InnerWeb
Skip to main content

FAQ: LoRa gateway capacity and number of watches

The IT department planning a watch fleet at a site asks first about the number: how many units one LoRa gateway can serve. The answer depends on the radio settings, the pulse period and whether every pulse is acknowledged. Below are the thresholds from the documentation and the signs that it is time to move to a campus gateway or LoRaWAN.

about 30
watches per channel at SF9, ACK and a 120 s period
Source: karta
30-300
watches after switching to SF7 and pulses without acknowledgement
Source: karta
about 0.2 %
airtime of a single watch at a 120 s pulse
Source: karta

How many watches does one LoRa gateway serve?

With the current settings, that is spreading factor SF9, an ACK for every pulse and a 120 s period, one channel serves about 30 watches. This is the reference value for a small site: a dozen or so people working in one zone fit comfortably.

An on-site gateway is designed for about 30-40 watches, and a campus gateway with an 8-channel concentrator for larger sites. Both types are served by the same backend, so changing the gateway does not force a change on the platform side.

How can capacity be increased without changing the infrastructure?

After switching to SF7 and a pulse without acknowledgement, the same channel serves from 30 to 300 watches. This is the largest gain without adding hardware, but it has a cost: SF7 has a shorter range than SF9, and a pulse without acknowledgement does not tell the watch that the frame arrived. Before the change, the IT department should therefore check the range at the furthest points of the site.

The second lever is the pulse period. The panel allows it to be set in the range 30-3600 s, with a default of 120 s, and the new value reaches the devices in response to the next pulse, without reflashing the fleet. A longer period reduces the load on the channel but slows the detection of position changes, so the value is chosen to match the type of work.

What about airtime?

The watch operates in the 869.4-869.65 MHz band with 500 mW ERP and a 10 % transmit duty cycle limit. With a 120 s pulse period, one device uses about 0.2 % of airtime, so the legal limit is not a constraint for a single watch. The constraint is the shared channel, on which the whole fleet transmits.

Capacity thresholds of a LoRa channel

Capacity thresholds of a LoRa channelBar chart: 30 watches per channel with current settings, 300 with SF7 without acknowledgement, 200 as the lower bound for considering LoRaWAN.0100200300watchesCurrent settings (SF9, ACK): 30 watches30Current settings (SF9, ACK)Upper threshold SF7 without acknowledgement: 300 watches300Upper threshold SF7 withoutacknowledgementLoRaWAN entry (lower bound): 200 watches200LoRaWAN entry (lower bound)
Source: Field card: channel capacity from the LoRa gateway documentation and the threshold for returning to the LoRaWAN option

What to do above 300 watches?

Above 300 units the documentation points to two routes: time slots or LoRaWAN. Time slots separate the transmissions of the watches in time so that they do not collide. LoRaWAN is a network standard that removes the job of organising the channel from the project.

LoRaWAN with ChirpStack is an option, not a decision. Small sites run on the IW v2 protocol, and the option returns when a site exceeds about 200-300 watches or when the client requires the standard. In the LoRaWAN variant, gateways based on the SX1302 or SX1303 concentrator forward frames to a ChirpStack server running locally, and a thin bridge converts them into calls to the existing API. The backend stays unchanged.

Does encryption change together with the network?

LoRa watch frames are encrypted with AES-128-CCM and replay counters. In the current version the network key is shared across the fleet. Per-device keys come only with the LoRaWAN option, so if the site security policy requires a separate key for each watch, that is an additional argument for this option, regardless of the number of units.

For the IT department this distinction matters: capacity and key security are two independent selection criteria. A small fleet can stay on the IW v2 protocol if the shared key fits the policy, while a large fleet can move to LoRaWAN because of capacity.

How can several areas be connected without adding internet gateways?

The repeater variant covers a harder-to-reach place without a new internet link. The watch transmits to a repeater gateway, which forwards the frame to a gateway with internet access. SOS remains deterministic and the backend needs no changes. The mesh variant is a scaling option, currently documented only, so no deployment is planned on it without a separate decision.

Good practice for IT: record the number of watches per area, the pulse period and the SF settings, then compare them with the thresholds of 30, 300 and 200-300 from this FAQ. Such a table is the basis for the conversation about the choice of gateway and for ordering the hardware.

Checklist

Video

Frequently asked questions

Is a campus gateway served by a different backend than an on-site one?

No. Both gateway types are served by the same backend. They differ in purpose: the on-site gateway is for about 30-40 watches, the campus gateway with an 8-channel concentrator for larger sites.

Does changing the pulse period require a visit to the watches?

No. The period is set in the panel and reaches the devices in response to the next pulse. The fleet does not need to be reflashed.

Does the 10 % transmit duty cycle limit a single watch?

No. With a 120 s period a watch uses about 0.2 % of airtime. The constraint is the total load of the whole fleet on the shared channel.

Are the repeater and mesh available straight away?

The repeater is a described variant in which SOS remains deterministic. Mesh is a scaling option that is currently only documented, so it requires a separate decision.

We will calculate the gateways for your site

Book a call with your IT department about the number of watches, the pulse period and the choice of gateway. We will configure it on your numbers.

Let's talk