Integrated System Build & Test
Table of Contents
Team Vision for Integrated System Build & Test PhaseThe Team’s Plan for this phase was to complete all of the subsystems and move as far as possible into full system integration. From the Mechanical side, all subsystems are complete besides the pontoons which are having issues with holding air. From the EE/CE side, Subsystem items have been almost all but closed out, and much progress has been made in integration. Towards the end of the phase, full system integration has started. The team has planned out the end of the semester, and also started on some of the final documentation. Upon the pontoons being able to hold air, testing will start ASAP, starting with the pool, and then hopefully to full scale testing at location.
CartA Buoy Deployment Card was built to be used for system-level tests in real open-water environments, and for transportation of buoy to and from vehicles. The black painted metal section of the cart was sourced from Brinkman Lab scrap. It was modified from original state first by cutting off its obstructing metal uprights, done using a Plasma Cutter with the help of Brinkman Lab staff. A wooden frame was then constructed onto the cart to allow the buoy to rest elevated from the ground. The wooden frame is bolted to the cart, and held together by wood screws.
Pontoon Grommets and mounting to frameGrommets were added to both pontoons to allow attachment to frame. The grommet holes were created using the punch included with the grommet kit, and the grommets were simply hammered into place. Gold colored grommets were used along the length of the pontoons, and one silver grommet indicates the aft direction for proper assembly. The grommets were placed in-line with bolt holes on the frame. Bolts attach each pontoon to the frame with wingnuts for fast and simple assembly and disassembly.
SteeringThe steering linear actuator is attached to the motor on one end, and to a frame upright at the other end. A ¼” clevis pin attaches the actuator to the motor arm. A mounting bracket with clevis pin assembly designed to mount the other end was purchased and used. A new, longer frame upright, where this bracket is mounted, was cut to make the linear actuator rest level. This subsystem has been tested and works as intended.
Electronics Box mounting
Brackets for mounting the electronics box were cut from 1/8 inch scrap sheet aluminum and 1/8 inch scrap aluminum angle stock. They were assembled and bolted to the frame so the box can rest above the winch.
Pontoon Build and Test
As they are one of the most critical parts of the prototype, a lot of care and attention has gone into the testing of the pontoons. In order the test for the pontoons holding pressure, a squirt bottle was used to spray the seams of each pontoon with soapy water. This water would then bubble in places where there were leaks.
After finding the leaks, the first solution was to go back over those seams with the vinyl cement. After several iterations of sealing and leak testing, the round seams attaching the cones to the cylinders no longer leaked. The straight seams still did however, both from in between the layers and out of the stitching. Because the cementin clearly wasn’t working for these areas, a new approach was taken, heat sealing. This was done using a typical household clothing iron, some aluminum foil, and waxed paper, resulting in the layers of fabric melting together. An initial test of half of one of the seams turned out quite successful, so we decided to move to trying this method for all of the straight seams.
After the iteration of heat sealing and testing, the pontoons did still leak, but we did notice much less leaking, and they were able to hold air much longer. The areas which seemed to fail were spots where the heat seal didn't go all the way down to the stitch. Therefore, a second iteration of heat sealing was conducted, going about .5” below the seam. Additionally, another line of vinyl cement was applied over the seams as the act of ironing may have damaged the original coating.
After doing another round of leak testing, it turned out that the second round of heat sealing was quite effective. The only remaining leaks on the pontoons are at the 4 corners where the straight seams meet the cones. Another iteration of sealing just these seams will be done. Also, Flex seal is on order as a backup method for sealing over these and any other tricky seams.
We designed and fabricated a bracket and mounted lights.
Misc. - Mechanical
As per the original design, bungee cords running from the corners of the solar panel down to the corners of the frame will provide structure to the visible cover that will go over the frame. At the same time, they will be resistant to impacts from a boom swinging by and hitting the buoy, as opposed to the metal to metal contact of a rigid support which could cause damage to the buoy, or to the boat.
Linear Actuator IntegrationIn previous phases, the linear actuator had been interacted with on a basic level. This involved having the arduino control the linear actuator to have it fully extend, fully retract, and come to a point in the middle of the two. This was done with code separate from the rest of the code, written only to test the functionality of the linear actuator. In this phase, work was done to integrate the written test code into a usable function to steer the buoy.
The function created takes in a percentage, denoted by a decimal number between 1 and 0, and extends the linear actuator to the given percent of full extension. For example, to have the buoy straighten the buoy, a percentage of .50 would be given so that the linear actuator would extend 50%, the point at which the linear actuator was set up by the MEs to have the motor pointing straight ahead. The function can be found in the navigation algorithm file here.
The linear actuator function was tested while it was hooked up to the motor and the frame of the buoy. The test was very successful as the linear actuator extended to the given percentages, turning the motor left, right, and center on command. A photo of the setup can be seen below.
IMU and GPS Integration TestingAn integration test of the IMU, GPS, and main navigation algorithm was executed. The arduino due, IMU, and GPS were loaded into the electronics box and a laptop was put on top of the box. Using hard coded goal coordinates, the program used the GPS to get its current coordinates and calculate the heading needed to point towards the goal coordinates. The IMU output the current heading that the system was facing, and by walking around campus and pointing the electronics box in different directions, the heading values were updated in real time. The components worked as they were supposed to, but there was a slight problem with the headings. The heading calculation and the heading measurement from the IMU were not offset to match values. We were able to determine if we were pointing the right spot if the difference between the measured and calculated headings plateaued at any value, but the desired effect was to have the difference equal 0 when we were pointing in the right direction. The testing was a partial success, and the failure section was pinpointed.
IMU IntegrationAfter the testing mentioned previously, more work on the IMU integration was done. The measured heading needs to be offset due to the calibration used. The exact calibration of the IMU doesn’t necessarily matter, but depending on the calibration values used, an offset needs to be added to make the readings match the calculated heading values. Both measured and calculated headings also needed to be normalized if they happened to be a negative value. These changes will make the heading difference equal to 0 when the buoy is pointing in the right direction, which will help greatly in the navigation algorithm.
Winch IntegrationMore time was spent integrating motor control with the winch motor. Specifically, the ability to detect when the anchor has returned to its starting position using current sensing. When the anchor is returned, it will be held in place, preventing the motor from spinning anymore, and increasing the current draw. Once the current exceeds a predetermined threshold value, power to the motor will be cut.
Using a 15 pound backpack as a makeshift anchor, we were successfully able to implement the current sense feedback control with the winch motor. The feedback current values from the Pololu driver appear to be inaccurate (possibly due to an offset in the measurements), but are at least consistent. With calibration efforts, we can confidently use current sense feedback for the anchor.
Overall System Circuit and Connections/ConnectionsThe circuit shown in the picture, shows the connectors that are going to be used in the full system integration. The circuit also highlights the parts that are going to be inside the electrical box (shown in the dotted red line).
A Terminal Block is going to be placed inside the box, through which all the connections are going to be made between the components outside the box and the ones inside it. Firstly, this makes it easier to make these connections and secondly, by this a strain relief can be added as well.
For the interconnections between the internal components two different methods are going to be used. The connections between the Arduino and the two motor controllers (Pololu and adafruit) are going to be using the Arduino shield. For the connections between the Arduino and the Xbee, the IMU and the GPS, a breadboard is going to be used. The driving transistors (BJTs) controlling the LEDs outside the box are going to be placed on the same breadboard.
Risk and Problem Tracking
The Phase 7 Risk Assessment document and the Phase 7 Problem Tracking document can be found here.