Hi, I built a bigger drone with BigQuad deck, however, one of the ESC is beeping.
Here is the log -
SYS: ----------------------------
SYS: Crazyflie 2.0 is up and running!
SYS: Build 34:71ebff013648 (2018.01.1-34) CLEAN
SYS: I am 0x333636353235510D002C0022 and I have 1024KB of flash!
CFGBLK: v1, verification [OK]
DECK_DRIVERS: Found 13 drivers
DECK_DRIVERS: VID:PID BC:05 (bcBigQuad)
DECK_DRIVERS: VID:PID 00:00 (bcRZR)
DECK_DRIVERS: VID:PID BC:01 (bcLedRing)
DECK_DRIVERS: VID:PID BC:04 (bcBuzzer)
DECK_DRIVERS: VID:PID BC:07 (bcGTGPS)
DECK_DRIVERS: VID:PID 00:00 (bcCPPM)
DECK_DRIVERS: VID:PID BC:08 (bcUSD)
DECK_DRIVERS: VID:PID BC:09 (bcZRanger)
DECK_DRIVERS: VID:PID BC:06 (bcDWM1000)
DECK_DRIVERS: VID:PID BC:0A (bcFlow)
DECK_DRIVERS: VID:PID BC:0B (bcOA)
DECK_DRIVERS: VID:PID BC:0C (bcMultiranger)
DECK_DRIVERS: VID:PID BC:FF (bcExpTest)
DECK_INFO: Found 2 deck memories.
DECK_INFO: Enumerating deck 0
DECK_INFO: Deck BC:05 bcBigQuad (Rev. C)
DECK_INFO: Used pin: 10011001
DECK_INFO: Driver implements: [ init test ]
DECK_INFO: Enumerating deck 1
DECK_INFO: Deck BC:0A bcFlow (Rev. E)
DECK_INFO: Used pin: 0000000C
DECK_INFO: Driver implements: [ init test ]
DECK_CORE: 2 deck enumerated
DECK_CORE: Calling INIT from driver bcBigQuad for deck 0
BIGQUAD: Switching to brushless.
DECK_CORE: Calling INIT from driver bcFlow for deck 1
Motion chip is: 0x
si pihc noitoM: 0xFF
MPU9250 I2C connection [O<F>
---
This happens on the M2 side of the BigQuad. I have tested the motors and the escs separately using another FC.
Any ideas?
BigQuad ESC Beeping
Re: BigQuad ESC Beeping
I think the problem is that the Flow and BQ-deck is not compatible without HW and software modifications. Have you tried removing the Flow deck?
One hack do get the Flow and BQ-deck working together:
1. Remove capacitor C1 on BQ-deck for SPI to work. The battery measurements can not be used then.
2. Change HW ChipSelect. Easiest I think would be to move it (cut trace and solder patch) to IO1 on the flow deck.
3. Change the CS pin in flowdeck.c (IO3 -> IO1)
One hack do get the Flow and BQ-deck working together:
1. Remove capacitor C1 on BQ-deck for SPI to work. The battery measurements can not be used then.
2. Change HW ChipSelect. Easiest I think would be to move it (cut trace and solder patch) to IO1 on the flow deck.
3. Change the CS pin in flowdeck.c (IO3 -> IO1)
Re: BigQuad ESC Beeping
So, I removed the flow deck for now and got a new big quad deck.
Now I am able to power the motors, however the drone flips during takeoff. The direction of the spin and the props match the wiki for big quad deck, still the drone flips.
Is it the PID? Any tips for calibrating pids for a biggquad Crazyflie (140mm) ? I have been trying to do it, still the drone flips and is really aggressive.
Will see how to get flow deck working with big quad.
Now I am able to power the motors, however the drone flips during takeoff. The direction of the spin and the props match the wiki for big quad deck, still the drone flips.
Is it the PID? Any tips for calibrating pids for a biggquad Crazyflie (140mm) ? I have been trying to do it, still the drone flips and is really aggressive.
Will see how to get flow deck working with big quad.
Re: BigQuad ESC Beeping
Often flipping is the setup or the sensors, most likely the setup. Double check that to start with. Maybe you oriented the BQ-deck the wrong way?
Normally the stock tuning works fine to start with as bigger quads have larger inertia. All the bigger quads we have build have worked with stock PIDs.
