10/4/2018, 12:00 PDT. Deployed system from Shana Rae with TreFoil doors open. During subsequent inspection dive the doors were closed. The scale factor was set to 0.6 due to the calm seas. The Gain schedule mode still needs to be turned on, previous deployments indicate a starting gain is 20 and an appropriate target for the StdDev is 1000 (the default). Through the afternoon the network connection was dropped once or twice, this is being worked on. Rich currently pings shore before issuing a Trefoil controller watchdog update, so an outage longer than an hour will cause the doors to open. Ultimately we should implement the doors open/closed based on streaming data, eliminating the reliance on the link to shore to properly operate the doors. 10/5/2018, 13:24 PDT: Overnight the cell link dropped between about 8PM and midnight, then came back. Looks like the beaglebone may have rebooted in that time, and anyway it did reboot for sure about 6AM. The network was down long enough that the Trefoil watchdog timed out and opened the doors, as expected. This morning we turned on the camera and that worked nicely. We did notice however that sometimes, turning on the camera stops the AHRS. Cycling power on the AHRS via the WebRelay starts it back up. The Camera and AHRS are both on the same 24V supply, switched independently by the web relay. 10/5/2018, 13:30 PDT: I turned on the power converter gain scheduler by issuing a "gain 20" command, this should adjust the scale factor to achieve a RPM standard deviation of 1000. Note: With the doors open as they are now, it'll drop the scale factor all the way to drive the motions up, as soon as the door closes the motion will increase immediately but the gain scheduler will take hours to respond. So that time-scaling may be an issue, it may be neccessary to actively set the scale to a larger number before closing the doors... 10/5/2018, 16:00 PDT: We opened the doors for the weekend. 10/8/2018, 10:15PDT: We closed the doors and will start the auto open-close software in "open only" mode. 11/8/2018, 9:40PST: We replaced the red switch caps today with new ones that have an o-ring seal in the hopes of eliminating the ground fault issue. The procedure to do this is - Open doors with a "tf_SetPos 0" command(in this case they were already open) - Stop the power converter from providing current to the battery by issuing "bus off", this makes the current in the switches much less for the moment when they are turned off. - Turn off switches, replace caps, turn switches back on. - Turn power converter back on by issuing a "bus auto" commmand. - Turn gain scheduling back on on the power converter by issuing a "gain 20" command. - Close the doors with a "tf_SetPos 1" command 11/29/2018. 3:45. - Ran the pump for two 5 minute periods. Will watch means to see change since we're not sure how long it needs to run. Hourly average before pump was run was 33.8 inches. Average after pump was run was 20.8". So about 1.3in/min. -