Is the "200 radios per repeater" figure realistic, and at what call rate does it break?

It is realistic only for very light users. Work backwards from the number and it implies each of those 200 radios transmits roughly three times an hour, for about five seconds a time. That is a plausible load for a quiet office or a small hotel on a slow shift. On a construction site, or a security team working an incident, the same repeater supports closer to 25 radios before the channel is busy more often than not. The figure is not wrong, it is quoted without the call rate that produced it.

Why a radio count alone cannot answer the question

Channel capacity is a traffic problem, not a headcount problem. What loads a repeater is not how many radios are programmed to it but how much time those radios spend keying up. The unit for that is the erlang: one erlang is one talk path occupied continuously for one hour.

A user's offered traffic is simply their transmissions per hour multiplied by the average length of a transmission, divided by 3,600 seconds. Three classes of user cover most buildings we survey:

• Light user (office, quiet reception, night-shift guard): 2 transmissions per hour at 5 seconds = 10 s/hour = 0.0028 erlangs

• Moderate user (hotel housekeeping, facilities, retail duty manager): 6 transmissions per hour at 5 seconds = 30 s/hour = 0.0083 erlangs

• Heavy user (construction site during a lift, security during an incident): 20 transmissions per hour at 6 seconds = 120 s/hour = 0.0333 erlangs

The spread between the light and heavy user is roughly twelve to one. Any single number for radios per repeater is therefore meaningless within a factor of twelve unless it names which of these it assumed.

How many radios one repeater actually supports

A conventional analogue repeater carries one talk path. A DMR repeater using both TDMA timeslots carries two. In land mobile radio we do not design to a telephony-style blocking figure, because a busy push-to-talk channel does not fail a call — the user hears the channel is occupied, waits, and presses again. The practical criterion is busy-hour channel occupancy, and past roughly 40% the waiting becomes noticeable enough that users start talking over each other and repeating themselves.

Taking 40% occupancy per talk path as the design ceiling, a two-slot DMR repeater carries about 0.80 erlangs of offered traffic in the busy hour, and a single-path analogue repeater about 0.40. Dividing by the per-user figures above: a light user population supports about 143 radios on a single-path analogue repeater and about 286 on a two-slot DMR repeater; a moderate population about 48 and 96; a heavy population about 12 and 24.

So 200 radios on one DMR repeater sits between the light and moderate cases. Solving it directly: 0.80 erlangs is 2,880 seconds of channel time per hour, and shared across 200 radios that is 14.4 seconds each — about three five-second transmissions per user per hour. That is the call rate the vendor figure assumes, and it is almost never stated alongside it.

At what call rate does it break

At roughly three to four short transmissions per user per hour, 200 radios works. At six per hour, which is an ordinary hotel operations load, the same repeater supports about 96 and the 200-radio system is running at double its comfortable occupancy. At construction-site rates it supports about 24. The break point is not gradual: because every user's retry adds traffic of its own, a channel that is merely busy at 50% occupancy degrades quickly into one where half the presses meet a busy channel.

Two things make it worse in practice. GPS location updates and text data share the same timeslots as voice on a DMR system, so a fleet reporting positions every 30 seconds consumes capacity that never appears in a radio count. And talkgroups do not create capacity — splitting 200 users across six talkgroups on one repeater leaves the same two talk paths carrying the same traffic.

How to size it properly

Measure rather than assume. A repeater's own logging, or a week of channel occupancy recording, gives busy-hour transmission counts and average transmission length per talkgroup. From those two numbers the figures above are arithmetic. If busy-hour occupancy is already above 40%, the answer is another channel or a move to DMR trunking, not better coverage — and it is worth confirming the problem is capacity rather than a weak uplink before spending anything, since a channel that sounds busy and a channel nobody can reach present very similarly to users.

Related questions

Does DMR two-slot TDMA really double capacity?

It doubles the number of simultaneous talk paths from one to two, which roughly doubles the traffic the channel carries at a given occupancy. It does not double the range, and it does not help if the load is data rather than voice contention.

Is busy-hour occupancy the same as blocking probability?

No. Blocking probability is a telephony measure that assumes a rejected call is lost. Push-to-talk users retry instead, so occupancy is the more useful criterion, and 40% is a stricter working target than it sounds.

Will adding talkgroups fix a congested channel?

No. Talkgroups partition who hears whom, not how much airtime exists. Only additional channels, or trunking across several channels, add capacity.

Suneast sizes radio systems from measured busy-hour traffic rather than a radio count. We log channel occupancy and transmission length on the existing system before quoting, so the channel count in a proposal is traceable to how the site actually talks, and we say plainly when a coverage complaint is really a capacity problem. Related reading: how many radios and users one repeater channel supports, and why a radio's advertised coverage rating does not translate into floors covered — the same class of error, a real figure quoted outside the conditions that produced it.