2024-09-28 Pegasus 2 Tuning/Range/Speed FT

 

Requesting team: Directors

Items TBD:

  • Test Procedures/Test Cards

Sections for flightline team to fill out:

  • Incident Procedures, communication preferences.

  • Test Cards

Admin Preparation

Requested By

@Nathan Green

Sub-Team Review

(To be checked once reviewed by sub-team representative)

Mechanical
Electrical
EFS
Autonomy
Operations

Date of Request

Sep 7, 2024

Goal Summary

Successfully fly Pegasus 2 without major incidents

Validate that all sensors and peripherals are working correctly

Evaluate tune and potentially retune

Status?

In PROGRESS

Desired Airframe

Pegasus 2

Desired Date(s)

September 28th 2024

Testplan - to be filled out by requesting team

 

 

 

 

 

https://ardupilot.github.io/MethodicConfigurator/TUNING_GUIDE_ArduCopter

Test 1 - Further Tuning

Procedure

Goals / Objectives

Knockoff criteria

  1. Ensure raw IMU data is being logged

 

 

  1. Set mode to Loiter, arm, hover ~10m above ground for 30 seconds

 

 

  1. Ensure no errors or failsafes

 

 

  1. Full stick deflection in each axis

 

 

  1. Land

 

 

  1. Evaluate filters using online tool, adjust if needed, repeat test

 

 

  1. If filters are good and were not good originally, repeat autotune on each axis

high (> 4.5) value for ATC_ANG_RLL_P, ATC_ANG_PIT_P, and ATC_ANG_YAW_P

 

  1. Disable raw IMU data logging

 

 

Test 1 - Range Test

Procedure

Goals / Objectives

knock-off criteria

  1. Check failsafe behaviour, auto speed, RTL altitude

 

 

  1. Set mode to loiter

 

 

  1. Arm, takeoff to 50m

 

  • Any EKF issues

  1. Use guided mode to set a waypoint on the other side of WRESTRC

 

  • Any EKF issues

  1. While the drone is flying, keep an eye on the link quality, tx power, and number of dropped packets

  • No connection loss, no dropped packets

  • Drone loses connection, will RTL

  • Any EKF issues

  1. Once the drone reaches the waypoint, enable RTL

  • No connection loss, no dropped packets

  • Any EKF issues

  1. Drone lands

 

 

Test 2 - Speed Test

Procedure

Goals / Objectives

knock-off criteria

  1. Set mode to loiter

  • Determine maximum flight speed, efficiency at each speed from log analysis

 

  1. Takeoff, climb to 50m

 

  • Any EKF issues

  1. Set a guided waypoint ~400m from the current location and fly there

 

  • Any EKF issues

  1. Increase the speed parameter by 2 m/s and fly another ~400m stretch

 

  • Any EKF issues

  1. Continue until speed maxes out, check max tilt value and increase if reasonable, then increase speed again

 

  • Any EKF issues

  1. If drone cannot reach max speed parameter after an increase, the old value can be said to be “max speed”

 

  • Any EKF issues

  1. Land

 

 

 

 

 

Necessary items:

  • Pegasus 2

  • Controller

  • Video receiver monitor

  • Video ground station

  • Batteries

 

 

 

 

Flight characteristics needed:

Comms / Support needed:

Attendees

Name

Phone #

Sub-team

Role

Name

Phone #

Sub-team

Role

@Nathan Green

 

 

  • Pilot

@Smile Khatri

 

 

  • Mech lead tingz

@Evan Janakievski

 

 

  • Mech

@Balaji Leninrajan

 

 

  • Auto

@Ben Lovegrove

 

 

 

@Camilo Artigas Alos

 

 

  • Mech

@Jane Zeng

 

 

  • Auto

@Daniel Puratich

 

 

 

@Victor Sun

 

 

  • Mech

@Vyomm Khanna

 

 

 


Flightline Team

Name

Phone #

Role

Reason

Name

Phone #

Role

Reason

Nathan Green

 

Pilot

Fly the drone

Daniel Puratich

 

 

 

 

 

 

 

Pre-Flight Preparation

Checklists

  • Software configuration correct

  • Compare against latest baselines and verify no params changed, or changes approved by comp-ft-team

  • LTE tested

  • Autopilot Orientation checked

  • Motors tested

  • Prop direction check

  • Sensors outputs healthy & within range

  • Controller input mappings correct

  • Controller input choices correct for pilot preferences

  • Wireless trainer tested & working per pilot preferences.

