1264Re: [Raspberry_Pi_4-Ham_RADIO] WsprryPi for WSPR15
- Apr 9, 2013Couple thoughts:
1 - most wall wart power supplies will NOT provide a GROUND, they can't if they are the simple 2-pin
ac plug type since they only use hot and neutral. Even if you found a 3-pin one, most likely the DC
output will be isolated from the AC connector.
2 - the ethernet connection will not normally supply a ground to the pi either.
3 - If you want the Pi grounded you are probably best off to supply a good ground to the pi separately.
4 - If your using the GPIO pins as OUTPUTS some isolation/buffer chips/transistors are highly
recommended, especially to drive an antenna.
5 - If using the GPIO pins as inputs, buffering/isolating is almost mandatory, especially since they
are 3.3v max.
6 - The micro-usb power connector is the only fused power input. Power applied anywhere else needs
to be fused externally from the board. (option - modify the board)
7 - good news, Pi's are in-expensive... Not like frying a $800+ computer.
-Jim- N1JMM, KB1YPL_B
Visit my Pi running ircddb, dvrptr, Apache & many other things:
On 4/7/13 8:02 PM, Robert Nickels wrote:
> Sad news to report ;-( After two days of amazing spots on WRPR it
> seems my Pi has died while in Wsprry Pi service - although I have no
> idea whether it was KIA or just coincidence.
> I'd switched the antenna to a vertical last night and started the beacon
> around 0445 UT in order to evaluate it's performance against the beam
> I'd been using on 20 meters. This morning I was surprised to only see
> no spots after 12:22 UT, and upon investigation I discovered that the
> beacon had failed and the Pi would not boot up. Upon further
> investigation using a bench power supply, I observed excessive current
> consumption (1.1 A from the 5 V supply) and the output of the 3.3 volt
> regulator is low, and drops even lower the longer power is applied - to
> less than 2 volts. One time I was able to observe a partial boot
> sequence before everything dropped out. The 3.3 volt regulator package
> is hot to the touch after such attempts and the 3.3 V line measures < 2
> ohms to ground. Fried Pi!
> My theory is that some part of the SOC has been damaged, and while the
> CPU part is still intact (and tries to run), the damaged part is
> drawing so much current that the regulator fails to achieve the 3.3 VDC
> needed (or drops below that level very quickly). The polyfuse is
> fine, and was not in the circuit at the time as I was feeding power to
> the Pi via the GPIO connector based on a suggestion I found on one of
> the R-Pi sites to eliminate having to use the microUSB cable. I
> wasn't aware until I looked at the schematic today that this method
> bypasses the fuse, although I'm not sure how that would have made a
> difference in this case.
> For Wsprry use, the low pass filter is capacitively coupled to GPIO pin
> 7 (R-Pi) which is connected directly to the GPIO4 (Broadcom) pin on the
> BCM2835. Hindsight being 20/20, I now wonder about the wisdom of
> coupling a outside antenna to a GPIO pin that goes directly to the CPU
> chip without any kind of isolation. Transient voltages such as
> lightning or static buildup as well as RF from other transmitters
> operating into nearby antennas are a concern. And, a ground loop is
> created between the shielded side of the antenna coax and an other
> device connected to the Pi that wind up grounded via the AC powerline
> neutral or grounding conductors. My theory is that the GPIO was
> damaged by one of these means, or in some other way that I haven't
> thought of yet.
> I think that one stage of isolation using a simple BJT amplifier would
> provide cheap insurance, as would galvanic isolation between the antenna
> and the Pi. I plan to try both when the replacement Pi arrives.
> Meanwhile, I thought others could benefit from this experience.
> Any thoughts or theories on this would be appreciated as well,
> especially by Guido and others who have been using Wsprry Pi longer than
> I have.
> 73, Bob W9RAN
> Yahoo! Groups Links
- << Previous post in topic