AC voltage meter, and LCD screen which provide additional safety for the operator and feedback. Additionally, the enclosure is a required 1.5’ x 1.5’ x 1’ (L x W x H), with the height being split up between 9” of metal enclosure and 3” of Lexan cover.

PCB Design

Each PCB consists of two layers, the top and the bottom. There were three designed boards: the load board, controller board, and the relay board. The PCBs contain many 45 degree angle traces to maintain ideal performance of current flow. Creepage [1] is the shortest distance between two conductive materials measured along an isolator’s surface in between. Clearance [1] is the shortest distance between two conductive parts as measured through the air. Both were taken into consideration when designing and laying traces. These two factors help determine necessary trace widths and pad diameters [1]. These design requirements are important because it prevents tracking and flashover occurrences. Tracking is when a current follows an undesignated path and flashover is the sparking between two electrical connections.

Controller Board

The controller board acts as the ‘brains’ for the system, containing the Arduino Mega microcontroller and controlling the functionality of the system. It automatically calibrates the control board, and tests various loads verifying the power board both of which make up the final unit to be tested. It controls the load board and relay board, dictating the proper task to execute. It decides which voltages and resistances to need to be used and determines if a system has failed in turn relaying a pass or fail message to the operator via the LCD Screen. The board also houses the SD card slot and the barcode scanner connection. Upon scanning of the UUTs’ barcode, this will signal to the Arduino that there is a unit to test and records the serial number on the SD card. Upon completion of test, other data is stored, like passing/failure of each test, reasons for failure, date/time of test etc. The board is 10” x 5” (L x W) and contains 115 parts. There is a combination of through-hole parts and surface mount parts with varying sizes and package types like 0805, SOIC-8, SOT-23, SOIC-14, DPAK.

Figure 5: Controller Board PCB Layout

Load Board

The load board houses constant current and constant voltage load systems, automatically calibrating to the necessary voltages and currents based on the needs of the test and the commands from the Arduino Mega microcontroller and controller board. This board tests the various voltage and current loads needed to verify the power board of the power supply unit. Various load resistances are also encompassed in this test. The board is two layers, 12” x 5” (L x W), and contains 143 parts. There are horizontal mounted TO-247 heatsinks placed on each of the ten power MOSFETs to help dissipate the generated heat. There is a combination of through-hole parts and surface mount parts with varying sizes and package types like, 2512, 0805, SOIC-8, SOT-23, SOT-23-8, SMA.

Figure 6: Load Board PCB Layout

Relay Board

The relay board is responsible for choosing the proper autotransformer based on commands from the controller board, and supplying AC voltage to the autotransformer chosen to either step up or step down to the necessary voltage. The tests all of the possible AC voltage inputs which allow for the power supply to be universal. There are two autotransformers used, one being the low-line of 85 VAC and the other the high-line of 264 VAC. This board is two layers, 2” x 2” (L x W), and contains 10 parts. There is a combination of through-hole parts and an SMA package type surface mount parts.

Figure 7: Relay Board PCB Layout

Feedback

Feedback to the operator is a dynamic subsystem to indicate if the test bench is ON/OFF and if the test passed/failed along with errors codes for better troubleshooting. An LCD Screen is placed on top of the enclosure to provide feedback for the system indicating if it’s ON/OFF, if a test is in progress, and feedback on the success or failure of a test. Simple Arduino commands were created and error codes were embedded into the software, signaling for proper feedback to the operator upon a test failure. If a test fails, a message displaying when it had failed and a reason for the failure. There is also an AC voltage meter displaying the voltage currently going through the system.