Hi thanks for the links, I checked pretty much all of them before asking. :)
It turns out that the device (ATSHA204A) has a bit more complex behaviour that I thought, in particular it aggressively goes into low-power/sleep mode, and had to be woken up before communication. When running the I2C device search, it has a side effect of waking the device up, while when I communicate manually, I need to send some zero bits (I think, but not totally clear) to wake it up. Now it works manually as well, though fortunately there's a library that communicate with this chip well, so after this debugging, I don't really need manual communication.
Hi, I'm using a Bus Pirate v4 that I just bought a few weeks ago. Wanted to do continuous ADC measurement in the HiZ mode (command D), and it works for a few seconds and then it hangs. Have to hard-reset the board to get it back.
I can duplicate the hang merely by connecting the GND and ADC lines together and running it for a bit.