One of the things I’ve always wondered was just how sensitive is an SDR compared to a modern super heterodyne receiver such as what is used in two-way radio equipment. I decided to break out my Motorola (General Dynamics/Freedom CTE) R2670A service monitor and configure it to test an RTL-SDR (Nooelec Nano 3 specifically).
I started off with clean install of Ubuntu 20.04 (over the last year I’ve really begun to avoid Debian based systems but needed Debian package support to play with OP25) on a Minix Neo Z83-4 Plus and installed the rtl-sdr, gnuradio and gqrx packages using apt. Next I plugged in my Nano 3 and fired up gqrx and set the monitored frequency to 151.5125 MHz. 151.5125 MHz (which I call 3D1 on my personal radios due to the P25 NAC I use with it) is one of my licensed itinerants and just so happened to be already plugged into my service monitor. I then connected the RF I/O port on the service monitor to the antenna port on the Nano 3 via a 6 ft coax jumper (about 0.2 dB loss at that frequency) and begun to generate a 1 kHz tone at 3 kHz of deviation at -90 dBm just to let everything begin to warm up and stabilize.
After about 10 minutes, I corrected the RTL’s frequency error (it was 1.0 ppm or 151 Hz off) as the Service monitor is a “calibrated” piece of test equipment and cranked the RF output to -50 dBm to set the computer for rated audio (technically should be done at -47 dBm or 1 mV but that require switching over to the RF Out port on the monitor) and I adjusted the volume on output of the Minix’s headphone jack until I obtained -3 dBm ( ~0.5 VAC).
I then reconfigured the service monitor to show me the SINAD meter so I could test the Nano 3’s receive sensitivity using the industry standard 12 dB SINAD (Signal, Induced Noise, and Distortion) for reference.
With the filtering in gqrx set to “Wide” (in terms of a 25 kHz channel or 5.0 kHz maximum deviation) the 12 dB SINAD receive sensitivity of the Nano 3 came out to -100.0 dBm (2.24 µV). The filters were set to narrow (as in a 12.5 kHz channel with 2.5 kHz maximum deviation) and the deviation set to 1.5 kHz resulting in a 12 dB SINAD sensitivity of -105.0 dBm (1.26 µV).
Now, why did I perform this test? I was curious to see how sensitive the RTL dongles everyone is using for abstract receiving actually are but I also am working on a project involving the development of a low cost and portable P25 repeater. I am aware of development boards such as the MMDVM but one of the product licenses for that board actually states that it not be used for commercial purposes (which I take to mean commercial radio) but right now I believe the public safety industry lacks a low cost field deployable repeater in the sub-$3,000 USD range (otherwise taxpayer money goes to purchase $30,000 units from major vendors). I really wanted to see if something like an SDR and a SBC running OP25’s repeater module would be a viable receiver/controller option but at the moment I do not believe it would be.
