Antarctica - Davis StationSeptember 14 2014
Accept contributions for your expediton by providing us a few details. We will create an account on your behalf at WePay. If you haven't already registered with WePay, they will send you an email to complete your registration.
Second trial dive completed...
After the results from the last dive I shifted some of my ballast weight further aft to see if would help prevent the ROV diving under forward motion. I also sorted out the vlc issue and setup screen castify.
This time we setup at a different location just of one of the rock outcrops near station (Torkler Rocks) for the testing. Once again the lack of natural light meant that visibility was limited, again everything went surprisingly well, the ROV handles better with the ballast change (now rises slightly under forward motion, which is what I was hoping for) and the software and recording went without a hitch this time (though again not a whole lot to be seen other than kelp).
Again the ROV was in the water (only around 5m depth at this site) for around 30 minutes and still had quite a bit left in the tank, I am surprised how well the LiFePO4 batteries hold up in the low temperatures.
One issue I have yet to sort out is the restriction in lateral movement that I suspect is caused by the drag of the floating line on the underside of the ice. I suspect this wont be less of an issue on deeper dives and I have a smaller diameter floating line that I will also try out.
First test dive through the ice completed...
Prior to the dive I had ballasted the ROV in fresh water and calculated the approximate additional ballast required for salt water and attached the tether to a 10m length of floating rope to be used as a recovery/safety line.
For the first deployment, we stuck close to station around 10m from the wharf. The aim was simply to sort out the basics for deployment and retrieval through the ice (currently around 1.2m thick) as well as getting the ballast correct, testing the ROV operation and control and getting an idea of battery life in the cold (air temps around -30degC, water temps -1.8degC).
On site we were able to open up a suitable hole by drilling three holes in a triangle with the 10 inch auger and busting out the remaining ice with crow bars and shovels, before dropping in the ROV. I had expected to spend a fair bit of time adjusting the ballast to get the buoyancy right in salt water, but it seems my guestimations turned out far more accurate than they had any right to be, with the ROV gently rising to the surface at a rate of a few centimeters per second.
Next I parked the ROV under the bottom of the ice a few meters away from the hole to see if there would be any issues recovering the ROV in the event of a loss of power/comms. Once again, everything went without a hitch, so it was time to test out the handling. The sun has only been back for about a week and is quite low on the horizon, this combined with around 20cm of snow on the surface meant it was pitch black below the ice making it a little tricky to work out where you were going when away from the bottom or the ice.
It seems that the ROV dives quite heavily under forward motion. I suspect that this is in part due to the floating rope being too buoyant, as well as the ROV having a slight forward bias in the ballast. This is one area where I need to make some improvements and have been considering adding a forward dive plane that can be adjusted to trim the handling. The ROV was also very slow to surface, admittedly I was only using power factor 2, and it normally probably wouldn't be a concern but when combined with the diving under forward motion it made the handling clumsy.
For the most part everything else worked without issue, the IMU (v1) was faultless and really valuable in the dark (although there are issues with the cockpit display on 1024x768 that I need to sort out) and the batteries held up better than I expected having dropped to 8.2V with 30 minutes of operation before the cockpit software crashed (v30.0.0). Initially I was concerned that the dropout was due to a leak so we began retrieving the ROV. I had set up a VLC script to display, transcode and record the stream from the ROV and looking at this I could see that the video continued to stream following the cockpit crash (I was in no-mans-land between the ice and the bottom but I could dimly see the camera lights blinking in the stream). I was able to restart the cockpit from the dashboard, but unfortunately the VLC script did not handle this well and began a new recording over the top of the dive footage (there wasn't anything terribly exciting in this, so I am not too bothered) when the cockpit restarted.
Once again, it has been far too long between updates but things are moving in the right direction again now. I have made a few more modifications to the rov in preparation for its first deployment and have received authorisation for working in the waters around station.
As a requirement of the the ATEP act, I had to get authorisation for operating the ROV in Antarctica, this required me submit an environmental impact assessment outlining potential environmental hazards (covering everything from wildlife disturbance to the environment contamination). As part of this process, I went through a bit of a design review looking at potential failure modes in the ROV design and how they can be overcome.
One of the things that came up was the battery containment, if a retaining o-ring breaks, or there is a failure of the battery tube or end-cap, the batteries could be lost to the environment. The o-rings used for retaining the e-tube and battery end-caps were already showing signs of deterioration (presumably drying out due the the climate down here), so I was looking for a new way to achieve this too.
