2023-11-04 Flight Tests

Owned by Daniel Puratich
Last updated: 2023-11-26
12 min read

Introduction

  • Pegasus Sensor Testing Flight

    • Test rangefinder and optical flow sensor on monster mount are functional

    • a step toward all hardware end to end Pegasus test

  • Pegasus ESC Telemetry and High Frequency Logging

    • Test ESC Telemetry

    • Log high frequency flight logs for FFT and notch filtering

    • This should put us on track for a fair comparison of four vs six batteries for an upcoming flight test

  • Houston Autonomy Camera Payload Flight

    • Fly camera over landing pad to gather more data for CV

    • required for model training progression

Flight Test was requested on Oct 31st as a part of AEAC runup plan by @Daniel Puratich .

Resources

List contains all possible human resources available for flight test support! Please add yourself to this if you want to go to the flight test and follow the format!

  • @Daniel Puratich

    • Role (what you will do at the flight test):

      • Flight Test Lead

      • Flight Test Director

      • Safety

      • Documentation Director

      • Electrical

    • Justification (why you’re qualified)

      • TC Basic

      • US HAM

      • Exec Director

      • Flight test Requestor

    • Availability (times on this date you can be at flight test)

      • All day

      • here 8am

  • @Georgia Vachon Westerlund

    • Role

      • Photography

      • Ops support

      • Mech support

    • Justification

      • Ops Lead

      • Mech Member

    • Availability

      • All day

  • @Emma Chan

    • Role

      • Ops support

      • Photography!!

    • Justification

      • Ops Member

      • EFS Member

    • Availability

      • All day

  • @Yuchen Lin

    • Role

      • flight test support

    • Justification

      • EFS member

    • Availability

      • all day

  • @Nathan Green

    • Role

      • Pilot

      • SDC Driver

      • Mechanical

      • Systems integration

    • Justification

      • Basic certificate

      • Mech lead

    • Availability

      • not filled out …

  • @Hardy Yu

    • Role

      • Flight Test Coordinator

    • Availability

      • All day

      • here 9:45am

  • @Mihir Gupta

    • Role

      • Autonomy Representative

    • Justification

      • Autonomy Lead

    • Availability

      • not filled out …

      • here 9am

  • Jacky

    • Autonomy

    • Please fill this out as others have above!

  • @Neel Patel

    • did not fill out in advance …

  • SDC Rav4

    • Booked by @Daniel Puratich 11/2/2023 for @Nathan Green to drive

  • SDC Dodge Caravan

    • unavailable, booked by Teertstra

  • @Georgia Vachon Westerlund

    • has a vehicle but would prefer not to use it

  • @Neel Patel

    • minivan, 6 people total

WRESTRC Booked, see Calendar , for Pegasus flight .

Worst case we do https://maps.app.goo.gl/XBr7u8oXZkMhzEAC8 as subject to weather constraints regarding timeline for Pegasus.

Logistic

  • Trip 1: Houston Autonomy Data Collection

    • Members:

      • GSO @Yuchen Lin FTC

      • @Hardy Yu [Flight Test Coord]

      • Auto Lead @Mihir Gupta [Autonomy Lead]

      • Driver @Neel Patel with minivan

      • Jacky

    • Car: Neel’s minivan

    • Time: We gather @ 10am, the flight test will run 10:45 - 12:00

      • departed 10:46am from bay

    • Location: WRESTRC

  • Trip 2: Pegasus Sensor Validation

    • Member:

      • FTL@Daniel Puratich [Exec Director]

      • Driver/Pilot @Nathan Green [Mech Lead]

      • Photo @Georgia Westerlund [Ops Lead]