Hardware (wiring)

  • Wires are not too tight or too loose (comfortable, not strained, not flapping around)

    • No cut/exposed wires

    • Shrouding is neat + presentable

    • Solder joints look good

      • no cracks

      • no corrossion

      • no exposed contacts (unless expected)

    • Specific requirement from @Daniel Puratich

      • no exposed pad without solder on ESC caps

        • i did this on peggy 2

        • someone will need to do with peggy 1

      • ESCs all have a touch of hot glue to secure cap to board

    • Connectors fully plugged in

    • Connectors in good health

Hardware (airframe)

  • Bolts secure (nothing loose)

    • Carbon fiber not cracked

    • 3D prints not failing / crushed / cracked

    • No loose items

Hardware (groundstation)

  • ELRS Gemini tested & Working

    • Logging correct

    • Data looks correct

    • Control tower tested & working

  • Tracking antenna x2 (control + video) (mechanical)

    • Servos secure

    • Hardware (antennas) secure

    • No cracks/bends/chips in polycarb/aluminum

    • Gears / shafts look healthy

    • Grease if necessary

  • Tracking antenna x2 (control + video) (electrical)

    • Electrical components secure

    • Power delivery / signal wires secure, no cracks/corrosion/damage

  • Video System

    • goggles(?)

    • Monitors(?)

    • Video tower tested & working

  • Laptop w/mission planner configured correctly.

    • LTE Link tested working

Flightline Kit

  • Spare motors

  • Spare props

  • Spare tools

    • Metric allen keys

    • Socket set

      Pliers (flatnose + needle nose)

  • USB-C cable (short & long)

  • Soldering iron + heat shrink + spare cables

  • Power supplies for ground station computer

  • Ground station computer

  • Spare bolts

  • Radio Controllers

  • Mission plan clear

  • flightline roles assigned

  • Communication hierarchy established

  • Incident procedures & roles in-place & briefed to all flightline team members

  • Timeline for day-of clear

  • Battery charging plan, equipment management plan clear.

  • Transportation clear

  • Batteries in safety bag

  • Equipment

    • Zip tie

    • duck tape

    • label maker

    • Type c cable

    • ELRS Gemini

    • Saleae Logic Analyzer

    • charged WARG laptop

    • spare bag for small electronics

    • controller box

      • two controller

      • battery tester

    • VTX box

  • execute

Total 5 main system checked:

  • Power system(battery and connection)

  • Propulsion system(motor and propeller)

  • Controller system(Control links and ground station system)

  • Sensory system(sensors)

  • Video system (VTX)


  • Debrief

  • clean / store equipment if necessary

Incident Procedures

Flight Test Timeline

Related to 2024-09-28 Autonomy Obstacle avoidance FT

Date/Time

Action

Notes

 

Date/Time

Action

Notes

 

10:00

Meet in bay

  • Tried to leave at 10

    • batteries were still charging

      • daniel woke up late

        • delegate someone to charge batteries

    • autonomy system was not wired

 

10:30

Leave for WRESTRC

 

 

11:00

Flights started

 

 

12:20

Finished

 

 

12:40

Bay Arrival

 

 


Flight Test Debrief report

Card 1 Test 1: Tuning

What Happened

  • Analyzed the initial tune, determined it was too aggressive

    • did not feel very stable in flight

  • Checked notch filter frequencies

    • took us a bit to setup parameters correctly

    • compared to in-flight FFT

    • Settled on using in-flight FFT as best outcome

  • Reset tune to default using initial tune config

  • Ran autotune again on all axes with aggressiveness of 0.07 (default is 0.1, range 0.05 to 0.1)

  • Evaluated new tune, new tune is felt less aggressive.

    • Needs more testing and log analysis

  • Ran all flights except final autotune with detailed IMU logging

Log Analysis

  • Stability

    • How stable was the drone with the new tune?

    • How good was the autotune solution according to the parameters?

    • We couldn't really fly it much because of the rain, but overshoots of a few degrees on all axies show up in the logs.

      • It felt much better than before to the pilot, but still doesn’t seem ideal when looking at the logs

      • Yaw looked good, but roll and pitch aren’t that great

    • Paths forward

      • run pitch and roll atune again → recommended

      • lower atune aggressiveness and run the full sequence again → maybe we get a better solution

      • learn to tune manually from this autotune solution → probably the best if we have skill or can learn

  • Vibration

    • Seemed okay in the flight

    • some clipping in VIBE reading above 30 m/s^2 but no problems with EKF, good most of the time

  • Packet Loss on Mavlink RC

    • Did the paccket

    • Have to check on mission planner laptop

    • zero damaged packet count

    • 100% link quality in mission planner

Action Items

  • Next week spend more time analyzing the tune and retune if needed.

  • Checking packet loss on laptop

Card 2 Test 1: Range Test

  • NOT ATTEMPTED

  • It started to rain and we did not want to attempt this.

Card 2 Test 2: Speed Test

  • NOT ATTEMPTED