BBB MMC / SD detection instability with U-Boot 2014.04 (CPU 1GHz)

[Bernd Walter]

On Thu, May 22, 2014 at 11:34:50AM -0300, Luiz Otavio O Souza wrote:
> Just picking one email to reply...
> 
> This is probably unrelated, but i can consistently corrupt my SD Card
> data if i connect (and use) a very small R/C servo (the 9g ones, used
> on foam planes) to the 3.3v on the expansion headers.

R/C servos usually are extremly harsh on power sources.
I know this because decades ago I'd build LED based akku charge displays
for R/C helicopters and I needed a lot of input filtering to not sweep
over the full display range when servos had been active.

> I'm not sure if it is because of noise or the power consumption, but
> looks like the BBB is really picky about the power conditions here.

It is the noise on the power consumption.
I don't know how good the 3.3V rail is stabilized, but I wouldn't
be surrised the the servo would produce a noise level of way more than
3.3V, so your rail even can be negative or more than doubled.
This can even damage components as they are not designed to handle
such voltages.

> Connecting the servo to the VDD_5V on the same header makes the problem go away.

The 5V rail is likely not directly used in any digital circuitry
beside the volatage regulators input and their output are
influenced.
The volatage regultaors however are also designed for small input voltage
ranges and can be damaged by invalid input voltages on the 5V rail.
An inductive load can be a power sink or a power source!
Effectively you are connecting something to the 3.3V or 5V rail,
which sometimes act as a high voltage battery.

> So be careful about what you connect on your boards :)

Absolutely.

-- 
B.Walter <bernd at bwct.de> http://www.bwct.de
Modbus/TCP Ethernet I/O Baugruppen, ARM basierte FreeBSD Rechner uvm.