Skip to content
Eliott65Sondewiki
Contents

Where the probe stands

In two days, the probe went from an idea to a tool that flashes Eliott65 over Wi-Fi. Here is what has been checked on the real board, and what so far only on the PC.

Checked on the board

ItemResult
Wi-Fi and namethe probe joins the network and answers to eliott65sonde.local; ports 4441 and 4442 open
SWDRP2040 identifier read (DPIDR 0x0BC12477), ROM read correctly, both cores examined
OpenOCD64 KiB read in 0.53 s at 4 MHz (120 KiB/s), identical to the read through the log
Flashing Eliott65v0.58.0 flashed over Wi-Fi at 200 kHz, verified, in 1 min 27 s
Local log10 s reading every 2 ms without error; 256 KiB of SRAM read in 4.7 s
Update over the network18 s end to end; several updates in a row succeeded
Return to the old firmwarea broken image replaced on its own by the old one in 52 s

Checked on the PC only

  • The trigger and the snapshot of the local log: tested against the simulated target. On the board, a first watch has been running since 16 October (Eliott65’s GRAPHE demo, read every second); its result is not yet known.
  • An image that starts but is never validated (3 times 2 minutes): covered by the boot loader’s bench and by the end-to-end tests on the PC, not tried on the Pico W.

What remains to be done or understood

  • Eliott65’s tools over Wi-Fi: done on the Eliott65 side since its v0.59 (one variable, E65_SONDE, sends all its tools through the probe). Their overhead compared with USB remains to be measured.
  • Burst errors on SWD: with the current wiring, long reads without pauses fail above 200 to 400 kHz. The cause is not established: probably the wiring, but an oscilloscope would be needed to prove it. See Tests on the board.
  • The boot loader can only be updated through BOOTSEL. It is small and almost never changes.

Timeline

VersionWhat it brings
0.1Portable CMSIS-DAP processor, simulated target, TCP server on the PC; OpenOCD connects to it
0.2Pico W firmware: SWD by PIO, Wi-Fi, server on port 4441, mDNS name
0.3PIO program tried in the simulator, against a bit-level simulated target
0.4Local log on port 4442; first test on the Pico W, then on the board
0.5Update over the network; SWD pins at 2 mA; fix for the restart after an update
0.6Boot loader: automatic return to the old firmware if the new one does not prove itself

The details of each step are in the repository’s development journal; this wiki takes up the essentials by theme.

Sources in the repository : README.md, docs/agile/BACKLOG.md, Eliott65 : docs/journal/15-retour-sur-la-carte.md, docs/journal/04-essai-reel.md, docs/journal/06-integrite-du-signal.md, docs/journal/07-amorce.md.