Why IoT Antennas Matter

The Part Everyone Overlooks

There is a predictable pattern in industrial IoT deployments that go wrong on connectivity. The project specifies a good cellular router. The SIM supplier provides a tested M2M tariff on a well-covered network. The VPN is configured. The SCADA or monitoring application is connected. Everything tests fine on the bench. Then the equipment arrives on site, goes into a metal cabinet, and the link quality is marginal, intermittent, or in some cases non-existent.

The cause, in the vast majority of cases, is the antenna. Either the stub antenna shipped with the router has been left in place inside a metal enclosure where it cannot radiate effectively. Or an external antenna has been specified but the wrong type for the installation. Or the cable run between the antenna and the router is long enough and low-quality enough to eat the signal before it reaches the modem. Or the antenna is mounted in a location that made installation convenient rather than one that maximises signal.

None of these are exotic problems. They are routine. And they are all preventable with a basic understanding of how cellular radio works, what an antenna actually does, and how installation decisions affect the outcome.

This matters more as deployments scale. A connectivity problem on a single device is an annoyance. The same problem across 50 substations, 200 remote monitoring sites, or a fleet of thousands is an operational and commercial disaster. Getting antenna specification right at the design stage is significantly cheaper than diagnosing and retrofitting at scale.

What an Antenna Actually Does

A cellular antenna does not create signal. It does not amplify signal in the way that an amplifier does – that requires external power. What an antenna does is focus the energy it transmits or receives in particular directions, and it presents the radio frequency energy from the cable to the air (and vice versa) as efficiently as possible.

A stub antenna – the short rubber-covered rod that ships with most routers – is a compromise. It is physically small, mechanically convenient, and electrically adequate in a good signal environment. It radiates in a roughly omnidirectional pattern in the horizontal plane and provides modest gain – typically 2 to 3 dBi. In a strong signal environment with an unobstructed radio path, it works. In a challenging environment – inside a metal cabinet, in a weak coverage area, through reinforced concrete, or where interference is present – it fails the deployment.

An external antenna improves the situation in two ways. It gets the radiating element outside of whatever enclosure the router is sitting in, removing the attenuation effect of the metal case. And a better-specified antenna provides more gain – focusing more of the transmitted and received energy in useful directions, effectively making the link stronger without consuming any additional power.

Understanding Signal Quality Metrics

Before improving antenna performance it is worth understanding how to measure it. Modern cellular modems report several signal quality metrics. RSSI is the one most people encounter first – it is simple, widely reported, and largely useless as a standalone quality indicator. The metrics that actually matter are RSRP, RSRQ, and SINR, which are specific to 4G LTE.

Reading these metrics from a Milesight router is straightforward – the Status > Cellular page in the web GUI reports RSRP, RSRQ, SINR, and RSSI for both SIM slots in real time. When diagnosing an antenna or installation problem, these are the numbers to look at. A device showing RSRP of -105 dBm and SINR of 3 dB needs an antenna improvement. A device showing RSRP of -85 dBm and SINR of 18 dB has a healthy link regardless of what the RSSI says.

Antenna Gain – What dBi Actually Means

Antenna gain is measured in dBi – decibels relative to an isotropic radiator. An isotropic radiator is a theoretical antenna that radiates equally in all directions simultaneously. Real antennas cannot do this – they concentrate energy in some directions at the expense of others. The gain figure tells you how much stronger the radiation is in the antenna’s preferred direction compared to the theoretical isotropic reference.

A 3 dBi antenna provides roughly double the effective radiated power in its favoured direction compared to the isotropic reference. A 6 dBi antenna provides four times. A 9 dBi antenna, eight times. Each 3 dB increase approximately doubles the effective radiated power – and 3 dB is also the point at which a human ear notices a clear difference in loudness. In radio terms, 3 dB improvement in signal level at the receiver is genuinely meaningful.

The trade-off is directionality. Higher-gain antennas focus their energy into a narrower beam. A 3 dBi omnidirectional antenna radiates in a roughly spherical pattern around the horizontal plane – useful when the direction to the base station is unknown or varies. A 9 dBi directional antenna focuses into a much narrower cone – more gain in the target direction, but needs to be aimed correctly and provides no gain at all in other directions.

MIMO – Why Modern Cellular Needs Multiple Antennas

MIMO stands for Multiple Input Multiple Output. It is a radio technology that uses multiple antenna elements simultaneously to either increase data throughput, improve signal reliability, or both. Modern 4G LTE networks use 2×2 MIMO as standard – two transmit and two receive antenna paths operating simultaneously. 5G networks extend this to 4×4 MIMO and beyond.

