During the previous year I have been working on another project, and haven’t had much time for working on the DPV. Still, I have been able to complete all the mechanical work, as well as building the battery pack. In fact, the DPV is ready for testing in water at this point, but the motor controller has stopped working so currently I am waiting for a replacement.
So, what have I done to complete the project? First, I started addressing the problem of eddy current losses in the magnetic coupling. As described in the previous update, the magnetic losses in the coupling were huge, and something had to be done. I decided to remachine the thickness of the barrier in the tailcone from 3 mm to 1 mm, at the expense of a lower depth-rating. Even though a deep maximum depth is nice, I am only diving to 30 meters depth and would rather have longer battery life than deep depth-rating.
Machining the barrier to a lower thickness reduced the total losses by approximately 75 W at 800 rpm and 115 W at 1000 rpm. These losses include the internal losses in the motor, as well as bearing friction. From this graph it is seen that adjusting the propeller to maximum pitch and keeping the rpm low during cruising will give the longest range.
After the tailcone was modified, all the machined components were completed. The parts were then sanded smooth, before they were anodized black to improve corrosion resistance (and look cool).
One thing that had bothered me for a long time was the weight of the motor. In order to achieve balance underwater, the center of gravity of the DPV has to coincide with the center of buoyancy. Since the tailcone is tapered, the center of buoyancy is shifted to the front. Having a heavy motor in the back is therefore problematic when trying to achieve balance, and the solution is to make the body longer and place balance weights in the front lid. I wanted to avoid this because it would increase the weight of the DPV, so I decided to shave some weight from the motor casing.
Upon removing the rear lid of the motor, I noticed that the circumferential casing was 8 mm thick steel. Taking this into account, I removed around 500 grams from the casing by turning it in the lathe.
The casing has probably been made by rolling a plate and welding it together to form a cylinder. The person who welded it has not welded it all the way through, so upon turning I noticed that a longitudinal crack in the casing had emerged. I hope this will not cause any trouble later.
Next, I started working on the battery pack. Sixteen 10Ah LiFePO4 cells were ordered from a company called Alternativ Energi AS in Norway, and the rest of the electronics (charger, cell monitors, cables, etc.) were ordered from Hobbyking and Ebay. I chose to go with the iCharger 208b from Hobbyking because it had the ability to charge and balance 8 cells at the same time, as well as having the possibility to burn-test the cells by connecting an external resistance.
When I got the cells, the first thing I did was to check that they fitted inside the body of the DPV. I knew it would fit if the dimensions of the cells were correct according to the website, but I didn’t have much tolerance.
The cells fitted nicely, so I started building the plates that will hold the cells in place.
I printed a drawing with the position of all the cells, and used a punch to transfer the hole layout to the acrylic plates that were going to become the cell mounting plates. Once the mounting plates were finished, I sandwiched all the cells between these two plates so that the threaded cell tabs were protruding on each side. Next, I made 16 cell connections by soldering ring crimp connectors on each end of a short wire. This way, I could connect all the cells in series to get the desired voltage.
Once the cell connections were made, I started wiring the leads that will be used to balance and monitor the cells. This was a lot of work, and it also needed sharp focus to avoid making a wrong connection.
Building the battery pack was a lot of fun and a good learning experience. I don’t think I saved any money by building it myself, but it was the only way to get a pack that would fit as compact as possible into the DPV.
At the same time the battery pack was finished, the aluminium parts came back from anodizing. They looked great! The only things left to do at this point was to fit the PE100 plastic body tube to the tailcone and frontlid, mount the latches to the stainless steel bands around the body of the DPV, make a main power switch and connect all the electronics together.
The PE100 plastic tube had a rather rough inner surface, and had to be machined. It was quite difficult to mount a Ø225 500 mm long plastic pipe in the lathe and get a nice result, but somehow it worked out by taking very light cuts.
The latches were pop-riveted to the stainless steel bands and tightened around the body of the DPV. This is the same method I used on my HID diving torch, and has proved to work fine.
I decided to incorporate a main power switch so that the DPV cannot inadvertently be switched on during transportation. The switch is mounted on the stainless steel band in the tail end of the DPV. It has a magnet that activates a reed switch inside the DPV, that in turn activates the main power relay. The switch is made so that it can only be in the on or off position. This was achieved by milling the curvature of the switch so that is matches the curvature of the body tube.
This concluded the mechanical work on the DPV. Connecting the electronics took some time, but I got everything working after a while. Unfortunately, the motor controller stopped working before I got to test it in water. After having tested the DPV in air, I disconnected the battery and set it aside a couple of days. When I connected the battery for testing in the bathtub, nothing worked. I did some measuring on the controller, and found out that the 5v output that drives the hall-effect sensor only gave around 1.6 v and was not stable. Searching the internet, I found that the controller was most likely defect. I have ordered a new one, but have not received it yet.
Having access to the composite lab at the university, I was able to vacuum test the DPV for leaks once everything had been assembled. No leaks were detected!
Finally, here are some pictures of the finished DPV:
And here is me with the finished DPV:
Thanks for reading!