Getting Crazyflie 2.1 to work out of the box

Discussions about quadcopters/multi-rotors
kimberly
Bitcraze
Posts: 1050
Joined: Fri Jul 06, 2018 11:13 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by kimberly »

Also, it might be good to also point you towards the Bitcraze toolbelt: https://github.com/bitcraze/toolbelt , which is a wrapper for all kinds of containers that you can also use for building the firmware (we use this for our CI and testing ourselves). The crazyflie installation instructions also have some instructions for this as well: https://www.bitcraze.io/documentation/r ... #compiling. With that you can truly eliminate any differences in your OS or compiler version.

As the toolbelt does not have any access to USB, you would still need to do make cload or use cfloader to flash the bin file:

Code: Select all

cfloader flash cf2.bin stm32-fw -w radio://0/40/2M/E7E7E7E703
sean112
Beginner
Posts: 16
Joined: Thu Nov 14, 2019 1:51 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by sean112 »

Last time I updated firmware with only the flow deck attached, and when I put the multiranger deck on, the state estimate became unstable again. Then I updated the firmware using cfclient with both flow deck and multiranger deck attached. After the firmware update, the state estimate is still unstable.
kristoffer
Bitcraze
Posts: 630
Joined: Tue Jun 30, 2015 7:47 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by kristoffer »

It should not make any difference flashing the Crazyflie with or without the flow deck and multiranger deck mounted.
sean112
Beginner
Posts: 16
Joined: Thu Nov 14, 2019 1:51 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by sean112 »

Oh I see. I thought there might be a difference, since at the end of the flash it showed detecting deck to update firmware. I thought having the deck detached, the firmware is not updating the corresponding deck driver.

In that case, compilation should not have been the issue. Is there another reason why updating the firmware would cause flow deck and multi-ranger deck not working together? More specifically, I believe the issue is state estimation instability caused by all range measurements not able to update (static measurement value during bootup).
kimberly
Bitcraze
Posts: 1050
Joined: Fri Jul 06, 2018 11:13 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by kimberly »

Hi!

I actually think it is the other way around, that the range measurements with the flowdeck and multirangers doesn't update and then the state estimation is not able to converge.

Ah sorry, I saw now that in a previous thread, you did mention that you flew without the multiranger and with only the flowdeck with a cfclient release.
If the flowdeck works without the mulitiranger, then there might be something going on with the deck connectors. Do you have perhaps an multimeter to check the connections from the decks to the crazyflie?

If that is not something that you are able to do, or if you don't detect any broken connection with the decks, the next best thing is to send an new multiranger to you to try out, so please send us an email at contact_at_bitcraze.io for that.
kristoffer
Bitcraze
Posts: 630
Joined: Tue Jun 30, 2015 7:47 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by kristoffer »

Oh I see. I thought there might be a difference, since at the end of the flash it showed detecting deck to update firmware. I thought having the deck detached, the firmware is not updating the corresponding deck driver.
That is a reasonable assumption and it is sort of correct :-) Decks that have some sort of firmware running in the deck it self are updated during this phase, but the Flow and Multi ranger do not (currently only the Lighthouse deck and soon the AI-deck).
More specifically, I believe the issue is state estimation instability caused by all range measurements not able to update (static measurement value during bootup).
The multi ranger is not used for state estimation, so that should not change the behavior. The distance sensors on the multi ranger and flow deck are running on the same same bus though, and the boot sequence is a bit special, there might be something here if the firmware is not OK possibly?

Could it be a poor connection in the expansion bus? Your problems sounds a bit random which could be pointing in this direction.

If you have access to other Crazyflies/decks it could be useful to switch them around a bit to see if you can find a pattern.

There is also a version of the push demo that runs in the PC, you could try it out with the 2021.06 firmware release (that we know is OK). You can find the push demo here: https://github.com/bitcraze/crazyflie-l ... ultiranger or here https://github.com/bitcraze/crazyflie-l ... 6/examples in the latest release.
sean112
Beginner
Posts: 16
Joined: Thu Nov 14, 2019 1:51 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by sean112 »

I checked the connection between soldered connection and the pins for the multiranger deck with a multimeter. All of them seemed fine.

I was able to run the push demo on both cf2 stably, until I update the firmware to 2021.06. Afterwards I updated the firmware on both cf2, the state estimate went unstable.

I will send over an email to see if a new multiranger would solve this problem.
sean112
Beginner
Posts: 16
Joined: Thu Nov 14, 2019 1:51 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by sean112 »

So I tried to run the flow deck and multiranger deck separately with 2021.06 firmware from cfclient. Both sensors work fine by themselves. With only flow deck attached, cf's state estimation is stable with zranger outputting continuous correct data. With only multiranger deck attached, cf's state estimation is stable (z reports -68 for some reason, but stable), all ranger data are continuous and correct (except for the left ranger.) Only when I have both decks attached at the same time, cf's state estimation goes unstable, and all ranger data are not changing.

So I am not sure whether this is a hardware problem.
kristoffer
Bitcraze
Posts: 630
Joined: Tue Jun 30, 2015 7:47 am

Re: Getting Crazyflie 2.1 to work out of the box

Post by kristoffer »

It sounds like a possible hardware problem, let's continue the discussion in the email.
Thanks for your patience!
Post Reply