      • Op member @Emma Chan

      • one extra spot

    • Car: RAV4 - Nathan drive

    • Time: Gather in bay @ 12:30pm, flight test will run 1:15 - 2:30

    • Location: https://maps.app.goo.gl/XBr7u8oXZkMhzEAC8

Procedure

Optical flow and rangefinder wiring @Mena Azab completed day before
Monster mount mounting @Smile Khatri completed days before
Confirm Raspberry Pi can collect images with the $200 CV Camera
Test the imaging repository with the Raspberry Pi and confirm images are taken and stored on the SD card @Mihir Gupta doing morning of. Has been done in the past, so it should work this time, will not be tested, following discussion at 10:09am
Edit the rc.local and ensure that image collection code runs on Raspberry Pi startup, completed by @Mihir Gupta yesterday
Raspberry Pi can be powered onboard the drone, conceptually all setup, confirmed by @Daniel Puratich 9:20am
Raspberry Pi and $200 CV Camera are mounted onto Houston. Confirmed mounted by @Daniel Puratich
Flight test location booked, WRESTRC booked a while ago by @Yuchen Lin
Houston → See Discord here
Pegasus → flying on project Vanguard registration confirmed @Daniel Puratich 9:30am
Ensure we have good props for Pegasus after last tests crash - we have plenty @Daniel Puratich
Ensure we have good props for Houston, we need to order more as confirmed by @Conall Kingshott here. The props we have no are slightly damaged but should be okay for these flights, we need to get the process going for ordering more, soft assigned to @Hardy Yu here.
@Hardy Yu switching a single chipped prop on Houston because we have one prop that works as cited by @Yuchen Lin in discord here at 10am. Confirmed done by hardy 10:16am, there is still one chipped prop, we’ll deal with it later
Telemetry Setup
@Yuchen Lin is uninstalling the antenna mount to add in the SMA connectors and the mounting it back in, 10:05am
@Nathan Green connect to the two that are labelled main and aux, do not connect to the one labelled GNSS. ensured connectors fit, but they’re kind thin
drone powered on by @Nathan Green 10:23am
Verify functionality, @Nathan Green 10:24am
@Nolan Haines Schematic for and delegating manufacturing of anti-spark connector configuration. completed by @Neel Patel a week ago, solder joints weren’t inspected by anyone else before heatshrink which may be a concern.
@Daniel Puratich to update template with all the new things he added to this template this time. template seems good, im running this checklist morning of and setup a week in advance when requested!
@Nathan Green add to Onedrive the logs from last flight test. completed but low frequency logging to not FFTable
Someone allocated for photographs → @Georgia Vachon Westerlund & @Emma Chan allocated
@Conall Kingshott potentially addressing slop on landing gear. This is not a short term task and will not be done for this flight test, nor is it needed for future flight tests. Marking complete for the time being and isn’t a big deal.
Pilots scheduled - @Hardy Yu , will be @Nathan Green & @Yuchen Lin
Test Plan - @Nathan Green for Pegasus, verified with discussion with @Daniel Puratich
Test Plan - @Mihir Gupta for autonomy, verified with discussion with @Daniel Puratich
Verify we have wrestrc card - confirmed 9:35am in holder @Daniel Puratich
Systems test - @Daniel Puratich running the checklist
Check flight controller orientation - @Nathan Green unchanged, confirmed working 11:51am
Optical flow not working @Nathan Green 10:29am
Check motor spin & rotation direction - @Nathan Green
Check DSHOT configuration (DS300, bidir?) - changed to bidir all params kept 11:52
Motors plugged into correct pins on pixhawk
Verify motors 1,2,3,4 in motor test go in the correct order
Verify motors in a,b,c,d configuration is correct order
verify rotational response of motors is correct (with more than just audio), move the drone around and check each motors reponse individually - skipping this
Check transmitter mappings
Verify communication links
VTX/RX - no video on any of these aircrafts so not a concern
Controller
Groundstation - LTE works
Ammeter working - @Nathan Green works 11:53am
Verify data logging - @Nathan Green checked SD card, its working, will check - we dont have SD card in bay yet will check later
Transportation logistics finalized
Booking SDC vehicles - @Hardy Yu to decide, @Daniel Puratich
Assigning drivers - @Daniel Puratich Done see above.
Renting any needed equipment (trailer? generator? etc.) - @Daniel Puratich not needed.
@Hardy Yu forgot to grab the Rav4 car key in advance, seeing if we an get into machine shop. confirmed we got it 10:16am by Hardy
Setting failsafes for Pegasus @Nathan Green
normal failsafe set to land
disabled battery failsafe, dont want it to randomly land
GCS failsafe set to always land after 5 seconds of no GCS
unchanged from last time confirmed @Nathan Green 10:50am
Initial frame parameters set - @Nathan Green same as last time, no worries, 10:50am
Houston battery bal charged started 9:20 am @Daniel Puratich
