Featured Articles
Announcing Pi
Technical Specifications
Raspberry Pi FAQs
Win a Pi
Raspberry Pi Wishlist
Actions
Forum Thread Details
  • State Suggested Answer
  • Replies 21 replies
  • Answers 2 answers
  • Subscribers 679 subscribers
  • Views 4499 views
  • Users 0 members are here
Related

Powering via the USB ports

Former Member

Have just bought a B+. Reading all the previous documentation on the B indicated I could power the Pi via a USB port if I wanted. I tried this on my B+ and nothing happened !.

 

My particular application has a USB module capable of supplying 2 amps at 5V and will be used for communication with the rest of my system. I do not want to use the micro-usb for power. The connector is inconveniently located and I have no 5V supply apart from the USB module. (This is for an embedded application).

 

Can I use the USB ports for input power on the B+? If not, why not?  The very sparse schematic for the B+ is no help. Is there a more comprehensive one in preparation? It doesn't show the USB ports at all, unlike the one for the B. If there is some 'block' on the USB 5V, could I patch from the USB ports to the 5V pins on the GPIO? (I am a retired professional EE with a long experience of microprocessor systems and quite happy to solder to my Pi if needed). I did note that my mouse, also connected into a USB port, lit up when I connected the power to another USB port - but the Pi itself didn't !

 

Also couldn't find how to send / receive data via a USB port using Python. How does the Pi allocate the USB ports? When booting, my Pi seems to recognise the external USB module  - without installing a driver, which is very surprising as it is a non-standard USB device. How do I check these ports? (Any 'device manager' like Windows?).  I have previously used USB ports as virtual COM ports in Windows. I need to be able to receive and send data (text strings) via a USB from a Python program (or maybe a C# and mono at some stage. Speed may be an issue for me)

 

Regards

 

Mike

Parents
  • 0

    Mike Bolton wrote:

     

    Have just bought a B+. Reading all the previous documentation on the B indicated I could power the Pi via a USB port if I wanted. I tried this on my B+ and nothing happened !.

     

    My particular application has a USB module capable of supplying 2 amps at 5V and will be used for communication with the rest of my system. I do not want to use the micro-usb for power. The connector is inconveniently located and I have no 5V supply apart from the USB module. (This is for an embedded application).

     

    Can I use the USB ports for input power on the B+? If not, why not?  The very sparse schematic for the B+ is no help. Is there a more comprehensive one in preparation? It doesn't show the USB ports at all, unlike the one for the B. If there is some 'block' on the USB 5V, could I patch from the USB ports to the 5V pins on the GPIO? (I am a retired professional EE with a long experience of microprocessor systems and quite happy to solder to my Pi if needed). I did note that my mouse, also connected into a USB port, lit up when I connected the power to another USB port - but the Pi itself didn't !

    I haven't heard of a full schematic for Model B+, but I haven't been paying much attention.  The Model B had problems with hot-swapping USB devices, so when B+ improved hot-swapping it lost the back-powering "feature".

     

    There's a thread about lack of back-powering in an unmodifed B+ at the RasPi forum: http://www.raspberrypi.org/forums/viewtopic.php?f=63&t=83014.  One of the moderators does suggest jumpering from USB +5V to GPIO +5V.  Doing this should work, but you lose fuse protection.  If it were my board, I'd solder from USB +5V to fuse F1 pin 1 so I still have fuse protection.  This is electrically equivalent to an external USB cable from a USB hub to RasPi's Micro USB power jack.

     

    Also, different USB hubs have different abilities to back-power.  Standard USB ports are limited to 500 mA, though cheap hubs just connect all the +5V lines together so you can get 2A.  The upstream connection for a hub may or may not be able to back-power.  They aren't supposed to, and some hubs have a diode to prevent it.

     

    Hope this helps!  Advice is offered without warranty - YMMV.

  • 0 in reply to johnbeetem

    John

     

    Thanks very much for this. I will give it a go.

     

    You suggest connecting the +5V from the USB connector to end 1 of the fuse, to give 'protection'. Although I don't have the schematic, it looks as if all the USB connectors have a common 5V. If I power the Pi from one of these, then all the others will have their 5V off my external supply? This is fine as it is a 2A supply so hot swapping should not be a problem, but, in that case, what is the fuse actually protecting? OK, it may be the GPIO 5V pins?

     

    If I connect the USB 5V to one end of the fuse, is the other end connected to the USB 5V line? In that case I will have shorted out the fuse?

     

    Regards

     

    Mike

  • 0 in reply to Former Member

    Mike Bolton wrote:

     

    You suggest connecting the +5V from the USB connector to end 1 of the fuse, to give 'protection'. Although I don't have the schematic, it looks as if all the USB connectors have a common 5V. If I power the Pi from one of these, then all the others will have their 5V off my external supply? This is fine as it is a 2A supply so hot swapping should not be a problem, but, in that case, what is the fuse actually protecting? OK, it may be the GPIO 5V pins?

     

    If I connect the USB 5V to one end of the fuse, is the other end connected to the USB 5V line? In that case I will have shorted out the fuse?

