Timeline of Operations for ROV Work at Nikumaroro, June 2015
by Walt Holm
As collected from the daily logs of:
|Operations off of the M/V Naia
||Seabotix VLBV950 ROV with 350m tether
Saturday, 13 June:
Nai’a arrived at the island at lunchtime. John went ashore, while Ron and I set up the DGPS (differential GPS- provides both GPS position and true heading) antenna, the ROV power supply, and the ROV workstation.
Sunday, 14 June:
We installed the 1Cam HD Camera/Scaling Laser set on the bottom of the ROV, and made some skid extensions to keep the camera from touching the deck with the ROV on deck. The USBL dunking transducer was installed on the ship’s keel. We installed the aluminum-housing HD primary camera that Ron had purchased for this expedition on the ROV (it had arrived just prior to shipment). Upon testing, we found the new camera to be inoperative right out of the box. We switched back to the titanium-housed HD camera that Ron had been using as a loaner since the purchase.
We spent some time playing with the power adaptors for Nai’a's 220V power sockets, and ended up taping them in place to achieve reliable power.
We set the ROV in the water off of the rear deck of Nai’a to do an initial adjustment of the weight and balance of the vehicle.
Monday, 15 June:
We measured the position offsets for the USBL dunking transducer and the DGPS antenna, and entered them into the navigation software. We also entered the GPS coordinates of the prime target/anomaly (“Richie’s Anomaly” at ~200m depth) into the navigation system as a target.
Set up for our first in-water test. Laid some cable on deck and went to place the vehicle in the water. Noticed a brief interruption of the video data when the ROV was hauled onto the ship’s rail, but it reset itself. The vehicle was lowered into the water. Driving the vehicle away from the ship caused a fault in the ROV power supply. We started investigating this fault by repeatedly placing the vehicle in the water. A consistent ground fault happened in the vehicle upon drawing ~150 mA of current from the main 345V power supply. The fault code on the power supply did not indicate whether the issue was on the AC side (Nai’a → ROV power) or the DC side (ROV power → ROV) of the power system.
After lunch, we checked, cleaned, and re-greased all of the power connections on the vehicle. The two vertical thrusters were replaced with spares as they had traces of cloudiness in their compensation oil. Satellite phone calls were placed 3 times to Seabotix techs, trying to interpret the fault codes being shown on the power supply. It was still unclear whether the ground faults were on the AC side or the DC side of the power supply.
Tuesday, 16 June:
We started the morning by talking with Nai’a’s engineer, trying to better understand the ship’s power supply. During this time we got an E-mail from Seabotix telling us that the fault codes showed that the ground fault was on the DC side (ROV power supply → ROV) of the power system.
At this point we started looking at the ROV for ground faults. We borrowed Lee Paynter’s largest Peli-Case, which was just large enough to fit the ROV; the case could be filled with water to create a basic test tank and simulate a dive while the ROV was still on the ship’s deck.
With the test tank filled with fresh water, no ground faults were observed. The tank was refilled with ocean (salt) water, and again no faults were observed. At that point we placed a ground wire in the tank, and connected the other end of the wire to the ships ground. In this configuration we were able to re-create the ground fault condition.
Tests continued to narrow down the cause of the fault, by submerging only portions of the vehicle in the salt water tank, and by using the ground wire as a probe. We found that the auxiliary HD Camera/ Scaling laser (a SubC Imaging “1Cam”) was causing the fault. This camera had come back from factory overhaul just prior to the trip, and there must have been some sort of short to the outer case inside the housing.
We switched to the spare 1Cam, tested that it did not have a ground fault condition, and broke for lunch.
After lunch we put the ROV in the water, and had a ground fault trip again. We pulled the vehicle out of the water and placed it back in the test tank. We narrowed the issue down to the port aft section of the vehicle by noting bubbles emanating from the tanks ground wire when the ROV power was on and the offending circuitry was placed in the tank. We pulled the ROV out of the test tank, and checked for leakage current between the ROV and ships ground using a micro-ammeter. Most metal parts of the vehicle showed between 5 and 20 uA of leakage; when the metal parts of the port aft thruster were probed, there was 1800 uA of leakage current. The port aft thruster was changed to a spare; examination of the bad thruster showed no signs of water intrusion.
At this point we turned our attention to configuring the USBL positioning (navigation) software, trying to get it configured the way we wanted for the ROV operations, with only partial success. We did a test dive with the ROV, driving it around Nai’a and checking for correct basic operation. We also did final trimming of the ROV weight and balance.