The two antenna elements in a 2×2 MIMO system are not simply redundant backups for each other. They exploit a property of radio propagation called multipath – the fact that radio signals bounce off buildings, ground, and objects and arrive at the receiver from multiple directions. A MIMO system uses the slight differences between what each antenna element sees to separate multiple data streams, effectively multiplying the throughput of the radio link without requiring additional spectrum or transmit power.

For this to work, the two antenna elements need to be physically separated – typically by at least half a wavelength, which at 4G LTE frequencies (800 MHz to 2600 MHz) is somewhere between 6 and 19 centimetres. They also need to have different polarisation or orientation so they sample the incoming signal differently. A single stub antenna, or two stub antennas mounted right next to each other, cannot exploit MIMO properly – the two antenna inputs to the modem are seeing essentially the same signal, and the MIMO gain is lost.

This is why MIMO antennas – devices containing two (or four) physically separated and independently connected radiating elements in a single housing – matter for anything beyond basic connectivity. A well-specified 2×2 MIMO antenna in a good installation can deliver 30 to 50 percent better throughput than a single element, and significantly better link stability in environments with strong multipath. For CCTV, high-frequency SCADA, and any application where throughput matters, MIMO antenna specification is worth the modest additional cost.

Antenna Ports and MIMO Configuration

Industrial routers list their antenna configuration in their specifications. A Milesight UR35 has two SMA antenna ports for the cellular modem – these are the two MIMO paths. For full MIMO operation, both ports need to be connected to antenna elements that are adequately separated. Connecting one port to an external antenna and leaving the other on an internal stub partially defeats the purpose. Connecting both to a single dual-element MIMO antenna – one cable to each port – is the correct configuration.

5G routers like the Milesight UR75 have four cellular antenna ports – supporting 4×4 MIMO on 5G bands. The same principle applies: all four ports connected to a 4-element MIMO antenna for full capability. In practice, many 5G router installations use a 2×2 MIMO antenna on the primary 5G bands and accept reduced MIMO performance – this is a pragmatic compromise that is often acceptable given installation constraints, but the performance difference is real and worth understanding.

Cable Loss – The Silent Signal Killer

Every metre of coaxial cable between the antenna and the router attenuates the signal. At 4G LTE frequencies, the loss per metre depends on the cable type – thin RG174 cable loses roughly 1.5 dB per metre at 1 GHz. Standard RG58 loses around 0.5 dB per metre. Low-loss LMR-195 equivalent loses around 0.35 dB per metre. LMR-240 equivalent, around 0.22 dB per metre at 1 GHz.

The practical implication is straightforward. A 5-metre run of RG174 cable loses around 7.5 dB at 1 GHz – more than the gain of most external antennas. That external antenna installation has made things worse, not better, by adding cable loss that outweighs the antenna gain. The correct specification for a 5-metre run is LMR-195 or better cable, which loses around 1.75 dB over the same distance – a net gain compared to the stub antenna it replaces.

Every connector in the run also adds loss – typically 0.1 to 0.3 dB per connector. N-type connectors perform better than SMA at high frequencies, which is why longer external antenna runs use N-type at the antenna end and adapt to SMA at the router. Minimising the number of connectors in any cable run is worth doing.

Antenna Types and When to Use Each

Omnidirectional External Antenna

The standard external antenna for most industrial IoT deployments. Radiates in a roughly horizontal donut pattern – good coverage in all compass directions, reduced gain directly above and below. Available in single-element (for single-antenna connections) and dual-element MIMO versions. Gains typically range from 3 to 9 dBi – higher gain versions have a narrower vertical beam, which matters if the base station is significantly higher or lower than the device. For most ground-level or rooftop installations, 5 to 7 dBi is the appropriate range.

Directional Panel Antenna

A directional panel antenna focuses gain into a sector – typically a beam width of 60 to 90 degrees horizontally and 30 to 60 degrees vertically. Gains of 9 to 15 dBi are common. The correct application is where the direction to the serving base station is known and consistent – a rural site with a single mast visible in a specific direction, or a private LTE deployment where the base station location is fixed. Directional antennas provide much higher gain than omnidirectional equivalents of the same physical size, but only in the target direction. Misaim them and performance can be worse than a stub antenna. Do not use directional antennas when the serving cell may switch between base stations in different directions.

MIMO Panel Antenna

A single enclosure containing two or four physically separated antenna elements for MIMO operation. The elements are arranged to provide the spatial diversity that MIMO requires – typically orthogonal polarisation (one element vertical, one horizontal) and physical separation within the housing. A 2×2 MIMO panel antenna connects to two router antenna ports and provides full MIMO operation in a single mounting point. For outdoor or elevated installations where a compact single-mounting-point solution is preferred over two separate antennas, MIMO panels are the right choice.

