search

kauailabs / navxmxp

Robotics Navigation Sensors. NavX-MXP designed for the National Instruments RoboRIO & FIRST FRC Robotics. Navx-MIcro designed for Android-based FIRST FTC Robots
MIT License
105 stars 87 forks source link

large delay in SPI, and conflicting comment about it #63

Closed PeterMitrano closed 6 years ago

PeterMitrano commented 6 years ago

https://github.com/kauailabs/navxmxp/blob/341c96f20816cc97a34341798f4f26c4cdb7ca3e/roborio/c%2B%2B/navx_frc_cpp/src/RegisterIOSPI.cpp#L50-L51

My first question is why is there a 1ms delay here? May I remove/shorten this delay or at least update the comment? 1ms is a very long time to be blocking in SPI world.

kauailabs commented 6 years ago

Since the SPI protocol is implemented on the navX-MXP STM32 F4 microconrtroller in an interrupt service routine, a small delay (worst case of 200 microseconds) is necessary; the only higher priority interrupt on the micro controller is the systick interrupt, and the SPI interrupt will preempt lower priority interrupts. If a method to achieve a 200 us delay is available on the roborio, the existing delay could be decreased to that. The practical minimum delay is not fully characterized, which was what the comment was meant to indicate. At the highest rate (200Hz) the existing implementation is not known does not miss any samples, so it’s not clear what decreasing the delay will achieve although lower latency is one possibility.

PeterMitrano commented 6 years ago

Ok, seems legit. I'm not actually using SPI I just wanted to bring it up in case it was unnecessary or a bug.