Day 2 of testing the automatic control modes (8 Feb 2008)
Martin made changes in the morning
	-  fixed a few software bugs
	-  reduced gains?
	-  zeroed integrator gains
P gain = 45
D gain = 230
I gain = 0

Sample rate = 4Hz (note:  we usually run at 5 Hz)


The first set of tests used the "step mode" input to issue commands to the ROV.  This mode was not working yesterday.

The goal of this first series was to observe behavior of the ROV with the new gains

Test1: 15:53:21
1 m forward

Test2:  15:54:09
1 m reverse

Test3:  15:55:00
2 m forward

Test4:  15:55:42
2 m reverse

Test5:  15:57:06
1 m to starboard

Test6:  aborted test (1m port)

Test7:  15:58:34
1 m port

Test8:  15:59:14
2 m starboard

Test9:  16:00:00
2 m port


The vehicle was pulled because oil was observed in the water. It was a shaft seal.  The housing had not been tightened.  The bolts were very loose.  Bolts were tightened, and the vehicle was put back in the water.

The next set of tests was an attempt to capture open-loop step responses.  The vehicle was put in auto-head, and autodepth (~2m).  Station keeping was off.  The pilot then put in "step" commands.  The results from these tests can be compared with the data being collected from the simulator (where similar tests were requested).

Test10:  17:06
28% forward thrust

Test11:  17:08:27
10% forward thrust

Test12:  17:09:33
35% forward thrust

Test 13:  17:10:34
66% forward thrust

Test14:  17:11:37
6% forward thrust --- the vehicle barely moved.  We think the command was less than the deadband.

The next two tests were lateral moves

Test15:  17:13:17
30% forward thrust

Test16:  17:14:05
50% forward thrust

The final two tests were rotations  I thought auto-head was off, but it looks like we were in closed heading loop= been chasing a command rather than looking at a torque input.  Cannot use these runs to identify an open loop model

Test 17: 17:16:59
rotate CW about 1/2 turn
Note:  Yesterday the vehicle was only able to rotate CCW

Test18:  17:17:30
rotate CW about 1/2 turn  