lucysrausch/hoverboard-firmware-hack

New Board, CPU crash

emente opened this issue · 1 comments

Hi Niklas,

i just got a brand new board from china, and they changed something. Major differences:

  • A 36V lead (2 wire, 0,1m2, 10cm) directly soldered to the mainboard (2 pin JST other side) with no obvious connection target
  • Sensor boards uses a 20 pin QFP chip labeled 744H343 as IMU and has an actually mounted cap after the 7805 :-O
  • Only one sense resistor (3 mOhm) per side
  • Cables for right side unsual short, left side unsual long
  • Came with a 2-wire, 3-pin jst charger cable instead of 4-wire, 4-pin (not the 21V 4-pin screw connector cables)
  • MOSFET wires cut too short so that reflow on cpu side is poor

That mainboard behaves extremely strange. The GD32 crashes/halts after a random (yeah, damn) time interval between 0-20 seconds. It happily outputs serial data on the sensor port until it crashes. It doesn't seem to react to speed/steer commands (tested with my unchanged ros node).

Have you already seen a board like i described? I'm currently unsure if i got a board with strange hw problems or the newest cheap-o-reversion that may have changed "something".

Cheers from Karlsruhe,
mnt

Oh, wrong CPU. Nevermind.