> The newly released RP2350 microcontroller has a confirmed new bug in the current A2 stepping, affecting GPIO pull-down behavior.. Presumably there will be a Bx stepping at some point, but for now it is clear that the RP2350’s A2 stepping is probably best avoided.
Latest investigation shows it has nothing to do with internal pull-downs. Merely switching IO to input is enough to latch up. RP2350 doesnt work with high impedance sources.
According the Lady Ada from Adafruit relatively few will run into it, but if you are affected it can be worked around by adding an extra resistor for the affected pins into your circuit.
So it seems it's an annoyance rather than a killer. A revised chip would be nice though.
It'd be interesting to know what happened there. I would have expected the pad hard block on that mature of a process to be pretty hard to mess up.
Too bad hard blocks like that are almost certainly wrapped in so much NDA that we won't hear outside of of bars adjacent to conferences even if rpi wants to tell us.
https://hackaday.com/2024/09/04/the-worsening-raspberry-pi-r...
> The newly released RP2350 microcontroller has a confirmed new bug in the current A2 stepping, affecting GPIO pull-down behavior.. Presumably there will be a Bx stepping at some point, but for now it is clear that the RP2350’s A2 stepping is probably best avoided.
> pull-down
Latest investigation shows it has nothing to do with internal pull-downs. Merely switching IO to input is enough to latch up. RP2350 doesnt work with high impedance sources.
This (and the linked GitHub issue) almost look enough to kill adoption of the chip? Or is the issue less severe than it initially appears?
According the Lady Ada from Adafruit relatively few will run into it, but if you are affected it can be worked around by adding an extra resistor for the affected pins into your circuit.
So it seems it's an annoyance rather than a killer. A revised chip would be nice though.
did not even run into it on the defcon32 badge
Too bad it's delayed the new buspirate boards. For anyone looking to debug or reverse engineer various signaling protocols buspirate is an amazing device. http://dangerousprototypes.com/docs/Bus_Pirate#Introduction
It'd be interesting to know what happened there. I would have expected the pad hard block on that mature of a process to be pretty hard to mess up.
Too bad hard blocks like that are almost certainly wrapped in so much NDA that we won't hear outside of of bars adjacent to conferences even if rpi wants to tell us.
Looks like this might have been a side effect of seeking an alternate pad design with 5V tolerance...
Related: Erratum E9 on RP2350 https://www.hackster.io/news/problems-with-the-raspberry-pi-...
[flagged]