Confirm Huston battery failsafe issue is fixed, @Yuchen Lin waiting for autonomy to be done with Houston as of 10:04am.
@Yuchen Lin Set to do nothing on batt failsafe.
@Nathan Green is helping out in this configuration of the new power module, reporting voltage and current 10:38am
@Yuchen Lin Pi Houston EE voltage check, confirm 5 V with DMM, confirmed 10:12am
Confirm the Raspberry Pi is securely attached to the drone prior to flight, confirmed by @Mihir Gupta 10:33am
Confirm the image collection software runs when the Raspberry is powered on, confirmed by @Mihir Gupta 10:33am
@Mihir Gupta Testing cam link, started 10:14am
@Nathan Green configuring the GPS so both GPS’s are used, still good from last time 10:55
@Nathan Green Setting the controller profile on the radio master
SDC Vehicle key collected → machine shop is closed, we will not be able to grab the key @Hardy Yu , should open at 10am, got it before 10:30am
Batteries Planned with respect to LiPo Battery Safety and Procedure @Daniel Puratich running the checklist
Drone batteries - confirmed we have them … idk what this item is @Daniel Puratich
Need to charge 4 Pegasus 6S batteries, started 10:10am roughly by @Hardy Yu
Confirmed 4 Pegasus 6S batteries charged (in progress 10:56am)
4 Houston 3S batteries put on charger at 9am by @Daniel Puratich
Confirmed 4 Houston fully charged by @Hardy Yu 10am
Transmitter confirmed charged to 70% at least. 8 V confirmed on controller, its a 2S so 4 V per cell is basically full 10:39am packed @Yuchen Lin
@Hardy Yu plugging in the transmitter, been a while since we’ve charged it, plugged in at 9:54am, wasnt actually charging bc usbpd cable
@Nathan Green @Yuchen Lin @Hardy Yu we learned transmitter cant be plugged into a USBCPD cable must just be a pure power connection, started actually charging at 10:25am
Battery tester/indicator, we have this in the control links box confirmed at 10am @Daniel Puratich
do not charge overnight - confirmed not done by @Daniel Puratich
On-site kit put together @Daniel Puratich running checklist
First Aid Kit
Ground station if applicable
Laptop - charging it half charged
Antenna / Receivers
Transmitter
Extra Water - non
Power cords / Extension cables - none needed
Wrenches - none needed
Allen keys (metric ones) - three sets - only have one set
Knives
Tape
Rope/string - not needed
Zip ties
Velcro -no
Safety goggles - not needed
Fire extinguisher - we don't have one, they’re out of our budget gg
Spare foam - not needed
hot glue gun w/ hot glue stick - not bringing that
harnesses - no spares
ratchet set
battery removed for velcro (thin 3d printed thing) for Pegasus
Wrestrc Key in possession of driver - hardy returning
Airframe assembled
ESCs/flight controller wired
Radio equipment tested
Channel mappings - confirmed workin
Flight modes - setup @Nathan Green 1:02pm
Motors tested - already checked
Check DSHOT configuration (DS300, bidir?)
Motors plugged into correct pins on pixhawk
Verify motors 1,2,3,4 in motor test go in the correct order
Verify motors in a,b,c,d configuration is correct order
verify rotational response of motors is correct (with more than just audio), move the drone around and check each motors response individually
Wiring secured (Harnessing) - all zip tied down to frame, @Daniel Puratich 11:05am
@Daniel Puratich secured monster mount sensor cables with zip ties 11:03am
End-to-end test (neglecting props)
Check the center of gravity - CG is off quite a bit but because only running two batteries and much less hardware then anticipated it will be reasonable for this test - unchanged, doesnt matter
All components not on the airframe that are required should be in the flight box (e.g. props & prop nuts)
Airframe payload - Raspberry Pi + $200 Camera
Mount to airframe pointing downwards (exact angle doesn’t really matter). Camera confirmed mounted by @Daniel Puratich 9:56am
Spare parts
Props
Houston props - we need to order more
Peggy props - in the big box
Airframe components
wings nuts - n/a for quads
prop nuts - in prop boxes
any detached components such as angle mount - none needed
Hardware to assemble these components - n/a
Spare Electronics - we dont have any of this so NA-ing
Ground station Software fully updated, firmware on 4.4.2 on pixhawk
@Nathan Green confirm op flow working on ground, confirmed 10:38
@Nathan Green confirm rangefinder working on the ground, requires being plugged and inplugged after the drone to make it work, but confirmed working 10:54am - there is a delay feature that fully works, configured 11:24am !
@Nathan Green confirmed new antennas for LTE are working from RPi LEDs 10:21am
@Nathan Green ESC telemetry confirmed
as of 11:27am not working, we couldnt get it to work last year
trying bidirectional dshot (reflashing pixhawk) as of 11:31am
bidirectional dshot isnt filterred but working for rpm (current voltage and temp do not work) 11:46am . data only sends when its running.
Take the WREST RC entrance card - must do!! not needed for peggy, back after
Only attach props directly before flight (keep off at all other times) For houston we’re keeping them on I guess - ez
