Gerald,
I would like to request a small improvement in the future revisions of
Beagleboard: Would it be possible to add a header (in similar fashion
as the McBSP2 header) or at least mounting holes for it, for power
supply input? It would not only increase robustness (no plugs and
wires sticking on the sides) of designs where the board is powered
form a daughter board, but also allow for more compact mechanical
construction.
thx,
j.
You can actually power the Beagle via 5V pins on the expansion header, so there's no need for an additional one. The (custom) expansion boards I've used do this to provide power from batteries.
--Tom
Thank you for your input. We just released the -xM so there are no near term plans to add a power header to the board, but we will take it under consideration when we do look to make any substantial design changes. In addition, there are no immediate plans to make any changes to the rev C4 version as well.
Gerald
That is fine, just don’t plug the OTG cable into a PC unless you have a dummy plug in the power jack.
Gerald
Cool!
Expansion header was my first bet, but before sending the email with
the suggestion I spent some time to check the schematic and I was sure
that the power supplied there would conflict with the power management
and conditioning circuits.
Thanks for the answer.
j.
My suggestion for tighter packing is to (unsolder the +5V connector and) solder replacement wires directly on the +5V pads. This work must be done with care!
Greetings to all,
siñ
That is fine, just don’t plug the OTG cable into a PC unless you have a dummy plug in the power jack.
Could you expand on that?
Lioric
Plug plugged into the power connector disconnects the power from the
USB connector to avoid a conflict between USB and power supply.
j.
Generally speaking, it is not a good idea to connect two power supplies together. The act of connecting the power to the DC jack, removes the ground from the USB power switch, keeping it from turning on. If DC is supplied from the expansion connector and the OTG port is connected at the same time, then those two rails will be connected together as the DC jack insertion was not present to turn off the switch.
The result will be sparks and possibly a damaged PC. It most likely will not harm the Beagle, but possibly the PC motherboard or at least cause the USB port to shutdown.
Gerald
Use forward-biased diodes on the output of each of the bridged
supplies to prevent supplies from swapping current. Of course, this
will lower the output of the supplies by a diode drop, which is a
fixed value depending on the diode used.
As this voltage is used by the USB host port power, that can cause issues with USB devices plugged into the board.
Gerald
Depends on how good you 5V is. If it is low say 5V, then the result will be at east 4.8V on the USB rail at that point which is right on the edge. Some USB devices may not work.
Gerald
A 0.2V drop across a Schottky diode is definitely on the low side and
will vary with current/temperature/device. This sort of thing could
be done pretty easily with a couple of SO-8 (or maybe smaller) P-FETs
and some logic (or a couple of voltage monitors). Otherwise, I would
be surprised if Maxim or TI don't already make an IC for exactly this
type of application.
On a related note, I didn't read the full datasheet, but wouldn't U1
pull SW_PLDN (and therefore SW_OUT) low if you plugged in power via
the USB client port and the barrel connector at the same time? Since
this has likely been done before I'm guessing the BB would
survive...probably a high Rds FET on the SW_PLDN output. IF I'm
right, then all this is doing is draining unnecessary power...which
could be fixed by lifting SW_PLDN from its pad...IF I'm right.
Also, I think if you provided power via the expansion connector and
the USB client port at the same time without a dummy plug in the
barrel connector, there is a good chance you'd either burn U1 or R6.
Seems like depopulating R6 and NOT jumpering J2 if you're concerned
about this type of situation might be a good idea.
The Beagle not only survives, but this is the way it is designed, to remove the ground from the switch when the DC jack is inserted, making the signal high, turning off the switch disconnecting the path from the USB power to the 5V rail. If you do not have the DC jack inserted, the switch is on. So, if you connect 5V from the expansion header and the USB the two 5V rails will be connected together in a shorting configuration which should not hurt Beagle, but it could damage the PC USB port or the power supply. We have shipped over 22,000 boards with this circut. So far is seems to be holding up.
You can read the circuit description in the System Reference Manual for a more detailed description of how the circuit works.
Gerald
I'm not suggesting its a total failure. I'm only saying it could be
improved and it isn't idiot proof.
No, it isn’t idiot proof. Then again most things aren’t.
Gerald
Sean Fuller wrote:
A 0.2V drop across a Schottky diode is definitely on the low side and
will vary with current/temperature/device. This sort of thing could
be done pretty easily with a couple of SO-8 (or maybe smaller) P-FETs
and some logic (or a couple of voltage monitors). Otherwise, I would
be surprised if Maxim or TI don't already make an IC for exactly this
type of application.
sure, but more fancy ICs means more cost and BB is trying to by lowcost
Just for the record, we did add a few fancy ICs to the -xM version so that if you do plug in a 12V supply a pretty red idiot light comes on and the board won’t power up. So we do try to help those in need!
Gerald
Hi Gerald,
A RED idiot LED - that is far toooo much…
Almost as cool as Seppo connector.
Seriously, did you ever consider a vertical +5V connector.
With expansion boards bottom-to-bottom with BeagleBoard
it would save some packing (and board) space.
Just a late night idea from Helsinki, Finland.
BTW, our 24-bit, 192 kS/s BB-DAC is alive and working well.
We now play 24-bit music from SD cards. Next over IEEE 802.11an WLAN.
I just cleaned the board lay-out a little and will have a decent prototype
to send to you in late September. Thank you for waiting!
Best regards,
siñ