assisi / casu-pic Goto Github PK

We should do extensive tests on fft measurements and based on the results we obtain, implement additional filters.

Reading of the temperature sensors fails if the SPI2 clock is set to a value higher than 1.25 MHz. This value is ok for reading temp sensors, but for reading an accelerometer, which is on the same bus, it is recommended to use clock frequency at least 2 MHz (for accelerometer sample rate > 1.6 KHz). The solution is to change the SPI2 clock frequency depending on whether the temp sensors or accelerometer is read.

Files interrupts.h/c which we should use to implement all interrupt routines. Definitions for interrupt priorities have already been added. Eventually, this change should make code more readable.

The goal is to have homogenous temperature responses of all CASUs.

• implement MRAC?

Instead of (only) triggering FFT with a 1s timer, it should (also) be triggered with a new vibration reference. This is intended for vibration pattern accuracy analysis, but the regular 1s functionality should still be available for constant vibration references.

The code should then be running as follows: do the fft with 1Hz frequency. If a new setpoint arrives or a timer counts 1s, start new data aquisition. In case of setpoint-triggered fft, discard and cancel the ongoing process if existing/neccessary. Start the timer in both cases.

Add a routine to detect faulty temperature sensors. Currently, this is done only during the initialization, but it should be done after each reading.

Add external interrupt as a trigger source of the DMA responsible for sending data to/from SPI2.

assisi/assisipy#61

Allow beaglebone firmware to put ir proxy sensors in standby state via i2c

Required for generating accurate vibration patterns.

@dmirkovic Please provide an estimate of when this could be implemented.

@kgrip @marselap is this the code currently deployed on the CASUs. If so, we should tag a release. If not, we should push the current version and then tag a release.

configurations.xml different for linux/win. We should keep both versions.

Add a file where all interrupt routines will be placed.

It has happened that idle CASUs (no BBG connection over I2C) would overheat (even above 40C). Working theory is that the PWM contrlling the Peltier heater was left active by mistake. Implement additional safety features in the firmware.

Add median filter and PT1 filter for each temperature measurement.