How does this work? Does it prevent problems when the Pi is running from it's own power supply?
Only two USB ports so a very early one.
It was definitely back powering, the hub had a PSU plugged in, I connected the hub's input plug to the Pi's USB socket, no PSU was connected to the Pi but it powered up.
I didnt have to supply power to the Pi, merely plugging a powered hub into the USB socket of the Pi powered it up.ProDigit wrote: ↑Thu Sep 12, 2019 6:15 pmOn a positive note, you could use one of the hub's USB ports and power a pi with that. As long as the powered hub gets more than 2,5A at 5V (1A at 12V) on the input, you can save cables there. (Meaning using data connection + power connection to the hub, for a more stable power).
The problems usually with powering 2 power supplies, is a ground lift issue of one PSU, can cause the other to blow up or fail. So using only 1 supply is important.
As far as USB hub power issue,
When using cables that transmit just data, no power, theres still a common ground.
I would just power the pi from the hub, using 2 cables, and make sure the hub has a sufficiently high rated PSU.
Did you connect the hub using one of the output ports (in which case, the behavior is expected for the Pi you've indicated), or was it by connecting the hub's uplink connection? The former is not back powering.
It was very early Pi - only two USB ports.W. H. Heydt wrote: ↑Wed Sep 18, 2019 12:27 amDid you connect the hub using one of the output ports (in which case, the behavior is expected for the Pi you've indicated), or was it by connecting the hub's uplink connection? The former is not back powering.
In any case, the very early Pis could be back powered, as can the Pi0 and Pi0W.
The PiHut hub definitely back powered the Pi. No connection was made to the microUSB power input socket on the Pi.
It's not seen as a problem, because the great unwashed use either a desktop or a laptop which take a great deal of power compared to the amount a powered hub could supply, and probably have protection built-in because there's plenty of room on board, and the cost is small compared to the rest of the circuitry.I never had even heard of the term 'back voltage' until the Pi was launched so why was it a problem for the Pi but not for conventional computers?
I'm not certain that it wasn't a "problem" for "conventional computers". Having contributed to previous threads about this (and elsewhere before acquiring my first Pi from the "second batch" in 2012) and checked out quite a few devices I already had at that time (see: http://www.cpmspectrepi.uk/raspberry_pi ... esChk.html ),I recall that in at least one thread it was stated that, according to the specifications for a powered USB 2.0 hub it should not "back-power" (ie. have a protection diode or equivalent built-in). At that time, AIUI, USB 2.0 was still a specification defined by a "consortium" of USB (chip) manufacturers rather that an international (ISO) standard (and I'm not sure if it's a "standard" now - or rather no one here has referenced such AFAIK).geffers wrote: ↑Mon Sep 23, 2019 9:09 amI'm a wee bit puzzled with this back power issue.
In the early days of Desktops, powered hubs were the norm otherwise, if used in passive mode the 4 (or 8) port hub could only use a maximum of 500mA being the current output of the single port the hub is plugged in to.
I never had even heard of the term 'back voltage' until the Pi was launched so why was it a problem for the Pi but not for conventional computers? Are later Pi devices protected, if so which one's? Is the Zero protected?
Reason I ask is that over the years I have acquired numerous models for various projects, I have four that have only got two USB sockets.