Trunked system misbehaving? Here's one reason why, maybe.

A Capacity Max (DMR Tier 3) system uses a control channel to process calls and other traffic on a site. The control channel also has another purpose: radios will use the received signal strength (RSSI) from the control channel, to determine whether this site can be used or not (a.k.a. roaming).

Similarly, a Multisite Capacity Plus system uses the Rest Channel to setup calls and this is also used by radios measuring the signal strength as part of their roaming process.

In both system topologies, the RSSI of the Rest Channel or Control Channel will have an effect on the roaming behavior of the radios.

In a system that requires a RF site to have an odd number of channels; with less than a few hundred kilohertz separation between the TX frequencies, a hybrid combiner may have to be used. Since the number of channels is not an even number (in this example 3), the combiner would look something like this.

The insertion loss per hybrid, in such a setup, is around 3dB. Therefore the system loss for Repeater 1-2 and Repeater 3 will each be different - this in turn will affect the TX power.
If all the repeaters are set to 25W, the measured power at the antenna will be around 6W when either repeater 1 or 2 are transmitting. When channel 3 is transmitting, the measured TX power will be around 12W [Ignoring all other losses!]
In this case, if the control channel is on channel 3, the radios will get a different RSSI reading than when the control channel is on either of the other two channels. For example, a VHF portable 25km away, might see -110dBm when the control channel is on channel 3 and -115dBm when the control channel is on either of the other two repeaters.

This may not seem so significant but consider a possible scenario where the above VHF portable can receive signal from three identical RF sites. Let's assume that the RSSI from all three sites is the same (so -110dBm) and that the control channel is on repeater 3.
Although rare, in a case like this, the radio would have to apply other criteria to choose which site to move to, if needed.
If the control channel or rest channel, on one of these sites moved to channel 1 or 2, how would the roaming behavior change?

Assuming the control channel (in the case of a Capacity Max system) is on channel 3, and a call setup requires the radio to go to one of the timeslots on channel 1 to participate in the call. While idle, the radio would see -110dBm but -114dBm while in the call.
This might make the radio roam to another site during the call. If there is no there site to roam to, the radio user might notice some bit errors (i.e. audio distortion; gurgling and holes).

The same might be seen on a Capacity Plus system, when the rest channel rolls over to the next physical channel.

There are two ways to solve this:

1. Turn the power of repeater 3 down to 12W, so that the RF power for all repeaters is 6W when measured on the output of the combiner.
2. Redesign the combiner system such that the insertion loss on all inputs is the same. One way to do this is to simply make the combiner a four-input model - this allows for future expansion. (see below)

Something like this?

I've used a hybrid combiner here but potentially, the same could occur if a cavity combiner was used. Although, as far as I know, the insertion loss on all inputs of a cavity combiner is equal, no matter how many of them (inputs) there are.



PS. As a rule of thumb, if the TX-TX separation is less than 300kHz, a hybrid combiner would be the better choice. You should, of course, contact the manufacturer during the design phase to confirm this.

PPS. Sometimes, during the draft design phase, it makes sense to refarm frequencies to allow a cavity combiner to be used. These may be bigger but they cost less and have a lower insertion loss.