Progress Report - Mechanical/Electrical Design (02/22/2012)

Since my most recent update, I have continued sedulous work on the AUV. I properly and securely installed all of the SEACON connectors to both the hull and camera enclosure end caps (adding silicone grease to each of the o-rings as well as the male plug connectors, and applying Loctite thread sealant to the middle threads of the female socket connectors), completed the construction of the camera enclosures (and successfully installed the forward facing camera in the top enclosure), designed and manufactured the grasp/release mechanism, waterproofed the "waterproof" (hobby servos like the one carried over from last year's team are not rated for sustained submersion of up to 16 ft--not even close) servo motor (Traxxas 2056), designed, manufactured, and installed the kill switch mount and kill switch (features a military grade rubber boot (IP86* rating--sustained submersion)), installed the newly arrived pressure transducer, developed a design for the amplification/filtering PCB for the hydrophones, and designed and manufactured the solenoid valve interface PCB using  a program called Dip Trace which I downloaded and then learned via a tutorial and introductory documentation in a day.

The grasp release mechanism design consists essentially of mirrored 3-bar (or 4-bar in a liberally general sense) slider crank mechanism. I briefly consulted Dr. Hollis (Mechanical Systems II and Modeling and Simulation Professor) in regards to a couple of the design details (e.g. the use of bushings over bearings, lubrication suggestions, etc.) as well as an approach to deriving the motion profile of the proposed linkage system. I ended up reviewing my Mechanical Systems I text book (I keep all my engineering and advanced mathematics books) and found a set of equations that could be directly applied to this model. It was during these calculations when I realized that the current design wouldn't output the intended range of motion of the jaws, and in fact would have induced significant stress build-up due to not physically being able to move in such a way in order to allow the single-acting air cylinder to translate the full 2" during actuation. I then set up a sketch on Pro/E, allowing me to make immediate design modifications by providing an efficient means of adjusting various design parameters (link lengths (i.e. distance between pivots) and claw shapes). Ultimately, I concluded that it was best to get a new ($18) single-acting air cylinder that has 0.5" of travel (it looks the same as the other single-acting air cylinder, but is about 2" shorter as well). This modification, along with the redesigned jaws provided the desired motion of the grasp/release mechanism, and without hindering the ability of the single-acting air cylinder to extend to its full capacity. The design features 8.5" of grasping range   (closing to a very slight out-of-plane overlap), a custom piston adapter, and a fixed pivot (achieved via minor additions to the frame, as well as another custom adapter piece). Furthermore, it uses bushings, 1/4" - stainless steel dowel pins, e-clips, and shaft shims.

I successfully tested the newly acquired Traxxas 2056 servo motor a few days ago as well, serving as final confirming evidence that the other two servo motors were, in fact, burned out as suspected. After extensive research on servo motor waterproofing methods, materials were ordered, and test were performed on these dead servos. The conclusion was to apply silicone grease to the output shaft of the gear train, submerge and assemble the servo in a high-viscosity mineral oil bath, ensure the screws were tightly fastened (the screws already had o-rings on them), clean the shell using alcohol, apply Duco Cement along the two interfaces of the servo shell (as insurance), add an lubricated (silicone grease) o-ring around the output shaft (externally), and then tightly screw the servo horn onto the output shaft. Finally, Plasti-Dip was used where the wires protrude from the back of the servo motor in order to ensure a definitive water tight seal.

The kill switch issue was the other waterproof task that was on my mind. The original source provided an astronomical quote that even earned complaints from their most recent customers--the Bellagio and the Navy. It was decided that it was far more economical (and necessary) to attempt to custom waterproof the toggle switch that was inherited from last year's team. Rough ideas were thrown around, including the implementation of a balloon around the toggle switch to prevent water while not restricting actuation. The general concept was decent, but I wasn't too fond of the sophomoric technique that was being implied. Thus, I conducted more extensive research and discovered an established company that manufactures silicone boots that are directly compatible with a vast array of toggle, push button, etc. switches. An appropriate boot for the toggle switch was obtained and this, combined with the addition of Plasti-Dip to the leads of the switch, should provide a reliable water tight seal. Furthermore, it has been decided to use this comfortable-actuation switch not only as our kill switch, but also as our mission start switch--preventing the need for an additional switch, and thus additional threaded hole in the rear end cap for the complementary set of SEACON connectors. The switch is conveniently mounted on the top of the back face of the AUV--directly behind the compressed air tank, and out of the way of thrusters or any mechanical subsystem.