Wednesday, 17 June:
Walt rode in a skiff out to Nai’a’s strobe buoy (this buoy is used at night to serve as a visual reference of the reef’s edge), and took GPS coordinates of its location. The buoy and its anchor will be used as a test target for calibration of the ROV navigation (USBL) system. During this period Ron and John prepared the ROV for diving.
Once Nai’a was in position and securely moored, we placed the ROV in the water. Very quickly we ended up with a communication failure in the vehicle. We pulled the ROV back onto the deck, and discovered that we had a failure of the main fiber in the tether. We swapped in the secondary fiber (the tether has 2 fibers, one used for control and SD video, the second used for our add-on HD camera), and still had issues with intermittent communications.
The afternoon was spent cleaning and re-seating all the fiber connections, including those inside the pressurized E-bottle of the ROV.
Once the ROV is working and tested, the tether will need to be re-terminated to get HD camera signals to the topside. This process will take most of a day. The spare tether is currently unterminated, so to switch to that we’ll need to terminate it as well. If we want to record HD video on the vehicle, we can run with the current tether as-is.
Thursday, 18 June:
We removed the 1Cam HD Camera/Laser Scaler from the vehicle, to make it lighter and easier to work with while we’re checking out the vehicle. We’ll replace it once we re-terminate the tether.
The final patches to the ROV comm system were made: we added a patch cable between the tether and the E-bottle, and it seemed to improve the ability of the tether to withstand sharp bends at the ROV.
We did a test dive from ~10am-12, drove out to the strobe buoy, and saw that ship-relative navigation was OK, but that earth-relative navigation was off. We studied the sonar and GPS manuals for a while, and eventually discovered that the DGPS antenna had its serial (RS-232) message for True Heading turned off. We found the software application to re-program the DGPS unit, and used it to turn the True Heading message back on. Once this was done, the USBL display switched to a north-up display (as desired), and otherwise seemed OK.
Just prior to 4 pm we got the ROV into the water; preliminary navigation checks looked OK. We drove the ROV out to the strobe buoy, and calibrated the navigation solution by rotating the DGPS antenna as necessary, and then securely lashing the antenna in place. Position error was ~5m at the ~150m lateral distance from Nai’a. We considered this to be sufficient accuracy for the mission.
We ran the ROV until dinner time, exploring the terrain between the strobe buoy and Nai’a. This got us down to ~150m depth.
Friday, 19 June:
This morning was the first attempt at diving at the primary anomaly, at a depth of ~200m. Currents were strong and we had difficulty staying in position; when the ROV would be turned to oppose the current, there were frequent resets of the vehicle comm system. Eventually we had to manually pull the vehicle back to the boat as it did not restart properly after a reset.
With the vehicle on deck, we found that there was intermittent power when wiggling the tether at the main power supply connection. This was discovered by watching the power supply ammeters – they would dip to zero upon certain deflections of the wire harness; this would not be the case if it were a comm system/fiber-optic issue.
Since we needed to re-terminate the tether anyway to get HD camera data, it was decided to stop here and re-terminate the tether. We cut the end off of the main tether, and set up a workspace in Ron’s cabin so that he could work on the re-termination. For the rest of the day Ron worked on the tether. John and I worked on setting up the OpenROV that I had brought along. We can potentially use this to investigate the sand shelf at ~60-70 meters depth (the location of the “Fender” anomaly).
Saturday, 20 June:
Ron did a quick test with the mostly-reterminated tether (the Kellems grip still needed to be attached). The video camera of the ROV still did not initialize upon powerup, and there was still an issue with intermittent power draw. Further debugging of the situation showed that the waterproof bulkhead power connector on the E-bottle (Seacon/Subcon type) had an intermittent pin. The rest of the morning was spent replacing this bulkhead connector, for which we had a spare.
After lunch, we noted during tests that the system power is now stable, but that the vehicle still does not initialize properly- the camera tilt servo does not cycle upon power up. We spent the afternoon performing various tests and sending E-mails to Seabotix.
Sunday, 21 June:
In the morning we worked on various tests with the ROV. In the afternoon Ron continued these tests, removing the ROV E-bottle and opening it up to look for faults. John and I took my OpenROV out for a dive on the sandy shelf at 60-70m depth.
Monday, 22 June:
In the morning John and I did another dive with my OpenROV, this time with a GoPro attached to obtain higher-quality video. Ron continued ROV debugging during this time.
We started doing initial planning on how to use the ROV hardware to make a drop-camera assembly, in case we can’t get the ROV fixed. Tomorrow will be Monday in San Diego (HQ location for Seabotix), and it will be one final chance to get tech support from them, as all of the staff will be around. We are getting limited support today, and their guidance suggests that we have an issue with the main circuit board in the E-bottle. We have no spare for this board.
