Tom describes his process of designing the frame:
Structural Design and Integration (Tom Gros)
The final design agreed upon for Project ROV-ITT was a 9 inch x 9 inch x 12 inch frame structure made of ½-inch PVC. Preliminary plans did not include a frame structure at all. Instead they called for the electronics of this project to be housed in everything from capped 3-inch PVC pipe to garden spray canisters. Each option was eliminated one-by-one for various reasons. For example, the option of using a garden sprayer was eliminated because, although it would be large enough to house all of the components and easily waterproofed, the opening was too small to mount the components. Generally, options were eliminated for one or more of the following reasons:
Access to the circuit board would have been difficult or even impossible.
The craft would not have been able to be sealed or resealed.
Securely mounting the bilge pumps without jeopardizing the structural integrity of the craft would have been implausible.
Waterproof routing of the power cables and umbilical to the craft’s interior would have been an issue.
Camera placement would have required the integration of a watertight lens.
Even the PVC frame underwent several design changes. During the first week, we discussed, as a group, different options for the frame’s construction. We concluded that, by mounting some of the components such as the lights, on the outside of the craft and locating others, like the battery, to the surface, a small, light-weight and watertight case would be the perfect option to house a circuit board and camera. The frame would therefore be constructed around the case containing the “brain” of the craft. The initial version was laid out and measured on a two-dimensional plane which resulted in the prototype being oversized and cumbersome—approximately twice the size of the final version. I determined the final design by removing half of the structure and rearranging the placement of the components to be integrated into a more efficient configuration.
The next major challenge consisted of mounting the case used to house the circuit board and bilge pumps used for propulsion onto the frame. The main problem in this endeavor was the fact that both pump housing and PVC frame are round. Merely attaching the pumps to the frame by means of wire ties provided no stability—causing the pumps to slide around the frame when producing thrust. One solution that I quickly eliminated was to dismantle the pump in order to attach the frame by means of bolts. The reason this option was eliminated was the fact that, considering that there are six pumps, this would have been too time consuming and there would be no guarantee that the problem would be solved. Rather, I decided upon a more efficient solution. It was much simpler to attach a floor to the frame to which the pumps could be mounted securely. The materials I decided on to construct this floor were plastic gutter screens and lightweight metal truss ties used to provide stability. The truss ties were also decided upon to be used as mounts for the plastic case.
The remainder of my involvement to this project, until the first test run, consisted of the construction of the umbilical cable and mounting the camera. The umbilical was constructed by zip-tying a section of Cat-5 cable, used for video transmission, to RG-45 cable, used for vessel control and power. Since the watertight case contained a clear lid, mounting the camera was easily accomplished by simply attaching it using double-sided tape. I also completed the final integration of the bilge pumps to the frame. The remainder of my time consisted of helping Ronnie with the integration of the electronics to the structure of the craft. This included routing the wiring from both pumps and umbilical through hose barbs screwed into 6 holes we drilled into the case housing.
The initial test run identified two problems with the structural design of the craft: It was unable to move forward in a straight line and the bilge pump responsible for upward movement did not provide enough thrust to allow it to surface. Emery and I repositioned the pumps to a horizontal rather than vertical orientation in an effort to stabilize them, allowing the craft to move forward without veering to the left. The pump intended to provide downward thrust was remounted to help resolve the lack of surfacing ability. The repositioning of all six pumps also allowed for the remounting of the case from a diagonal to a horizontal alignment, reducing drag and, theoretically, further addressing the veering problem. Other than these two issues, the design merely required some trimming to level its position in the water.
(note: Tom mixed up his cables! The Cat 5 sends power and RS232 data, while the RG-45 sends video.)