Hi Bill
The minikits LPF7 has lands for sma sockets but you can also direct connect the coax.
The output arrangement is limited by the available coax already installed in the bg7sgm. We cut this off as close as possible to the circuit board and bared a new short section for attachment to the LPF7.
On the input side we initially tried a short section of coax directly soldered between the output of the bg7sgm and the input of the LPF7 but got disappointing results in terms of suppressing the 900mHz harmonic. My technical adviser attributed this to io couplng..
We then tried some different cable lengths and routing and installed the sma socket initially to make trying different cable lengths easier, but in fact the socket itself seems to help isolate the input from the output. In the end we settled on the illustrated arrangement as by experiment giving us acceptable results. Further experimentation could get further improvement.
The LPF7 is a low pass filter, it starts to cut in at 450mHz and is designed to provide -54dB at 900mHz, The unit we built was measured and pre-installation was found to meet this spec within a dB or so. In Aust 900mHz is a tdma mobile phone band so it is important to minimise potential interference at 900mHz..
In relation to higher density, a problem with DMR and the proliferation of cheap handsets is that the tdma pulsing paradigm has potential cause cross channel interference. We have examined the spectrum of a number of popular DMR handhelds and in some cases found significant spurs mostly immediately either side of the nominal channel frequency. I think tdma mobile phone networks can get away with this as the cell towers use power management to minimise interference but in an amateur setting it is a bit more of a free for all so cross channel interference potential could be greater. In such cases a bandpass filter rather than low pass filter might be more appropriate.
We have done some field testing comparing DMR, DSTAR and FM over similar long paths (about 20km) using an ID51 and Anytone 858 handset to handset.. We found FM is better than either of the digital modes in fringe situations (digital cliff problem). DMR and DSTAR are about on par. However, DMR appears to have a bit better punch through at the fringe but audio quality is compromised. It is a subjective judgement but most situations DSTAR audio quality generally seems to edge out DMR, possibly because it is a continuous carrier system using forward error correction and is less reliant on handshaking.
In making comparisons another consideration is that DSTAR is easier to amplify, standard 2 and 70cm linears give a good result. Using higher powers DSTAR has been demonstrated over DX length paths of hundreds of kms whereas DMR is hard limited to about 120km - the two systems in some respects don't directly compare, its horses for courses.
Jim VK2JHG