I also discovered, installed, and briefly learned a PCB design program called Dip Trace. I was able to derive the solenoid valve interface PCB (containing four identical circuits in a compact arrangement) which has dimensions of 3.35" x 1.45", and also has built-in mounting holes. The transistors, resistors, solenoid valve wires, and protection/snubber/flyback diodes will simply need to be soldered to this board. Lastly, I sat down with Antony on Friday and confirmed my initial determination that we in fact need 25 dB of amplification in the pre-amplifier for the hydrophones. I have derived circuit design with the aid of National Semiconductor's Webench, which will consist of of four identical circuits containing a third-order active band-pass filter connected to two gain stages (i.e. two-stage amplifier using op-amps). The active filter should sufficiently filter out noise from the delicate hydrophones, and the amplifier should provide the necessary gain in signal so that as the AUV enters into its pinger detection state (following completion of the Kill Caesar mission), the signal amplitude should be about 0.4 V, and as the AUV approaches about 1.5 m, the signal strength should be about 4.5 V. Since these signals will be sent to A/D converters for eventual digital signal processing, it was critical to design the circuit so that the signal would not saturate the A/D converters within the desired range from the target pinger. I will construct this complex and robust circuit using Dip Trace as well. However, due to the severe magnitude of this particular PCB, I will likely end up sending the file off to a company (possibly National Semiconductor) to construct the custom PCB using more advanced milling equipment than is currently available at the FAMU - FSU College of Engineering.

*We are also currently pursuing further funding from the college, and also are in the process of submitting/have submitted the competition entry fee, and requested the lending of a transponder (pinger) from the competition staff (they have items that teams can borrow for up to four weeks, including IMUs, PCs, pingers, hydrophones, etc.).

I will update my blog again at the start of spring break. Until then, I will be continuing the progress of the AUV until the task is done, we go to competition, and we win the competition.

      -Eric Sloan (Mechanical Engineering Project Manager - 15th Annual AUVSI Robosub Competition)

Progress Report - Mechanical Design (02/12/2012)

Since my latest blog post, the AUV has continued to progress significantly. The compressed air tank, tank/high-pressure regulator, and secondary/low-pressure regulator were successfully installed their simple, elegant Delrin mount. The tank was filled to capacity (i.e. 3,000 psi) at a local scuba diving shop named Coral Reef, and successfully integrated with the network of nylon gas distribution tubes. Furthermore, the single-acting air cylinder for the grasp/release mechanism was successfully installed on its recently designed and manufactured mount. The simple, efficient angle-bracket mount was intelligently designed for its efficient use of material, relative ease of manufacture, and with dimensions that enable the jaws of the grasp/release mechanism to successfully pick up the rescue object (i.e. laurel wreath) without the bottom camera enclosure interfering. The single-acting air cylinder, as well as the torpedo launchers (with the redesigned (embedded neodymium magnets in place of the appended low-pull bar magnets), low friction, proper density, properly balanced torpedoes) were successfully tested in air.

An aluminum heat dissipation platform and the electronics rack were also successfully redesigned and manufactured. This redesign was prompted by the recent decision to use the Zotac mini-itx board in place of the Beagleboard-xM as the main control unit due to the need for greater processing power (particularly for the computer vision algorithms). The new heat dissipation platform was designed and manufactured to slide snugly into the hull end cap slots, and also to provide a wider region in the center (between the two lithium-ion batteries) on which to mount the new 6.7" x 6.7" Zotac PC. The revised electronics rack features one level which spans the majority of the interior hull, and essentially forms a bridge over the batteries and Zotac PC. This electronics rack not only features an efficient use of material and an aesthetic appeal, but also provides much easier access to the circuit boards than in the previous multi-level design. Furthermore, the redesigned electronics rack uses the same amount of material as the original design, and also is very cleanly welded at the joints (in place of cumbersome, interfering nuts, bolts, and angle brackets) to provide a strong, durable, elegant result.