When I have build a bigger quad I start by testing it holding it in my hand above my head. Then it is pretty easy to feel if the stabilization works by applying some thrust.
Normally the stock tuning works fine to start with as bigger quads have larger inertia. All the bigger quads we have build have worked with stock PIDs.
When I have build a bigger quad I start by testing it holding it in my hand above my head. Then it is pretty easy to feel if the stabilization works by applying some thrust.
Re: BigQuad ESC Beeping
@shreks7 The flipping over is definitely your PID tuning! Mine did that too. Totally un-fliable off the bat. Build a cage, tie your build down and go to work calibrating the PID. You can adjust the settings in crazyflie-clients-python, and then when you find some that work, program it into the firmware. I am interested to hear how you accomplish the flow deck mod, because I would love to have that functionality on my big drone build so I can send it down the long hallway at work, skirting above the floor at 70 kph! (probably dangerous, be safe with your dangerous big quad) But lets get ya flying first!
It is a similar process to other drones. Start at zero (drone will just flop and rev motors) and work your way up until you hit limits and it oscillates, then derate by 10-50ish I've found. I can post a link I have open in my tabs at home but I am out of state right now. You'll find a result faster if you google this yourself.
I did not yet post my build because I blew something on the Crazyflie or my BQ deck (haven't made time to test) so I'm working on ground vehicles since I haven't found something as open and awesome as the CF for controlling ground vehicles.
@tobias, @shreks7
This post caught my attention because of the beeping. All four of mine did that, and it is because the crazyflie must be perfectly still when you turn it on to do some calculation for the accelerometers etc. If this step fails, there will be no output, or warning in the cfclient. Your ESCs will not run, and when mine don't receive a signal, they beep. I thought my crazyflie died when I first bought it but someone told me this! It bit me again with my ground vehicles and not getting UART Tx from evoggy's crazycar I am hacking apart. I keep meaning to post about this, it would be nice to see a readout of the failure or have the crazyflie loop and try to do it again a few times. I would be surprised if this happened and resulted in just one ESC not getting a signal and beeping though. Unless the other 3 think they have seen input and shut up but the crazyflie is actually not producing output. Definitely make sure your build has solid landing gear and you turn the CF on while on a stable surface.
@shreks7
I recommend getting an oscilloscope if you like embedded projects, it is an invaluable tool to have to see what you cannot see. You would be able to see the changing PWM signal to your ESCs to rule out anything happening with the crazyflie. I like my Rigol ds1054z, though I wish I had a logic analyzer (they make one with it built in but does not work with riglol I don't think). Might get a dedicated one, currently trying to use my bus pirate for this purpose.
It is a similar process to other drones. Start at zero (drone will just flop and rev motors) and work your way up until you hit limits and it oscillates, then derate by 10-50ish I've found. I can post a link I have open in my tabs at home but I am out of state right now. You'll find a result faster if you google this yourself.
I did not yet post my build because I blew something on the Crazyflie or my BQ deck (haven't made time to test) so I'm working on ground vehicles since I haven't found something as open and awesome as the CF for controlling ground vehicles.
@tobias, @shreks7
This post caught my attention because of the beeping. All four of mine did that, and it is because the crazyflie must be perfectly still when you turn it on to do some calculation for the accelerometers etc. If this step fails, there will be no output, or warning in the cfclient. Your ESCs will not run, and when mine don't receive a signal, they beep. I thought my crazyflie died when I first bought it but someone told me this! It bit me again with my ground vehicles and not getting UART Tx from evoggy's crazycar I am hacking apart. I keep meaning to post about this, it would be nice to see a readout of the failure or have the crazyflie loop and try to do it again a few times. I would be surprised if this happened and resulted in just one ESC not getting a signal and beeping though. Unless the other 3 think they have seen input and shut up but the crazyflie is actually not producing output. Definitely make sure your build has solid landing gear and you turn the CF on while on a stable surface.
@shreks7
I recommend getting an oscilloscope if you like embedded projects, it is an invaluable tool to have to see what you cannot see. You would be able to see the changing PWM signal to your ESCs to rule out anything happening with the crazyflie. I like my Rigol ds1054z, though I wish I had a logic analyzer (they make one with it built in but does not work with riglol I don't think). Might get a dedicated one, currently trying to use my bus pirate for this purpose.