Get snacks - na, short test
Mission Briefing - by @Hardy Yu
Read the flight plan to the team (that was made a week ago)
Purpose of flight
Plan of action
Success criteria
Delegate and brief on flight line roles
Safety Briefing
Key flight characteristics to observe
Key system characteristics to observe
Measurements to be made
“Knock off” criteria (anyone can call)
Specify Pegasus Flight area planned, hard land and disarm boundaries set. Ensure visual observers are aware of this and can visually verify boundaries. - @Hardy Yu
Air Frame
Check all antennas, ensuring they are secure and in good condition
Check the battery emplacement and secure attachment and ensure there are no cracks
Check that all lights are operating normally
Check that all sensors are secured and clear of any obstructions
Check that the GPS is receiving satellites and providing a navigation solution
Inspect top and bottom of air frame arms for cracks, loose parts, or signs of damage
Landing Gear
Check landing gear is secure
Inspect skids or wheels, especially attachment points
Motors
ensure all motors free-spin correctly
Ensure wires secured and not at risk of shorting
Ensure motor mounts secure.
Motor test with props
Propellers / Rotors
Ensure propeller is secured
Check for nicks, chips or cracks
Ensure prop direction is correct
Ensure prop mounting is secure
gold rings to props
Batteries
Inspect to ensure there is no signs of swelling, external leaking or other defects
Battery wiring and connectors from the battery are connected securely
Batteries (and spares) are adequately charged
Wires are not pinched
Control Station / Receivers / Transmitter
The battery and spare batteries if required are adequately charged
Control Station device and cables are properly connected and functioning
All flight interfaces are functioning normally
Safety
Good visibility and safe wind speed
Surrounding is free of obstacles
Crew far from drone
General Site Survey
Are there other people on the property
Ground Recording / drone footage on
Set the parameters for high frequency logging before takeoff!
Check optical flow health is working still in flight
Check rangefinder output is working still in flight
Hover in loiter for 1 minute to measure vibration with IMU batch sampler
Land.
Upload Logs
Verify that SD telemetry works
Remove Props
@Hardy Yu running the checklists allocated 10:45am during briefing
@Yuchen Lin running the Ground station allocated 10:45am during briefing
Air Frame
Check all antennas, ensuring they are secure and in good condition
Check the battery emplacement and secure attachment and ensure there are no cracks
Check that all lights are operating normally
Check that all sensors are secured and clear of any obstructions
Check that the GPS is receiving satellites and providing a navigation solution
Inspect top and bottom of air frame arms for cracks, loose parts, or signs of damage
Landing Gear
Check landing gear is secure
Inspect skids or wheels, especially attachment points
Motors
ensure all motors free-spin correctly
Ensure wires secured and not at risk of shorting
Ensure motor mounts secure.
Propellers / Rotors
Ensure propeller is secured
Check for nicks, chips or cracks
Ensure prop direction is correct
Ensure prop mounting is secure
Batteries
Inspect to ensure there is no signs of swelling, external leaking or other defects
Battery wiring and connectors from the battery are connected securely
Batteries (and spares) are adequately charged
Wires are not pinched
Control Station / Receivers / Transmitter
The battery and spare batteries if required are adequately charged
Control Station device and cables are properly connected and functioning
All flight interfaces are functioning normally
Safety
Good visibility and safe wind speed
Surrounding is free of obstacles
Crew far from drone
General Site Survey
Are there other people on the property
Ground Recording / drone footage on
Autonomy Camera Mounted
Raspberry Pi is powered on the drone
Autonomy image collection code runs on RPi startup
Autonomy Tasks
Connect to Raspberry Pi + camera using Camlink
Verify camera is working (image is displaying)
Set up landing pad #2 on pavement
Ensure camera focus is set to 30m
Ensure 2 landing pad (blue side facing up) are positioned on the ground
Images should be taken on grass and tarmac
move pads if we have time between flight tests
Takeoff
Hover over landing pad #1 at height of ~30m to allow camera to take photos
Hover over landing pad #2 at height of ~30m to allow camera to take photos
Land within ten minutes
Battery Swap if we have more batteries
potentially move pads
repeat
Upload Logs
Remove Props
Dismount Raspberry Pi + camera
Check SD card for saved image
drive back to bay
post mortem
Analytics added to this document
action items delegated
Autonomy Tasks
Dismount Raspberry Pi + camera
Check RPi for saved images
Make note of the best camera settings.
batteries discharge is delegated -
batteries at storage voltage confirmed
Upload logs, @Nathan Green
Confirm wrestrc card is back in the holder

