16-Jun-2021

Attendees

  1. @Thomas Watteyne

  2. @Austin Patel

  3. @Kristofer PISTER

  4. @David Burnett

  5. @Fil Maksimovic

  6. @Martina Balbi

  7. @Said Alvarado-Marin

Action items

@Thomas Watteyne to write an nRF UART pass-through snippet
@Thomas Watteyne to write an VDDD tapping snippet
@Austin Patel to coordinate with @Andrew Fearing to get 60x boards assembled with a regulator and an IMU
@Thomas Watteyne to work on letter for @David Burnett
@Austin Patel to tell @David Burnett turn-around time and cost for new board manufacturing

Minutes

  • @Thomas Watteyne

    • Farnell order (60x XLs, 2x dev boards) should be received on Jun 17, 2021

    • Will hold onto chips, just let me know where we should send it

    • @Austin Patel will do!

  • @Said Alvarado-Marin

    • DotBot runs!

  • @Martina Balbi

    • BLE-based connectivity working

  • @Fil Maksimovic

    • LH2: working on calibration using a calibration board (4x LH2 receivers)

    • FPGA: thanks @David Burnett for point to https://github.com/PyFive-RISC-V/caravel_pyfive ! implements exactly what was needed, peripheral bus (Wishbone), DMA and HW acceleration

    • @David Burnett https://opencores.org/

    • @Kristofer PISTER great, as previous experience wasn’t fantastic with other approaches

  • @Kristofer PISTER

    • VDDD tap: would prefer not haven’t an external regulator, breaks the vision a bit

    • we would like SCuM itself to work as a true single chip

    • do we get different results depending on the test (tapping a wire vs. using a power supply)

    • 2 tests:

      • use a real wire

      • use the nRF:

        • pin to VDDD LDO out

        • high z, set 1, strobe from high z to active output, to high z, emulating tapping behavior

    • @David Burnett it that doesn’t work, we can use a voltage divider

    • discussion about previous tests

      • tapped at 1.1V: @aus

      • @Austin Patel

        • test 1 leave tap connected, record whether programming works, optical programming, on McCoy with the IMU

        • test 2: remove tap when seeing program running

      • constant 1.8V didn't work; a tap at 1.8V didn’t work (board wouldn’t program)

    • @Kristofer PISTER what did we do?

      • @Fil Maksimovic program, doesn’t execute, tap, it executes

    • @Kristofer PISTER looks like the only difference is the duration of the tap

    • @Thomas Watteyne we could use the nRF to vary (sweep) the duration of the tap

    • @Kristofer PISTER we need to get the nRF up and running, and try with the 3-wire bus

    • @Austin Patel probably the next step is to use the nRF.

      • thoughts: people integrating on micro-robots may only need cold boot

      • low confidence this will work

      • looking to having the board work

    • @Kristofer PISTER let’s

      • have @Andrew Fearing put a regulator to get things going

      • study this with the nRF by disconnecting the regulator

    • @Kristofer PISTER can we put a resistor divider on the board so voltage from nRF goes down to 1.1V

    • @Austin Patel another pin which has the voltage divider, do can choose between straight to nRF, or voltage divider

    • @Fil Maksimovic use a break-out board of a regulator?

    • @Austin Patel will ask @Andrew Fearing

    • @Thomas Watteyne to write a UART pass-through snippet

    • @Thomas Watteyne to write a strobe snippet

  • @David Burnett

    • writing proposal