As I remember it, I had to first disable the File Source block and the Throttle block, they were causing problems for some reason. Then of course the proper source had to be enabled. Then the first FIR filter in the chain had to be connected to the USRP or RTL2832 source (in this case I'm using a RTL device). I received a warning about that and the flowgraph would not compile/run without it being connected.

One of the issues with the OP25 flowgraph example from gr-baz github, in that it seems to be two circuits (flowgraphs) in one. In the flowgraph, there's one circuit that uses the Combined OP25 Decoder block, and a separate circuit which uses the Simple OP25 Decoder block (it may have been much simpler to have two distinct flowgraphs).

I decided to use the circuit with the Combined OP25 Decoder block, and carefully and meticulously clear away any blocks that were not necessary for the operation of that circuit. I cross referenced it with another flowgraph using the Combined OP25 Decoder block and everything looks correct.  When I run the flowgraph and observe the symbol decoder graph (DiBits), at first I could not see the symbol stream lines at all. - There are three parameters which can be adjusted. The Fine Offset, Xlate Bandwidth and Gain. After fiddling with these parameters for a time, I started to see two symbol lines appearing, the uppermost and lowermost, although they weren't perfectly well defined. I also noticed that by slowly adjusting the Fine Offset and BW, I was able to get one symbol-stream at a time ("scroll through" the four symbol streams), but never all four at once (you need all four to reconstruct the audio).

I believe that sample rate has a lot to do with it, since every block in a flowgraph depends on the master samp_rate block. At best, I was able to get the four symbol streams, but never parallel with each other. In other words, the symbol stream was jumping between all four positions, fragmented, but never all four streams flowing concurrently.

I attempted this with different P25 radios, different antennas on my RTL devive, high-power low-power etc, but still am not able to decode it. The flowgraph seem to be very finicky and requires fine adjusment. But it also clearly works, as shown in an older video of Balint's. He makes it look very simple.

 
World's cheapest P25 receiver w/ decryption: GNU Radio + OP25 + $20 RTL2832 DVB-T Dongle
https://www.youtube.com/watch?v=wShOLgW2tmI


Rob