If things go poorly, flight test director will command someone to do specific tasks on this last (this is stated to avoid bystander effect):

  • Fire

    • Call 911

    • give location

  • Property Damage

    • if in e7, email Graeme

    • if at WRESTRC, email them

  • Fly away

    • call TC

  • Injury

    • first aid

    • call 911

  • For more details see Operating Procedures

 

Debrief

After the flight test summarize the timeline of what happened with timestamps if timestamps aren't given in checklist. I’ve moved to trying to mark dates and all info in the checklist itself so this section is fairly minimal

  • Flew all four Houston batteries for data collection

  • Ran no filters but data logging, op flow and rangefinder worked flew for a bit (log didn't work here)

  • turned on ESCTELEM 3 notch filter flow for a bit

  • upped max speed and flew at loiter near max (limited by mode/stability) 14m/s max

  • 14m/s at the fast run

  • 30 A max

  • 12 A hover

@Daniel Puratich Directing this section of flight test post mortem!

  • The drone flies for 3 auto missions

  • We collected usable data for the Autonomy ML team

  • Process of collecting data and running Raspberry Pi was very streamlined

  • No crashes

  • didnt have to recalibrate compass at all

  • FFT looks good? Will let @Anthony Luo analyze

@Daniel Puratich Directing this section of flight test post mortem!

  • We should be planning drivers and cars better, need to ping people and ask in advance

    • this is a big one, we shouldn't get surprised by anyone who has a car bc it affects things a ot

  • Accidently hit the arm button without double check and flew to 100 meters altitude

  • WRESTRC a lot of birds in the air, if drone exceeds 100 meters altitude could risk the chance of getting hit with a bird

  • Better time management so we can fly more auto missions

  • Better idea on how many auto missions we want to run to collect data

  • More predetermined waypoints at flight test

  • The first log we recorded on Pegasus didn't save, other two did just fine, not sure why

@Daniel Puratich Directing this section of flight test post mortem!

  • None.

All of these major issues should be addressed via Root Cause Analysis Forms:

@Daniel Puratich Directing this section of flight test post mortem!

Ordering new propellers for Houston (discussed in above checklist & Discord) @Hardy Yu Should be done before next flight Houston flight which will likely be next week for more photos for autonomy
Logs to be uploaded and deciding if notch filtering is necessary @Nathan Green @Anthony Luo should be discussed in discord before next Pegasus flight test
Creating next flight test confluence @Daniel Puratich gg ez Confluence monkey grind. Next flight test will be 4 vs 6 batteries, make it heavy and run autotune. @Nathan Green unavailable next few weeks.
Order a new USB splitter @Mihir Gupta “the one we have now sucks”
@Nathan Green uploading the three logs