The SEACON connectors also arrived, so I carefully drilled and tapped (using a mill tool plate attachment that operates in polar coordinates) each of the aluminum end caps to enable the installation of the female bulk head connectors. In addition, I cut the wires of each of the peripheral subsystems to about 3", and then measured and cut each of the corresponding male SEACON cables to length in order to ensure the reduction of excess cable (and thus weight). Each of the cables was then stripped to prepare them for soldering, heat shrink, and Plasti-Dip on Monday. The male cables for the SBT150 thrusters were also soldered to the leads of the thrusters inside the back end caps, and the cable was then potted using a waterproof epoxy. Plast-Dip will also be brushed over the external epoxy in order to provide a slightly cleaner look, as well as another layer of protection against water leakage.

The new acrylic camera enclosure boxes also arrived and were carefully mounted to their redesigned supporting platforms. I conducted exact measurements of these enclosures in order to ensure that their respective fixed end caps will fit firmly and securely, reducing the potential for leakage following the addition of the 100% silicone caulking around the four outer edges. The redesigned mounts for the cameras were also successfully installed (in the center of each of the enclosures (and the vehicle)) recently. I discovered a product called Duco Cement which was proved to be a fantastic find. It is a completely clear, very strong adhesive that dries within about 15 minutes. It has been used to adhere the camera enclosures to their respective aluminum base plates, as well as to adhere both of the camera enclosure mounts to the inside surface of the acrylic enclosures, and to adhere the polycarbonate pressure transducer supports to the center line of the top of the acrylic hull.

By the end of this upcoming week, I expect to have the redesigned fixed end caps for the camera enclosures manufactured and installed, all the male cables successfully connected to their respective peripheral subsystems (including the pressure transducer, which should arrive by Friday, but excluding the kill switch and the servo motor of the marker dropper (both will be ordered Monday)), the electronics rack successfully installed inside the hull with holes drilled in line with the circuit board mounting holes, the thrusters re-installed on the vehicle, the files required to make the solenoid valve PCBs obtained and submitted, the grasp/release mechanism jaws water cut, and the pre-amplifier circuit for the hydrophones tested on a breadboard. The kill switch (which will now be mounted to the frame on the back of the vehicle behind the compressed air tank), new servo motor, and new SEACON connectors (for the kill switch and to enable code to be uploaded to the main control unit without having to repeatedly open the end caps) will also be ordered on Monday, and should arrive the following week.

The vehicle is still on pace for mechanical and electrical completion by spring break (i.e. March 5th), but there is considerable work to do until that point in order to reach this goal. Currently my main concern is interference in the signal from the hydrophones, and the successful design and integration of the custom pre-amplifiers for the hydrophones. I will be drawing on my Mechatronics II knowledge, experience, and resources in order to make this a success as well. Also, once the vehicle is successfully completed and refined in preparation for underwater mission testing in March and April, I will direct my attention to helping Hang Zhang develop control algorithms for the thrusters using information from the IMU and pressure transducer as input sensory information.

Lastly, the vehicle was recently weighed, and is on pace for a projected 87 lb. Due to the 84 lb threshold, we might lose a few points (110 lb is the max weight in order to qualify to compete). This is not of major concern as this was both anticipated, and has a relatively small impact in reference to the entire mission. Furthermore, there is a potential that either (a) the weight limit will be increased slightly, or (b) this projected weight ends up being a slight overestimate.

I will provide another update on the status of the AUV in the coming weeks.

                      -Eric Sloan (ME Project Manager, 15th Annual AUVSI Robosub Competition)