5 ms·

>And many (I dare say "most," or "nearly all") USB-PD power supplies top out at 3A when in 5V mode. 5 A at 5 V seems to be sufficiently unusual for USB-PD that

by cge 3y ago

>And many (I dare say "most," or "nearly all") USB-PD power supplies top out at 3A when in 5V mode. 5 A at 5 V seems to be sufficiently unusual for USB-PD that Wikipedia's section on the specification doesn't mention it as being part of the specification at all, giving 3 A as the maximum in the specification for 5 V, and only mentioning 5 A with EPR and marked cables at higher voltages [1]. Discussions about 5 A at 5 V seem to come up essentially only in the context of the rpi5, and there most references seem to be to footnotes in the specification mentioning that more than 3 A can be advertised if a 5 A cable is detected. The rpi5 also apparently doesn't support 5 A at 5 V with PPS; it only supports it as a fixed voltage profile. The result would seem to be that essentially no USB-PD power supply other than theirs, and one other SBC power supply, actually offers 5 A at 5 V in a way that the rpi5 will accept. [1]: https://en.wikipedia.org/wiki/USB_hardware#USB_Power_Delivery https://en.wikipedia.org/wiki/USB_hardware#USB_Power_Deliver...

ssl-3 3y ago
It is unusual. I'd even go as far as to say that it is an awful and stupid decision to do this at a time when USB-PD operating at higher voltages is finally becoming somewhat commonplace (common, as in: we might have extra junk laying around that can do it), but I'm not trying to make 90,000 widgets per month and build them down to a price. But a casual Google search where I am located at least two alternative suppliers that are already shipping appropriate power supplies that can do 5@5. And the usual (US) distributors (places like Adafruit and also places like Digikey) are selling the official power supply for $12, which I personally consider to be very cheap. --- And remember, wanting peculiar power supplies is normal in the RPi space. The very popular Pi 3 was weird (5.1V@2.5A over MicroUSB, with many unofficial recommendations suggesting 3A), for instance -- in a world where common "high-power" USB power bricks topped out at 5.0V@2.4A on a good day, with the wind from the South, and under the direction of an unladen sparrow. Many of us are familiar with the struggles that this entailed as we worked on whatever we were working on. But other suppliers did come to the front and offer alternative products that reliably delivered 5.1V@3A, and there's no indication that this won't be the case for this round as well -- especially if they're able to produce the quantity of RPi5s that they say they will. At a projected 90k units per month, then: Every 7 seconds, an opportunity to sell another RPi5 power supply opens up. (The Pi 4 has a different version of the same story.)
- RetroTechie 3y ago
  > It is unusual. I'd even go as far as to say that it is an awful and stupid decision to do this at a time when USB-PD operating at higher voltages is finally becoming somewhat commonplace (common, as in: we might have extra junk laying around that can do it) Not even "junk drawer gear". USB-PD is recent enough that any gear supporting a good cross-section of that spec will be recent if not brand new. As 1 data point: I recently bought a USB-C power adapter that supports USB-PD up to 35W at a variety of voltages & negotiation methods. A RPi5's power needs are well within that 35W. And yet, beside supplying 5V @3A (which comes with its own set of issues / limitations), it'll be largely useless as a RPi5 power supply. Regardless of cable used. Many USB-PD power bricks will be in same boat. So indeed this is a stupid design decision on RPi's behalf. I've read & understand their rationale for that decision. But it doesn't make it less stupid. Other SBCs use USB-PD as intended (in this case, @ higher voltages like 9, 12 or 15V, or even variable) no problem, and without needing special 5A rated cables. The onboard electronics / ICs to do this are cheap & plentiful. Okay - RPi's rationale is an excuse. Not a good one. Supplying a 5V/5A brick themselves, doesn't do much to paper over that fact.

by cge 3y ago

ssl-3 3y ago

It is unusual. I'd even go as far as to say that it is an awful and stupid decision to do this at a time when USB-PD operating at higher voltages is finally becoming somewhat commonplace (common, as in: we might have extra junk laying around that can do it), but I'm not trying to make 90,000 widgets per month and build them down to a price. But a casual Google search where I am located at least two alternative suppliers that are already shipping appropriate power supplies that can do 5@5. And the usual (US) distributors (places like Adafruit and also places like Digikey) are selling the official power supply for $12, which I personally consider to be very cheap. --- And remember, wanting peculiar power supplies is normal in the RPi space. The very popular Pi 3 was weird (5.1V@2.5A over MicroUSB, with many unofficial recommendations suggesting 3A), for instance -- in a world where common "high-power" USB power bricks topped out at 5.0V@2.4A on a good day, with the wind from the South, and under the direction of an unladen sparrow. Many of us are familiar with the struggles that this entailed as we worked on whatever we were working on. But other suppliers did come to the front and offer alternative products that reliably delivered 5.1V@3A, and there's no indication that this won't be the case for this round as well -- especially if they're able to produce the quantity of RPi5s that they say they will. At a projected 90k units per month, then: Every 7 seconds, an opportunity to sell another RPi5 power supply opens up. (The Pi 4 has a different version of the same story.)

RetroTechie 3y ago
> It is unusual. I'd even go as far as to say that it is an awful and stupid decision to do this at a time when USB-PD operating at higher voltages is finally becoming somewhat commonplace (common, as in: we might have extra junk laying around that can do it) Not even "junk drawer gear". USB-PD is recent enough that any gear supporting a good cross-section of that spec will be recent if not brand new. As 1 data point: I recently bought a USB-C power adapter that supports USB-PD up to 35W at a variety of voltages & negotiation methods. A RPi5's power needs are well within that 35W. And yet, beside supplying 5V @3A (which comes with its own set of issues / limitations), it'll be largely useless as a RPi5 power supply. Regardless of cable used. Many USB-PD power bricks will be in same boat. So indeed this is a stupid design decision on RPi's behalf. I've read & understand their rationale for that decision. But it doesn't make it less stupid. Other SBCs use USB-PD as intended (in this case, @ higher voltages like 9, 12 or 15V, or even variable) no problem, and without needing special 5A rated cables. The onboard electronics / ICs to do this are cheap & plentiful. Okay - RPi's rationale is an excuse. Not a good one. Supplying a 5V/5A brick themselves, doesn't do much to paper over that fact.