You would need enough dongles to cover the frequencies that the site uses.
RTL dongles can cover 2.4 MHz of contiguous spectrum. I believe from some of your previous posts that you're decoding the McHenry County Simulcast, correct? That site has frequencies running from 851.350 to 857.1125, which is approximately 5.8 MHz of contiguous spectrum. Most of the channels are grouped between 851.350 and 853.4375, which happens to fit within the bandwidth of a single dongle, but you would still need to cover the remaining four channels between 854.5375 and 857.1125.
For that site, you would essentially need three dongles to cover everything.
SDRTrunk looks at the entirety of the 2.4 MHz of spectrum for each dongle you allow it to use, and anything it finds related to the site within all of that bandwidth, it will decode. The key is that it never leaves the control channel to chase traffic (voice) channels. It continues decoding the control channel the entire time, while also decoding the traffic channels using the same dongles, as you would presumably provision enough dongles to cover all of the spectrum being used at the site, as discussed above.
In your situation, one dongle will lock onto the 851.7375 control channel and will never leave it, effectively decoding all call grants that come across the site. That same dongle will also decode all voice traffic between 851.350 and 853.4375, as it's all well within the 2.4 MHz of bandwidth. Note that the dongle doesn't really "tune" to the voice channels, as it's already looking at the entirety of the spectrum, but rather creates virtual "receivers" on the fly when a voice channel is active so that it can capture and decode that small slice of spectrum that the voice channel occupies while active. The other two dongles will be available to cover the rest of the spectrum above 854 MHz.
We've veered pretty far off topic here, but to get back around to my original post...once you get SDRTrunk up and running, you can then create streaming profiles for Calls and/or traditional feeds, and are able to assign any talkgroup(s) you want to each streaming profile. You could send every talkgroup to Calls, but also have three traditional feeds for Fire North, Fire West, and Fire South each of which have only the appropriate talkgroups assigned to each. SDRTrunk will take care of keeping everything separate and will feed the audio to each using the audio that it's decoding anyway.