We took pictures and wrote descriptions of the various internal lights and indicators on the ROV electronics, and sent this to Seabotix so that they would have as much information as possible on their Monday morning.
Later in the day we started setting up the 1Cam as would be necessary if we were using it as a drop camera. It needs to wake up in HD mode, putting video data out its fiber-optic port, and with the laser scaler on. We were able to successfully reprogram it to do this.
Tuesday, 23 June:
Ron and I spent most of the morning on the satellite telephone with Seabotix, trying any last tests that the support staff there could come up with. We were not able to make any progress, and the Seabotix techs ran out of ideas for things to try.
After lunch we began the process of turning the ROV into a drop camera. We stripped off the thrusters, the float block, the main camera, and the tilt servo. We also stripped off any unnecessary brackets and wire harnesses. We strapped on the 1Cam HD camera/laser scaler, and mounted the USBL transponder on a right angle bracket on the aft portion of the assembly, so that it would be properly oriented when the camera was hanging downwards. We made a custom wiring harness for the external lights that I had on my OpenROV, and added them to the vehicle.
Wednesday, 24 June:
We performed final clean-ups and tests on the new drop-camera assembly that was derived from the ROV, and added weights to the front side of it.We flaked about 250m of cable onto the deck of Nai’a, and at 0900 performed a first test dive. The vehicle was too heavy for extended hand-holding of the tether, so we brought it back up and removed 6 lbs of weights.
At 0930 we began deploying the drop camera over the area of the principal anomaly at ~200m. The USBL positioning system worked down to about 175m (it is rated for 150m), and we were able to make contact with the ridgeline at about 200m and take video data. After a period of filming we pulled the camera up to ~175m to note its position, and saw that the camera was drifting well north of Nai’a’s position due to currents.
To pull the camera southward we passed the mid-section of the tether to a waiting skiff, which pulled the drop-camera southwards. On the Naia we monitored the position of the camera using the USBL system. After towing the camera about 2/3 of the way to the anomaly site, we lost the video feed from the camera. We retrieved the camera and took a break for lunch.
After lunch, tests of the tether showed that both fibers and all three copper conductors had parted. The failure was due to a disbond between the Kevlar braid and the tether outer jacket.
Ron started studying the prospect of re-terminating the tether yet again, and Walt did a dive with the dive team. The ship Fiji Princess arrived on site. The entire TIGHAR crew from Nai’a was invited to dinner by the captain of the Fiji Princess.
Thursday, 25 June (last day at Nikumaroro):
We spent the morning stowing the ROV equipment for departure: the DGPS antenna was taken down, the USBL dunking transducer was removed from the keel of Nai’a, and the ROV power supply and nav console were packed away. The whole area around where ROV operations took place was policed up to make it ready for going back to sea.
After lunch we made up a new drop camera, for a final try at photographing the anomaly at 200m. For this attempt, we reprogrammed the 1Cam HD Camera/Laser scaler to wake up in photo mode, taking a photo once every 5 seconds and storing the result in its internal memory card. The laser scaler would be on, and the internal flash would light the scene. Power for the camera came from two 9.6V battery packs that I cannibalized from my OpenROV, and wired together to provide 19.2V to the camera. We wrapped the assembly with rubber padding to protect it from coral hits, attached it to a long line, and added a ~5´ weighted line out the front of the camera to provide for a “feeler” to estimate distance from the reef surface.
Ron, John, Rob Barrel, and I took a skiff out to the anomaly site and spent ~40 minutes taking photos of the bottom, in the vicinity of the anomaly. Upon retrieving the drop camera, the battery tubes were crushed (they are only rated to 100m), and the camera was not powered. However, later analysis of the stored photos showed that the battery tubes failed on the way to the surface, so no photos were lost.
A “snail trail” of GPS values was recorded, showing the boat position as a function of time during the photo shoot. The photos taken by the drop camera are individually tagged with the time they were created. Post-shoot comparison of the GPS to the camera’s internal clock showed that the camera clock was 8 seconds fast compared to GPS time.
Upon returning from the anomaly, dinner was served, and after this we immediately left Nikumaroro for home.
Summary of ROV technical issues:
- Ground fault in 1Cam HD Camera/Scaling Lasers
- Ground fault in ROV port aft thruster
- Broken fiber in tether
- Intermittent pin in bulkhead power connector
- Failure of main ROV circuit card
- Complete failure of tether due to Kevlar/jacket disbond (during drop-camera operations)