Tricopter/Tricopter2016 Change Log (I didn't continue this all the way).txt at master · steflayanto/Tricopter · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
2/21/2016
New Frame

2/23/2016
Updated Prop Vectors
Updated ConvertOutput Section
ISSUE Function may cause values over 100% to be written to motor.
Examined Values in Program
Added New tab to tricopter excel file for program information

2/25/2016
Set all PID output limits to -50 until 50
Updated writeServo Command to new servo values
Set heading PID output to convert to -40% to 40% range (percentage of total motor thrust) so that at hovering motor thrust (~80%) servo angle will not exceed 30deg
Set Calibration for radio inputs (pitch, roll, heading only), set to calibrate upon completion of gyro calibration
Set Servo to 110 - 180 degrees with 145 as midpoint, rather than use writeMicroseconds();
Inversed Heading PID controller to align with controller.

2/26/2016
Connected everything and ran program. significant amounts of servo wobble.
Added landing gear
converted servo control back from degrees to microseconds
Added Battery mount with ziptie
Rough test without proper setup, balancing point foudn to be around 2200-2250. Adjusted servo dowel shorter
Main problem is servo wiggle. considering MAF.

3/1/2016
Tried connecting servo to BEC instead of arduino 5v to reduce noise. ESC/Servo did not initialize at all
Changed thrust, pitch setpoint, roll setpoint, heading setpoint to integers to reduce noise. Seems alot more stable
Modified backprop funciton such that vertical component of back prop must be greater than horisontal component
Updated servos specs onto excel doc
Programmed all ESCs to LiPo, and soft acceleration start ups
Built frame for heading PID tuning by taping flat board to ice skating spinner
limiter set at 75.

3/5/2016
Plan to convert floats to ints
Converted floats to doubles
Tuned PID to approx 0.2
Set heading trim to 50 (on servo range of 1600+trim-2400+trim)
Observed, able to yaw clockwise easily, but anticlockwise is slow.
Yaw is held, despite oscillating. Values of PID are 0.2, 0.2, 0.01 respectively
Small, low frequency oscillations observed even when no visible yaw present. suspected cause is gyroscope noise, rather than any float-double-int conversions. also possibly caused by transmitter input noise.
transmitter input noise ruled out. attempted testing with setpoint permanently set to 0, oscillations still evident.
changed Gyro MAF length from 5 to 10

Added rollRatePID
set roll set rollsetpoint to 0.0 permanently
reversed rollPID controller direction

3/11/16
Set roll setpoint to stick input
set rollrate PID setpoint to roll output. set roll output to roll setpoint.
reorganized PID program to compute roll and pitch first, then compute rollRate and pitchRate
Tested while holding, appears to be a success.
Set sample time to 50 instead of 100
Setting PITCH pid to identical settings to test.
test flight. BEC not functioning. plugged Vin to LiPo battery charger socket. tricopter unresponsive despite motors being initialized. copter crashed.

3/12/16
Realized that arduino requires more than 5V for external power. second guessing need for BEC
Ran test flight with copter still connected to comp. Back lifts off too fast. front is too weak/heavy.
After umbilical tests, determined that copter always tilts right and forward even with maximum stick correction.
When printing center of thrust values, found that values never exceed much more than 10.
Findings:
1. Set trims for centre of thrust (since copter is asymmetrical)
2. Perform static tuning for roll/pitch (after trying to set trims)
After adding trim, values still tend to stay near 10.0
Trim affects motor startup. Trimming back and left props made them start after right one.
Programmed convertOutput tab to factor in CoT coordinates ONLY if total thrust is over 25
Noted that lift-off value is approimately 60-70% total thrust
trims noted to be approx 5 for X, and 10 for Y.
SMALL SUCCESS!!!!! First technical lift-off. all three legs off the ground for a second, despite obvious instability.
Tried to replicate feat, but battery died before I could do so.
Noted that battery can come loose and shift forward/backward.

3/13/2016
Ran test, backprop went insane and went on full throttle. problem found to be heading PID.
Changed left stick to be throttle and roll, and right stick to be pitch and yaw.
Ran test flight. mixed success. Hard to control.
Inversed heading setpoint.
switched back yaw - roll swap

18/03/2016
Setup apparatus for PID tuning of tricopter
disabled yaw for roll trimming.
disabled computation of yaw PID

09/04/2016
Secured Arduino to frame with two screws

11/04/2016
Decided to try replace Center of Thrust idea with accurate mathematics
"Perhaps it isnt the fine tuning of PID values that is preventing the copter from flying"
For Stability, need to investigate ALL factors, not just PID values, such as:
 - Delay between esc.write & motor RPM response
 - Delay between detection of angular change & actual angular change
 - Delay caused by process of calculation of PID response
 - Date update caused by length of gyro moving average filter
 - Sampling time of PID input
 - interval time of PID output
 - Length of arms of tricopter
 - center of gravity
 - numerical approximation errors in storage type conversion (float-int etc.)
realized I have been caught up in trying to fine-tune the variable when in fact the problem could lie in a great many other weaknesses.