So the aim was to:
- Retain the batteries in the event of a battery tube failure
- Replace the o-rings and prevent the end-caps from coming out during a dive
On top of this I also wanted to achieve the following:
- Securely mount the battery tubes to the main structure
- Allow for easy removal/access for battery replacement
- Require no modification to the existing ROV components (because I hate making irreversible changes only to find out it doesn't work well)
- Stay in keeping with the ROV aesthetics
After trying out a few different ideas I eventually settled on acrylic outer tubes/mounts, these were formed from 3mm acrylic sheet moulded over a knock-up form. It took a few attempts to get right, but I am happy with the end result. They clip into the upper cable-tie holes and are secured with nuts on the threaded rod, when in place I found they do a really good job of mounting the battery tubes and actually stiffen up the whole structure. There was a bit of a tolerance issue with the E-tube so I had to file out a small divot of the top of the mounts but otherwise they fit perfectly.
I also made up some battery sleeves to make inserting and removing the batteries a bit easier, these also help keep the batteries in line improving their contact with one another. The sleeves are glued to the poly-fuses at one end, have a small gap for the positive lead to run along and little tabs to make it easy to pull out the sleeve/batteries.
The E-tube mounting was changed over to velcro straps with rods that run the length located at each end in the holes for the O-rings, this also gets rid of the issue with the O-rings interfering with the lip of the end-caps (which I was never really comfortable with).
Apologies to those following for the lack of updates but after an initial flurry of activity on the ROV build front at the start of the season the project got shelved as some of my summer work projects were dominating my time. Now that the winter is here I will hopefully have a little more time to keep the updates coming.
The good news is that my ROV is now basically complete and will hopefully be making its first dive in the near future, all that it left to be done before this is
For the most part the build went fairly smoothly with most of the issues I came across already covered in the forums. I am not 100% happy with the e-tube end caps as I had some troubles getting the right amount of acrylic cement, but given most of the waters around here are less than 30m, I am not too concerned.
- For my initial testing I was using a benchtop power supply which caused a few issues. It seems that the inrush current of the ESCs and to a lesser extent the beaglebone and camera would cause a brown-out of the ROV on startup. Using a 6A power supply would allow the BBB to boot normally if the ESCs were switched off prior to startup, and turning on the ESCs one at a time after the BBB has booted after which the ROV operated normally.
- As previously raised in the forums, I found during initial leak testing that water was getting in under the o-rings. Some light sanding, a brush of acrylic cement and a coat of vasoline on the o-ring sorted this out.
- The usual ESC programming hiccupps
- I was having trouble getting the electronics tube to slide over the boards, largely due to the ESC switches getting in the way, and finding no suitable place to secure them (I didn't like them hanging loose as they may wind up shorting out the electronics) so I decided to just cut them off. With a few crimp terminals, some heatshrink and some header strip I knocked up a fly-lead connector, so I can plug the switches back in if I need to reprogram them in the future, or plug in a jumper to short out the power switch in normal operation.
- The USB connector for the webcam was a very tight fit and the stiffness of the cable coming from the rear of the webcam was adding load to the tilt servo so I modified the cable. First, the rear housing was removed and the hot glue holding the cable to the rear housing picked out. This allowed the connector to be removed from the camera board and passed through the rear housing. Next, the USB connector was cut off (I shortened the USB cable too) and a short right angle connector was made from the original. Leaving the rear housing off the camera means the tilt servo only has to bend the four very light gauge wires that go to the PCB connector instead of the much stiffer USB cable.
- Instead of securing the props with epoxy (as I wanted to be able to re-use them if I need to replace the motors in future) some heatshrink was applied to the motor shaft and glued in place before the prop and a locking nut was threaded onto the shaft.
- To reduce the glare in the camera produced by the lights, I used some adhesive vinyl/felt stuff to blackout the area around the camera lens and e-tube.
We arrived at Davis last week, and have been kept busy with resupply operations, with around 800,000 litres of fuel to be pumped ashore and several hundred tons of cargo to be delivered to station, sorted and stowed. Most of the ROV kit has come out of cargo (the acrylic cement is still sitting in a container somewhere on station) so I am hoping to get some spare time over the next few days to start building.
The good ship Aurora Australis, parked up in the fast ice around 2km off Davis:
Preparations continue - The OpenROV kit, materials and a few additional bits and pieces I have ordered for potential add-ons are now all consigned to cargo. At this stage the ship is scheduled to arrive at Davis around the 6th of November though by all accounts it has been a bumper year for the sea ice off Antarctica, so I wouldn't be surprised if it takes a little longer.
In the mean time training continues, so far I have completed the medical training (one of my secondary roles on station is as a lay surgical assistant), some technical training and training for the various policies and procedures. Next week I will be doing the fire fighting training with the rest of the winter crew.
Here is some of the kit issued by the Australian Antarctic Division.
I am heading back to Davis for an 18 month stint as the Electronics Engineer and purchased an OpenROV kit to build and explore while down south. The Aurora Australis is scheduled to depart on the 22nd of October and is expected to take around three weeks to make her way to Davis. Once there I plan to build and test the ROV before hopefully deploying it at a number of locations around the Davis area.