Page 1 of 1

Light House Deck Damaged?

Posted: Tue Mar 24, 2020 7:52 am
by stupid_moron
I have 2x Light house deck with crazyflie 2.1.

Lighthouse deck 1 works perectly well every single time
Lighthouse deck 2 works ocassionally, (just power off and swapping decks only,SAME setup, same placenmet, same lighting condition etc)

When i run tools/lighthouse/get_bs_position.py, I get the error in the following lines:

rvec_start, tvec_start = calc_initial_estimate9sensor_sweeps) in ine 341
yaw = estimate_yaw_to_base_station(sensor_sweeps) in line 185
return direction[hash_sensor_order(sensor_order)] in line 146
KeyError; 120

but when i run exact same setup, with just different lighthosue deck, it gives the error. The deck is less then 1 month old. No visible damaged either...

Re: Light House Deck Damaged?

Posted: Wed Mar 25, 2020 9:51 am
by kristoffer
The best way to check if the deck is OK is probably to set up logging in the client for the angles of all sensors and check that they change when you move the deck around.

Log these:
lighthouse.angle0x
lighthouse.angle0x_1
lighthouse.angle0x_2
lighthouse.angle0x_3

Re: Light House Deck Damaged?

Posted: Thu Mar 26, 2020 3:12 am
by stupid_moron
I have ran the logging check on both of my decks

1) I move the CF orientation in 90Degree angle at the same spot
-The deck which is working perfectly; Every 90degree angle change results in around 0.01 value change
-The deck which i suspect is damaged: Every 90 degree angle change results in 3.4 value change(lighthouse.angle0x_1)

2) Leaving the deck position and orientation untouched, logging values taken every 30 secs apart
-The deck which is working perfectly; Aprox 0.01 value change
The deck which i suspect is damaged; Approx 2.6 value difference every logging script is ran

I also tried the following for both deck;
Logged the data via script, ran get_bs_position.py, ran a simple swarm hover at 0.5m
Results:
CF with perfect lighthosue deck; No issue
CF with suspect damaged deck; Flew uncontroabably

Somehow running the logging data script, followed by "get_bs_position.py" does not give me any hash errors. But leaving the CF position and orientation untouched, switching off and on again, running the sript "get_bs_pposition.py" returns an hash base station error.

I am not sure if the data values 1) and 2) confirms if the deck is damaged? I presume the values are not normal as there is a huge difference between the 2 decks: I do not see any visible damage to the TS4231 Sensors. Thus, probably not sensor issue. What are the chances the FGPA chip is faulty instead?

Note: The CF was left alone with clear line of sight with the base station for at least 10sec before every "get_bs_position.py". So i doubt it was an issue with CF unable to sync with lighthouse base station.

Re: Light House Deck Damaged?

Posted: Thu Mar 26, 2020 9:07 am
by kristoffer
It looks like something is not working as expected with one of your decks. Could you please contact us on support@bitcraze.io, we will handle it there.
Thanks!