You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Reported in #36 and by others, there seems to be more packet loss when connecting Crazyradio to a USB 3.0 port and communicating with many Crazyflies. The issue seems to be gone if the radio is connected to a USB 2.0 port or if a USB 2.0 hub is used.
Since this seems to only affect multi-crazyflie setup my assumption is that there is a problem with the control transfert: the difference between controlling one or many Crazyflie is that when controlling many the computer will make control transfer between packets to change channel or radio address.
I have also noticed that programming other board using DFU is very slow on USB3, programming in DFU is using control transfers as well. This would confirm my assumption that there is something fishy with full speed control transfers and XHCI (the USB3 controller). Though I could not find any information confirming this assumption on the Internet.
If there is any other report of problems with USB3 please add to this ticket with a description. I will try to reproduce the problem and see if avoiding the use of control transfers (for example by implementing #4) fixes the problem.
The text was updated successfully, but these errors were encountered:
Reported in #36 and by others, there seems to be more packet loss when connecting Crazyradio to a USB 3.0 port and communicating with many Crazyflies. The issue seems to be gone if the radio is connected to a USB 2.0 port or if a USB 2.0 hub is used.
Since this seems to only affect multi-crazyflie setup my assumption is that there is a problem with the control transfert: the difference between controlling one or many Crazyflie is that when controlling many the computer will make control transfer between packets to change channel or radio address.
I have also noticed that programming other board using DFU is very slow on USB3, programming in DFU is using control transfers as well. This would confirm my assumption that there is something fishy with full speed control transfers and XHCI (the USB3 controller). Though I could not find any information confirming this assumption on the Internet.
If there is any other report of problems with USB3 please add to this ticket with a description. I will try to reproduce the problem and see if avoiding the use of control transfers (for example by implementing #4) fixes the problem.
The text was updated successfully, but these errors were encountered: