Skip to content
Ben Dyer edited this page Sep 4, 2015 · 12 revisions

Release notes

Updated 2015-09-04

Hardware revision 3, firmware commit 63433cb (2015-09-04)

Changes — firmware

  • Removed PWM control option.
  • Added support for audio via the UAVCAN BeepCommand, including while the motor is running.
  • Added UAVCAN bootloader.
  • Added open-loop startup mode to improve compatibility with a wider variety of motors.
  • Added automatic Kv identification, so most motors require no configuration at all.
  • Added thrust controller, so if propeller/rotor parameters are available the ESC can compensate for changes in wind/inflow velocity automatically.
  • Added power and thrust estimation in extended ESC status messages.
  • Added rotate-while-armed support.
  • Transitioned UI to use only UAVCAN messages.
  • Added support for PX4ESC 1.6 hardware.

Known issues — firmware

  • [FIXED] Configuration of motors with unknown pole counts is difficult.
  • [FIXED] Very low speed operation is difficult to achieve with the speed controller; this would require a position controller to get right.
  • [FIXED] UAVCAN support has not been tested extensively. In particular, it's possible that busy CAN buses will increase processing latency to an unacceptable degree (hardware filtering is not yet implemented).
  • [FIXED] For flash capacity reasons, the ESC uses a custom UAVCAN implementation. It would be better to re-use the libuavcan [de-]serialization code with custom framing/scheduling.

Known issues — hardware

  • [FIXED] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit f15ac82 (2015-05-06)

Known issues — firmware

  • [OPEN] Configuration of motors with unknown pole counts is difficult.
  • [NEW,FIXED] Max speed parameter is not respected in speed or torque control modes.
  • [FIXED] Start-up in torque control mode is difficult to tune for some motors.
  • [FIXED] In torque control mode, the initial torque setpoint needs to be set high enough to start the motor. For compatibility with standard PWM ESCs, it's possible that we should increase torque to whatever is required to get the motor moving, and then transition to the torque setpoint once running.

Known issues — hardware

  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit 1f3191f (2015-05-04)

Changes — firmware

  • PWM control parameters default to more reasonable values.

Known issues — firmware

  • [OPEN] Start-up in torque control mode is difficult to tune for some motors.
  • [OPEN] Configuration of motors with unknown pole counts is difficult.
  • [NEW,FIXED] Parameters are not updated when flashing firmware via CAN.
  • [NEW,FIXED] Empty CAN messages are sent every millisecond.
  • [NEW,FIXED] Motor Kv conversion factor is wrong for motors with low pole counts.
  • [NEW,FIXED] PWM control mode and control curve parameters can't be set via the UI.

Known issues — hardware

  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.
  • [FIXED] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [FIXED] UAVCAN to UAVCAN cables are too fragile.

Hardware revision 3, firmware commit 4a1d232 (2015-03-28)

Known issues — firmware

  • [FIXED] Start-up in reverse is not reliable.

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit ac3db3a (2015-03-27)

Known issues — firmware

  • [OPEN] Start-up in reverse is not reliable.
  • [NEW,FIXED] Some motors have fewer than 4 poles, below the minimum for the motor_num_poles parameter.

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit 0354930 (2015-03-26)

Known issues — firmware

  • [OPEN] Start-up in reverse is not reliable.
  • [NEW,FIXED] UAVCAN RawCommand values above 4095 result in negative rotation.
  • [NEW,FIXED] UAVCAN EscStatus messages are not given IDs compliant with the spec.
  • [NEW,FIXED] PWM input throttle curve and min/max setpoints are out of order.

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit 681bfcb (2015-03-23)

Known issues — firmware

  • [OPEN] Start-up in reverse is not reliable.
  • [FIXED] Unloaded motors running at maximum speed sometimes fail to spin down when the "Stop Motor" button is clicked. To spin down a motor in this state, change the setpoint mode to "torque", enter "-100", and click "Start Motor".

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit 18df8c7 (2015-03-22)

Known issues — firmware

  • [NEW] Unloaded motors running at maximum speed sometimes fail to spin down when the "Stop Motor" button is clicked. To spin down a motor in this state, change the setpoint mode to "torque", enter "-100", and click "Start Motor".
  • [OPEN] Start-up in reverse is not reliable.
  • [NEW,FIXED] User interface displays a completely blank screen when no ESC is found.
  • [FIXED] Rotor angle estimation can be unreliable when using high-Kv motors unloaded at high speeds. This can cause ESC damage.

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit c09a957 (2015-03-21)

Known issues — firmware

  • [NEW] Rotor angle estimation can be unreliable when using high-Kv motors unloaded at high speeds. This can cause ESC damage.
  • [OPEN] Start-up in reverse is not reliable.

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit 14a58d5 (2015-03-20)

Known issues — firmware

  • [OPEN] Start-up in reverse is not reliable.
  • [FIXED] Start-up with high speed setpoints and high-inertia loads is not bulletproof. Maximum torque during start-up should be limited, possibly requiring another configuration parameter.
  • [FIXED] The PWM input may require additional configuration options to be useful, e.g. minimum and maximum pulse width, and mapping between pulse width and torque or speed setpoints.

Known issues — hardware

  • [OPEN] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [OPEN] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit 4f04458 (2015-03-18)

Known issues — firmware

  • [OPEN] Start-up in reverse is not reliable.
  • [OPEN] Start-up with high speed setpoints and high-inertia loads is not bulletproof. Maximum torque during start-up should be limited, possibly requiring another configuration parameter.
  • [NEW,FIXED] Multi-frame UAVCAN transfers fail due to data type signature checksum mismatches.
  • [NEW,FIXED] Node ID updates made via the configuration UI can't be saved.
  • [FIXED] Transition between high-frequency injection angle estimate and EKF-only estimate is rough. This occurs at half the motor's Kv, i.e. 450 rpm for a motor with Kv = 900 rpm/V.
  • [FIXED] EKF can lose sync when stopped from high speeds due to a sudden increase in load torque. This results in vibration and a speed estimate above the HFI speed threshold. Sync is generally restored when load torque is decreased.
  • [FIXED] In some cases, on startup motor moves backwards for up to 90° before switching to forward motion.
  • [FIXED] HFI can yield sub-optimal results for high-impedance motors (especially low-Kv pancakes). This generally involves rough start-up, but once the HFI speed threshold is exceeded the motor runs normally.

Known issues — hardware

  • [NEW] The Fischl USBtin adapters we shipped on or prior to the 18th March do not have CAN termination enabled. Depending on your network configuration, this may cause unreliable communications. If you're not able to connect to the ESC(s), try shorting the two large vias closest to the ground (black) wire on the USBtin PCB to enable termination. If this solves the problem, solder a link between those vias.
  • [NEW] UAVCAN to UAVCAN cables are too fragile.
  • [OPEN] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.

Hardware revision 3, firmware commit aa124d5 (2015-03-17)

Known issues — firmware

  • [NEW] Transition between high-frequency injection angle estimate and EKF-only estimate is rough. This occurs at half the motor's Kv, i.e. 450 rpm for a motor with Kv = 900 rpm/V.
  • [NEW] EKF can lose sync when stopped from high speeds due to a sudden increase in load torque. This results in vibration and a speed estimate above the HFI speed threshold. Sync is generally restored when load torque is decreased.
  • [NEW] Start-up in reverse is not reliable.
  • [NEW] In some cases, on startup motor moves backwards for up to 90° before switching to forward motion.
  • [NEW] HFI can yield sub-optimal results for high-impedance motors (especially low-Kv pancakes). This generally involves rough start-up, but once the HFI speed threshold is exceeded the motor runs normally.

Known issues — hardware

  • [NEW] No thermal solution available; if running above 16 V, 20 A ensure adequate airflow is available and keep an eye on temperature. The suggested limit with no heatsink is 27 V, 20 A but thermal issues may be encountered below this limit depending on airflow and ambient temperature.
Clone this wiki locally