Conversation

Notices

1. Embed this notice
   F4GRX Sébastien (f4grx@chaos.social)'s status on Friday, 27-Oct-2023 11:15:33 JST F4GRX Sébastien

   in reply to

   • Mike Knell
   • Pierre Bourdon
   • Graham Sutherland / Polynomial
   • S Nielsen

   @gsuberland @m @simonlbn @delroth "the vast majority of people with an existing [3-amp] power supply will plug it in, set the override in config.txt and just forget about it."

   • feld likes this.
   • Embed this notice
     Graham Sutherland / Polynomial (gsuberland@chaos.social)'s status on Friday, 27-Oct-2023 11:15:35 JST Graham Sutherland / Polynomial

     in reply to

     • Mike Knell
     • Pierre Bourdon
     • S Nielsen

     @m @simonlbn @delroth running the numbers, it's ~1Wpk of additional dissipation on a decent buck reg tuned for peak efficiency at peak load, so an additional 6.5% over the existing thermal output at peak load. by the time you subtract the losses of the existing VBUS 5V -> regulated 3.3V conversion, which would be replaced, it'll be even less.

   • Embed this notice
     Graham Sutherland / Polynomial (gsuberland@chaos.social)'s status on Friday, 27-Oct-2023 11:15:37 JST Graham Sutherland / Polynomial

     in reply to

     • Mike Knell
     • Pierre Bourdon
     • S Nielsen

     @m @simonlbn @delroth yeah but the real reason is always gonna be cost. fitting a sufficiently capable USB-C PD compatible buck converter on there while hitting the board size constraints would've required a buck design with high power density, and likely would've caused knock-on density increases on the rest of the design.

   • Embed this notice
     Mike Knell (m@blat.at)'s status on Friday, 27-Oct-2023 11:15:38 JST Mike Knell

     in reply to

     • Pierre Bourdon
     • S Nielsen

     @simonlbn @delroth The USB-PD spec supports a maximum of 5A at all voltages, but most implementations top out at 3A at 5V because apparently nobody is going to want 5V at 5A, right?

     Pretty sure they’re not violating the spec - it’s more that a lot of chargers violate the spec by making assumptions. Given the ferocious heat budgeting that must go on in RPi design and the fact that buck/boost converters will always generate heat (and the higher the current, the more heat) I can understand not wanting to do that on the board.

   • Embed this notice
     S Nielsen (simonlbn@infosec.exchange)'s status on Friday, 27-Oct-2023 11:15:40 JST S Nielsen

     • Pierre Bourdon

     @delroth And seems to be not just non-standard as uncommon, but actually not following the USB-C and Power Delivery standards (assuming my quick skim is correct)... and considering you need better cables for normal 5A USB-PD this is going to be "fun"...