[LAU] JACK on Pi... almost working

Simon Wise simonzwise at gmail.com
Wed Jul 30 09:21:16 UTC 2014


On 30/07/14 17:50, Ken Restivo wrote:
> On Sat, Jul 26, 2014 at 08:44:30PM +0200, Jeremy Jongepier wrote:

>>
>> Did you try backfeeding the RPi via a powered USB hub?
>>
>
> I'm not sure what this means.

the 5V rail in the RPi A and B rev 2 (but it has improved quite a bit in the new 
B+ models, and was quite different in the original rev 1) seems directly 
connected to the USB outs, the 5V pins and the 5V holes while the miniUSB power 
inlet has a 6V 1.1A polyswitch limiting power in (which may be the reason you 
can get reboots when hotplugging USB devices). So if you connect power to a hub 
that sends that power back through its USB inlet you can power the Pi that way, 
plus whatever else is on the hub, without the 1.1A limit going via the Pi. A 
proper hub shouldn't be connected that way (I'd guess it should have a diode or 
some way of switching its source) and you should be able to power each 
separately. Hubs in theory shouldn't do it, but many do. You are bypassing the 
over-current protection on the Pi also, so better do it correctly, I'd guess the 
hub has some protection and the Pi alone won't draw too much, but you could 
certainly draw too much from the 5V pins on the header.

USB leads to the miniUSB will often have considerable voltage loss if you try to 
put 1 amp or so through them, that is well over the spec for USB, so better make 
sure it is actually a power lead not just a USB lead, maybe full sized USB leads 
tend to be not so lightweight (so again better luck back through the USB outs)?

Re 2A claims on power supplies ... when trying to set up a system that needed a 
good power supply the labels meant nothing, like labels on domestic audio amps I 
guess. I tested some by how long they took to charge my Samsung Tablet ... a new 
Apple power supply using the USB lead supplied by the Samsung took 16 hours, and 
almost anything plugged into the USB outs of the RPi caused trouble if I used it 
in the setup, it is labelled 5.2V 2.5A while the Samsung supplied one is 
labelled 5V 2A and takes about 1/3 that time to charge (with the same lead) ... 
this makes sense for an 8 amp hour battery, and it runs the RPi setup well.

The new model B+ has a 2A polyswitch. It is said to need less power and now 
there is some kind of regulator circuit between the miniUSB and the 5V rail 
(there always has been one between the 5V and the 3.3V of course), so it is 
quite different. Very clean power would be vital for the analogue audio out, 
since it is just a very simple low pass filter on the board, filtering a bitstream.

Simon


More information about the Linux-audio-user mailing list