Embedded Antenna

An antenna integrated into the device enclosure or PCB, with no external connector. Common in low-cost IoT sensors, trackers, and consumer devices. Performance is limited by the physical constraints of embedding an antenna into a small enclosure, and is highly sensitive to nearby metal, moisture, and the human body. For industrial deployments in metal enclosures, embedded antennas are generally not appropriate – the enclosure itself defeats them. They are appropriate for small plastic-cased devices in benign environments.

Vehicle and Magnetic Mount Antenna

Magnetic mount antennas attach to the vehicle roof and use the metal vehicle body as a ground plane, which actually improves their performance significantly – the ground plane is a key part of the antenna system for this type of design. For vehicle telemetry, fleet tracking, and mobile IoT applications, a magnetic mount or through-hole vehicle antenna provides excellent performance. The same antenna bolted to a plastic surface or used without a ground plane will perform much worse than its specification suggests.

Installation Height and Placement

Signal propagation follows the inverse square law – doubling the distance between transmitter and receiver reduces signal power to a quarter. Antenna height partially counteracts this by increasing line-of-sight range. An antenna at 10 metres above ground level has a radio horizon roughly 11km away. At 3 metres, that drops to 6km. In areas where the serving base station is distant, getting the antenna higher can make the difference between a marginal and a reliable link.

Height also matters for avoiding ground-level obstructions. Buildings, vegetation, and terrain features that absorb or diffract signal between the device and the base station can cause 5 to 20 dB of additional path loss that a few metres of antenna height can eliminate. For outdoor cabinet installations, mounting the antenna at the top of the cabinet or on a short mast above it is worth doing regardless of whether the signal is currently adequate – the resilience under adverse conditions (wet vegetation, temporary obstructions) improves.

For LoRaWAN gateway antennas – which need to cover sensors across a wide area rather than reach a single base station – height is even more critical. A Milesight UG65 gateway mounted at 2 metres covers a radius of perhaps 500 metres to 1km in an urban environment. The same gateway mounted at 10 metres on a rooftop may cover 3 to 5km. The coverage difference is orders of magnitude for a relatively modest installation change.

Antenna Separation and Interference

A device with both a cellular antenna and a WiFi or LoRaWAN antenna needs adequate separation between them to avoid co-site interference. Cellular and WiFi operate on different frequency bands – 4G LTE at 700-2600 MHz, WiFi at 2.4 GHz and 5 GHz – but proximity between antennas can still cause intermodulation products and receiver desensitisation that degrade link quality.

The general rule is a minimum of 20 to 30 cm separation between cellular and WiFi antennas, and at least 10 cm between two cellular MIMO elements. For GPS antennas – relevant on devices using GPS-disciplined NTP or asset tracking – separation from cellular antennas of at least 15 cm reduces GPS receiver noise floor impact. These are not hard boundaries but practical guidelines based on common deployment experience. Tighter spacing is manageable with good antenna design; the problems emerge with cheap antennas in close proximity in high-interference environments.

How Installation Environment Changes Everything

Verifying Antenna Performance After Installation

The only way to confirm an antenna installation is working correctly is to measure signal quality metrics after installation and compare them with a baseline. The baseline is the signal with a stub antenna or before the external antenna was connected. The post-installation metrics should show improved RSRP and SINR values.

For Milesight routers, signal metrics are available via the web GUI in real time and via the Milesight Development Platform dashboard for remote monitoring across a fleet. Setting up RSRP and SINR monitoring as platform alerts – triggering a notification when RSRP drops below -105 dBm or SINR drops below 5 dB – provides ongoing visibility of antenna health across all deployed devices without manual checking. Signal degradation that develops over time (a cable connector corroding, vegetation growing in front of the antenna) will be caught by an alert before it becomes a connectivity outage.

For large fleet deployments, logging signal metrics over time and comparing the distribution across sites quickly identifies outliers that need investigation – sites where signal is worse than typical for the coverage area, which usually indicates an antenna or installation problem rather than a genuine coverage issue.

The Economics of Getting It Right

A good external MIMO antenna costs between £20 and £80 depending on specification. The cable and connectors for a typical cabinet installation add another £15 to £40. Call it £100 total in antenna hardware per site. That compares to the cost of a return site visit to investigate and fix a connectivity problem – typically several hundred pounds once travel, labour, and downtime are accounted for – and the cost of the downtime itself, which in utility and critical infrastructure applications can be significant.

Specifying the antenna correctly at project design stage is not an optional refinement. It is the difference between a deployment that works reliably and one that generates a long tail of support calls, site visits, and customer complaints. For any deployment where connectivity reliability matters – which is most industrial IoT deployments – antenna specification deserves the same attention as hardware and SIM selection.