I had an old weatherproof enclosure meant for wireless access points and I re-purposed it to house a RaspberryPi and rtl-sdr dongle. I was able to reuse the PoE circuitry to power everything over ethernet.
This will allow me to install the antenna somewhere it will get good reception and then I can connect to it across my network and experiment. It would be neat to setup an automated system to scan all radio frequencies (that it can) and build a database of background anomalies in my area.
The enclosure already had internal standoffs and an aluminum plate for mounting everything to. Here you can see the RaspberryPi v2 Model B with added heatsinks. I opted to solder an old 3-pin fan connector to provide power from a 12v to 5v regulator. I also used an old RCA cable to provide additional grounding to the case because why not.
Under the aluminum plate you can see the 12vdc to 5vdc voltage regulator and PoE breakout board. You can also see that I cut the antenna cable from the rtl-sdr and soldered a rp sma connector to adapt to the enclosure pass-through.
Here is a closer picture of the 5v regulator and PoE breakout board with questionably soldered ground wire.
Here is what it looks like with everything connected. Note that the lid is not installed, just sitting on top and rotated 90deg. I did this to provide rf shielding but increase airflow while testing. I concluded that the rpi and sdr would get too hot in normal operating conditions so I'll have to do something about that.
To combat the building heat, I added an internal 12v fan. I chose a low power fan since I am already reaching the maximum current that I would like to draw over PoE. The lower the current, the longer you can have your cable. Using CAT6 cable also helps since it is a larger gauge wire. I added a 4700uf capacitor to the bottom of the mounting plate by soldering it directly to the 12v output from the PoE circuit. This is not required but I think it will help with any current draw spikes from the rtl-sdr or RaspberryPi.
I also added a little bit of foil tape inside to reduce potential noise and secure loose cables.
The fan helps a lot. I ran the system at the hottest part of the day and the enclosure itself was much warmer but the internal components were cooler. The fan is doing its job of mixing the air and transferring the heat from the electronics to the metal case. I will have to perform some testing to see what type of noise/interference it introduces.
I am trying out a planar disk antenna made of two pizza trays from the dollar store. It is recommended on reddit.com/r/rtlsdr and is very cheap to make, costing only $2 USD.
Using this setup, with no filters or amplifiers, I was able to run rtl_power for 12 hours to survey 50MHz-1.2GHz. I could then generate this output with heatmap.py. Each pixel represents 1Mhz average across 73seconds at that frequency.
The next steps include tuning the antenna and mounting this in the attic or somewhere more permanent. Also adding filters should cut down on the noise that you can see in the above output.