update with altimeter, swimfile.txt endleg.txt, etc see changes_13sep.txt also reset_PI()
Dependencies: mbed MODSERIAL FATFileSystem
new_changelog.txt
- Committer:
- joel_ssc
- Date:
- 2019-05-09
- Revision:
- 91:916ee0fea9ff
- Parent:
- 89:0b8291dbf340
- Child:
- 96:0048e13ca2c3
File content as of revision 91:916ee0fea9ff:
Modified FSG PCB V_1_1 - Freezes when doing a dive or any timed sequence (commented out SD card references) - commented out sdLogger().appendLogFile(current_state, 0); //open SD file once - commented out sdLogger().appendLogFile(current_state, 1); //writing data - commented out sdLogger().appendLogFile(current_state, 0); //close log file - reduced timer to 20 seconds for bench testing - modified ConfigFileIO for rudder() - added in getFloatUserInput function from newer code - changed LinearActuator & batt() in StaticDefs to match new pinouts from Bob/Matt/Troy fixes - slowed down battery motor because it's silly fast at 30 volts (bench test) * BCE gain is proportional 0.1 now 0.01 - BATT was moving in the wrong direction, gain was P: 0.10, I: 0.0, D: 0.0 * change gain to P: -0.10 (gain was flipped, I think the old circuit board had the voltages flipped ? ?) - StateMachine changes for testing * added keyboard_menu_STREAM_STATUS(); * added keyboard_menu_RUDDER_SERVO_settings(); - modified the zero on the battery position from 610 to 836 - BMM (batt) slope may be incorrect, seems low, currently 0.12176 - modified the zero on BCE from 253 to 460 - Pressure readings are wrong * added readADCCounts() to omegaPX209 class to see channel readings * modified omegaPX209 class to use filtered ADC readings from SpiADC.readCh4() - fixed rudderLoop to headingLoop from newer code Modified FSG PCB V_1_2 - added init headingLoop to main - added pitch and heading outputs to STREAM_STATUS NOTE: Flipped motor controller output on connector side with battery mass mover (BMM) - Motor direction was opposite the BCE motor (because of gearing) - BMM P gain is now positive 0.02 (from -0.10) Modified FSG PCB V_1_3 - added timing code for interrupt that drives the rudder (testing with o-scope) - PID controller replaced with newer version from 5/29 code branch - StateMachine hanged style of variables to match convention in code Modified FSG PCB V_1_4 - adc tests Modified FSG PCB V_1_5 - IMU update - Testing print outs Modified FSG PCB V_1_6 - new BMM zero count of 240 (confirmed manually) - new BCE zero count of 400 (confirmed manually) - Modified emergency climb to go to position 10 on the BMM, not zero (almost relying on the limit switch) - fixed keyboard input; including typing in the timeout (can enter exact times) Modified FSG PCB V_1_7 - removed data logger references - fixed bug where I was logging data twice in the interrupt code and the main file - fixed bug where Multi-Dive sequence wasn't restarting (the counter used to get the current state was not reset) Modified FSG PCB v_1_7C - fixing a bug with the data transmission (output had newline character) because of packet lengths - new version blasts data and Python program requests lost packets, dramatically faster over USB, slightly faster over XBee - fixed some format and terminal errors Modified code 2018-08-06 - fixed bug with heading save feature, wasn't updating deadband properly - modified input on servo sub-menu Modified code 2018-08-07 - Float broadcast changed to command rudder to center in that state (whatever center is in the file) - Was sending the wrong command to the position dive (BMM) Modified code 2018-08-14 - Fixed the deadband on "case Float_Broadcast" so it exits properly, I was seeing the BCE stop around 317 so I made a fixed deadband of 4.0 (anything 316 or above will exit the state) - FIND_NEUTRAL was code was double-checked: 1) Starts with an offset of 25 mm from the manually found neutral BCE position (still buoyant), BMM starts at that neutral position 2) case NEUTRAL_SINKING: Every 5 seconds it retracts the BCE 5 mm. When it reaches the input depth setpoint it switches to NEUTRAL_RISING 3) case NEUTRAL_SLOWLY_RISE: Every 5 seconds it extends the BCE 2 mm. When it starts to rise (depth rate is negative) (depth rate positive is sinking, negative is rising) it will immediately stop and go to case NEUTRAL_CHECK_PITCH. Removed an incorrect line of code that was setting BCE position with depthloop output. 4) case NEUTRAL_CHECK_PITCH: Every 10 seconds the pitch is adjusted +/- 0.5 mm depending on whether the nose is pitched up or pitched down (the timing is because the PV may move around in the water before it stabilizes) - Homing Sequence now moves at -0.5 not -0.1 (was way too slow) Modified code at Carderock (2018-08-15) - Removed -1 multiplier from IMU outputs - Pitch is incorrectly oriented now, pitch down - Heading/Yaw is off by 90 degrees I believe (north is not north) - Did not check roll - KNOWN ISSUE of log file not having consistent line length (will throw off python data receiver) Modified code at Carderock (2018-08-16) - Observations at LASR lead to fixes on the save PID files, zero offsets were saving incorrectly, fixed 2019 feb 15 Modified code jcw - LogOO?.csv name and version number is set with logvers.txt - DIAG00?.txt likewise set with logvers.txt - diagnostic logging to diagnostics file - legfile.txt sets parameters ( max-depth, min_depth, heading, timeout, yo_time) for a dive leg, after which mbed program ends to allow monitor computer to access via usb filesystem - added states: Leg_position_dive, leg_position_rise, fb_exit to handle above yo-yo leg and exit have not yet added these states in menu operations. STILL TO DO: - test size of log file and set new log file name and update logvers.txt - allow for something other than 0=leg in legfile.txt so that state will NOT drop into legfile.txt operations mode - add access to leg_position_dive ( with exit??) to menu operations -maybe 25 feb 2019 DONE: jcw - added state ENDLEG_WAIT - will drop to keyboard/FLOAT_BROADCAST if char is read over xbee() while waiting - test time for log file to be cycled - in main.cpp - convert from feet to meters in depth().getposition() calls ( deep inside outerloop() ) 07may2019 DONE as of now jcw - fixed version of posvelfilt that can deal with heading variables - state FLYING_IDLE that is to be ended by a pulldown on a DigitalIn line - presently has a timeout for testing - state START_SWIM that is started automatically at end of FLYING_IDLE and splashdown wait - Find_Neutral will run immediately if splashdown ends with FSG upside down, but START_SWIM will exit the mbed program and return control to Raspberry PI for communications. I have not yet exercised code that would run if the FSG submerges and does not self -right under water. It should happen soon. I think I do that on the bench. - FIND_NEUTRAL will be run if flag and file showing neutral found has not been set - this is important in my view. if find Neutral fails, file is written to warn raspberry PI - lots of flashing leds() to indicate still running. - runs on bare board. still to be tested with FSG full system, or even pieces - such as imu().