    Just to be clear, I'm using the partial schematics published here at element14: http://www.element14.com/community/docs/DOC-68376/l/raspberry-pi-b-v12-schematic.

     

    The fuse is protecting RasPi's SoC, DRAM, and everything else powered from 5V (or lower voltages generated by U3).  The fuse is protecting them from an overvoltage condition on the power input.  For example, suppose a cheap external power supply connected to the Micro USB power input jack decided to lose regulation and instead of producing 5V +/- .25V it produced 8V.  Well, this would probably destroy U3 which could short out and supply that same 8V to the SoC and LAN9514 and destroy the chips on the board.

     

    What's supposed to happen instead is that Zener diode D5 prevents 5V from going above 5V (plus a safe margin).  However, that erroneous 8V supply could produce a lot of current and if D5 is dropping 5V with too much current it will heat up and destroy itself.  Polyfuse F1 prevents this from happening: if there's more than 2A going through F1, it heats up and becomes more resistive, so that the current drops to a safe level before D5 gets too hot.

     

    Similarly, if you hook up a 5V supply backwards, D5 conducts with a voltage drop less than 1V.  Again, if there's too much current F1 heats up and turns into a resistor.

     

    Now, if you hook up your USB +5V power supply to RasPi's 5V at the GPIO connector, if your supply misbehaves and turns on D5 there's nothing to protect D5 from overheating and starting a fire (or desoldering itself).  Hooking up USB +5V to the F1 fuse input allows F1 to do its job and product RasPi from a bad power supply.

     

    Addressing your other questions:

     

    1.  I suspect that all the USB +5V pins are connected.  You can verify that with an Ohmmeter.

     

    2.  If you add a jumper from USB +5V to F1 pin 1, you're not shorting out F1.  There must be a diode or "ideal diode" circuit between RasPi's main 5V rail and USB +5V.  Otherwise you'd be able to back-power.  So here's your current path: Your USB +5V goes to F1 pin 1, which goes through F1 and Q3 to power RasPi's 5V rail.  However, there will be a small voltage drop across F1 and Q3, so the 5V rail will be less than USB +5V.  This means the diode between the 5V rail and USB +5V won't turn on, and the 5V rail won't power the other USB ports.  Your USB +5V source will do that instead because it's a higher voltage than the 5V rail.

Reply
  • 0 in reply to Former Member

    Mike Bolton wrote:

     

    You suggest connecting the +5V from the USB connector to end 1 of the fuse, to give 'protection'. Although I don't have the schematic, it looks as if all the USB connectors have a common 5V. If I power the Pi from one of these, then all the others will have their 5V off my external supply? This is fine as it is a 2A supply so hot swapping should not be a problem, but, in that case, what is the fuse actually protecting? OK, it may be the GPIO 5V pins?

     

    If I connect the USB 5V to one end of the fuse, is the other end connected to the USB 5V line? In that case I will have shorted out the fuse?

    Just to be clear, I'm using the partial schematics published here at element14: http://www.element14.com/community/docs/DOC-68376/l/raspberry-pi-b-v12-schematic.

     

    The fuse is protecting RasPi's SoC, DRAM, and everything else powered from 5V (or lower voltages generated by U3).  The fuse is protecting them from an overvoltage condition on the power input.  For example, suppose a cheap external power supply connected to the Micro USB power input jack decided to lose regulation and instead of producing 5V +/- .25V it produced 8V.  Well, this would probably destroy U3 which could short out and supply that same 8V to the SoC and LAN9514 and destroy the chips on the board.

     

    What's supposed to happen instead is that Zener diode D5 prevents 5V from going above 5V (plus a safe margin).  However, that erroneous 8V supply could produce a lot of current and if D5 is dropping 5V with too much current it will heat up and destroy itself.  Polyfuse F1 prevents this from happening: if there's more than 2A going through F1, it heats up and becomes more resistive, so that the current drops to a safe level before D5 gets too hot.

     

    Similarly, if you hook up a 5V supply backwards, D5 conducts with a voltage drop less than 1V.  Again, if there's too much current F1 heats up and turns into a resistor.

     

    Now, if you hook up your USB +5V power supply to RasPi's 5V at the GPIO connector, if your supply misbehaves and turns on D5 there's nothing to protect D5 from overheating and starting a fire (or desoldering itself).  Hooking up USB +5V to the F1 fuse input allows F1 to do its job and product RasPi from a bad power supply.

     

    Addressing your other questions:

     

    1.  I suspect that all the USB +5V pins are connected.  You can verify that with an Ohmmeter.

     

    2.  If you add a jumper from USB +5V to F1 pin 1, you're not shorting out F1.  There must be a diode or "ideal diode" circuit between RasPi's main 5V rail and USB +5V.  Otherwise you'd be able to back-power.  So here's your current path: Your USB +5V goes to F1 pin 1, which goes through F1 and Q3 to power RasPi's 5V rail.  However, there will be a small voltage drop across F1 and Q3, so the 5V rail will be less than USB +5V.  This means the diode between the 5V rail and USB +5V won't turn on, and the 5V rail won't power the other USB ports.  Your USB +5V source will do that instead because it's a higher voltage than the 5V rail.

Children
No Data