I think peterlite's idea is to make more 5V power available at the USB socket, by injecting 5V from the PSU directly under the socket. Without knowing the Pi5's circuits, it is hard to know whether that is safe, but it is certainly not a good scheme because it will short out the USB power protection.Why?Using official Pi 5 power supply.PD capable PSU with a 5V/5A mode
Maybe the board would stay cooler with big wire soldered from power to USB port.
If the current is entering the Pi one one side and exiting on the other it is passing through the Pi circuitry. Any resistance or inefficiency in that path will lead to heat build-up. If your supply wire is getting warm then maybe it would be wise to replace it with something bigger, but that's not what's happening here. What you observed was the Pi remained cool when two peripherals (keyboard and monitor) were powered externally, but got warm when they were powered through the Pi. That's exactly what you would expect (unless the Pi was a perfect conductor).
What you could do is split the power path before the Pi. Thus, power to the peripherals reaches them via a more direct route, and perhaps with heaver cables, than through the Pi. How you do that is up to you, but it's complicated by the 1-to-1 relationship between power supply and device which is used to negotiate the power delivery. If you choose to supply 5V directly to the Pi GPIO header then you can avoid that to some degree, but you still have to take care with how much power is available, and where it's going.
Yes to splitting the power at a suitable (dual outlet) PSU, as long as the outputs are suitable (what current does the Pi need when working without the monitor?)
No to applying power via the GPIO header, if it can be avoided. It introduces a whole new set of problems (wire gauge, contact resistance, configuration if the Pi plus peripherals will exceed 3A).
Statistics: Posted by davidcoton — Sat Jul 13, 2024 8:03 am