Please enable JavaScript.

Coggle requires JavaScript to display documents.

Discussion, ADC, General, GPIO, Data Processing / Computation,…

- - - - Except they didn't inlcude the sources of error from the measurement equipment
- - - - We are measuring via a sensor value instead of a supplied
    - - Long cable
      - Multimeter, power supply and oscilloscope tolerances
      - Lack of differential signalling usage
- - - - We are measuring via a sensor value instead of a supplied
    - - Long cable
      - Multimeter, power supply and oscilloscope tolerances
      - Lack of differential signalling usage
- - - - Jonas & Sebastians suggestions
        
        Address gain errors from voltage divider tolerances through scaling factors.
        
        Use higher-quality operational amplifiers and components with lower tolerances
        
        Apply software calibration to correct offset errors.
        
        evt. læs Sebastians nye rapport
        
        Issues with Torque Pedal Voltage Buffers:
        
        Voltage buffers saturate due to low supply voltage (3.3 V).
        
        Adjusting resistor values, though this may reduce accuracy
        
        Increasing supply voltage to 12 V (requires different op amps).
        
        Current design places Sallen-Key filters after voltage dividers and near CMC connectors
        
        Placing filters before voltage division may reduce noise but eliminates the need for torque pedal buffers
        
        Testing may reveal whether placing filters closer to the Zynq improves performance by reducing noise into the XADC.
        
        Mikkels kommentar: disse fejlkilder ligger på PCB'et, og brug af differential signaling mm. vil ikke fjerne disse.
        
        the more accurate, the finer the adjustment of acceleration, safety & datalogging. Max margin of error, according to Jonas & Sebastian ~2.5% --> 1/40
        
        the only thing we can change, and see if it affects the accuracy, is the usage of differential signals, since we don’t want to mess around with hardware.
      - Make xADC inputs P & N true differential instead of just N = GND.
- - - - o wire two pairs of 2 wires same distance near it simultaneously  (one differential pair, and one single-ended pair) o Generate an input voltage (either sinusoidal or constant) to measure at one end. Wire the two pairs through the noise zone and out to the ADC.  DC: Make it a differential signal by switching the terminals of the power supply  AC: o measure the difference in accuracy after the ADC output (input vs. Output)
      - o Use the car with the motors running at different (varying speeds) while measuring in a single go. o Make sure the wires are placed in approximately the same area of the car.
- - - - see Testing Ideas or Result & Product Ideas document
- - - - in general: have pins available for extra SDC nodes
- - - - ○ Watermark-interrupts & HPB interrupts
        ■ Don’t interrupt on all incoming messages
  - - - Snooping on the CAN bus
      - Normal mode
      - Why can't we ?
        
         Goes into a reset
         Error & event interrupts
        
        Noise?
        
        transceivers seemed to do what they should
        
        Maybe we shouldn't use a CAN bus that is managed by the X90 PLC? --> setup a new CAN bus between dashboard and Zynq Based MC
        
        ○ Jonas & Sebastians values don’t output anything
      - Jonas & Sebastian also didn't succeed
    - - Showcase the robustness of CAN in the car future work
  - - - does it not really have an impact?
  - - - only major thing is diffrerence in speed