There is a conversation that happens on every IoT deployment, in every site survey, on every support call where signal quality comes up. It goes something like this:
“The signal is fine. I’m getting -70.”
“RSSI or RSRP?”
Silence.
That silence – that moment where the person on the other end of the call realises they have been quoting a number without being entirely sure what it measures – is the gap this article exists to close.
I have spent over 25 years in cellular IoT and M2M. I have seen antennas specified on the basis of RSSI readings that told only a fraction of the story. I have watched engineers spend significant money on high-gain external antennas, fit them to routers, and then wonder why throughput barely improved – because the problem was never signal strength in the first place. It was interference. And they were not looking at the right metric.
This is not a beginner’s glossary. There are plenty of those. This is an attempt to give you a working understanding of what each metric actually represents, where people go wrong, and how to use this knowledge to make better decisions about connectivity, antenna selection, and deployment.
Why Signal Metrics Matter More Than People Realise
Modern cellular networks – LTE and 5G NR – are genuinely complex radio environments. The signal arriving at your device’s antenna port is not just “strong” or “weak.” It carries information, it competes with other signals, it gets reflected off buildings, it changes depending on the time of day and the number of nearby devices connected to the same cell.
The metrics we use to describe that signal are not interchangeable. They measure different things. They tell you different stories. And if you use the wrong metric to diagnose a connectivity problem, you will almost certainly reach the wrong conclusion.
The four metrics that come up most often in LTE and 5G IoT deployments are:
- RSSI – Received Signal Strength Indicator
- RSRP – Reference Signal Received Power
- RSRQ – Reference Signal Received Quality
- SINR – Signal to Interference plus Noise Ratio
Each one is a layer of understanding. Used together, they paint a full picture of what is happening in your radio environment.
RSSI: The Number Everyone Quotes and Barely Anyone Understands
What people say
“We have good signal. The RSSI is -65 dBm.”
What they think it means
That the device is receiving a strong, usable signal from the network. Closer to zero equals better. Simple.
What it actually measures
RSSI is the total power received across the entire channel bandwidth at the antenna port. That sounds specific, but here is the critical detail: total power means everything. The signal you want, background noise, interference from adjacent cells, interference from other systems operating nearby – all of it, added together.
RSSI is an inherited metric. It comes from the 2G/3G era when radio environments were simpler and when the things competing for spectrum were fewer. In those days, measuring total received power was a reasonable proxy for signal quality because noise and interference were relatively predictable.
In dense LTE and 5G deployments, it is far less useful. A strong RSSI reading can mask a terrible signal environment. You might be receiving significant total power, but the proportion of that power that is actually your network’s useful signal could be small. The rest is noise and interference from other sources.
RSSI is expressed in dBm (decibels relative to one milliwatt). As with all these metrics, the scale is negative – 0 dBm would represent one milliwatt of received power, which in a cellular context would be extraordinarily strong. In practice you are dealing with values from around -50 dBm (very strong, close to a cell tower) down to -110 dBm or below (very weak, edge of coverage).
The reference scale people use (and its limitations)
| RSSI Range | Common Label |
|---|---|
| -50 to -65 dBm | Excellent |
| -65 to -75 dBm | Good |
| -75 to -85 dBm | Fair |
| -85 to -95 dBm | Poor |
| Below -95 dBm | Very Poor |
These ranges get quoted everywhere. They are not wrong exactly, but they are incomplete. An RSSI of -72 dBm in a quiet rural environment with a clear line of sight to a single cell tower is a completely different situation to an RSSI of -72 dBm in an industrial estate surrounded by machinery, competing cells, and RF noise sources. Same number. Very different actual connectivity.
What RSSI is actually useful for
RSSI gives you a quick, directional indicator of received power. It is useful for coarse comparison – is this location better than that one? It is useful for antenna orientation where you are trying to find the direction that maximises received power. It is a starting point, not a conclusion.
Do not use RSSI alone to assess whether a site has adequate signal for a demanding IoT application. You will sometimes get it right and sometimes get it badly wrong.
RSRP: The Metric You Should Actually Be Using First
What people say
“My RSRP is -95. Is that okay?”
What they think it means
Often, not much. RSRP is less intuitively understood than RSSI because it requires knowing a little about how LTE signals are structured. Many people treat it as “another signal strength number” and reach for the same good/bad thresholds they use for RSSI.
What it actually measures
RSRP – Reference Signal Received Power – measures the average power of the Reference Signals within a given LTE or 5G NR carrier. Reference Signals are specific, predefined pilot tones that the network transmits at known power levels. Your device uses them for synchronisation, channel estimation, and handover decisions.
Because RSRP is measuring a known, specific signal rather than total broadband power, it is far more meaningful than RSSI in a modern radio environment. It isolates the useful signal from the noise and interference. It tells you how much of the network’s reference signal is actually reaching your device.
RSRP is what your device and the network actually use to make decisions. Handover thresholds, cell reselection, beam selection in 5G NR – these are all driven by RSRP. If you want to understand how the network sees your device’s situation, RSRP is the metric to start with.
RSRP reference scale
| RSRP Range | Typical Assessment |
|---|---|
| Above -80 dBm | Excellent |
| -80 to -90 dBm | Good |
| -90 to -100 dBm | Fair / Acceptable |
| -100 to -110 dBm | Poor |
| Below -110 dBm | Very Poor / Edge of Coverage |
The relationship between RSSI and RSRP
You can often observe the relationship between RSSI and RSRP and learn something from it. In a clean radio environment with low noise and interference, RSSI and RSRP will track relatively closely – the total received power is dominated by the useful signal. In a noisy or heavily loaded environment, RSSI will read higher relative to RSRP – you are receiving a lot of total power, but the reference signal component of it is a smaller proportion.
If your RSSI looks decent and your RSRP looks poor, you are almost certainly in a high-interference environment. That changes the solution completely.
What RSRP tells you about antenna choice
RSRP responds well to antenna gain improvements in scenarios where the primary problem is path loss – where your device is simply too far from the cell, or there is significant attenuation due to building materials, terrain, or other physical barriers. A high-gain directional antenna pointed at the serving cell will improve RSRP.
RSRP is less responsive to antenna improvements in scenarios dominated by interference or noise. You can add all the antenna gain you want, but if the problem is that you are surrounded by strong signals from multiple competing cells, your RSRP improvement from a better antenna will be modest. What you actually need in that case is to address the interference – and that requires looking at RSRQ and SINR.
RSRQ: The Metric That Reveals Interference
What people say
“I never look at RSRQ.”
What they think it means
When engineers do look at RSRQ, they often treat it as a quality score – higher (less negative) is better, lower (more negative) is worse. Which is true, but it misses the specific information RSRQ contains.
What it actually measures
RSRQ – Reference Signal Received Quality – is defined as:
RSRQ = (N x RSRP) / RSSI
Where N is the number of resource blocks in the carrier bandwidth.
In plain terms: RSRQ expresses the proportion of total received power that comes from the reference signal, scaled by the bandwidth. It is a ratio. A high RSRQ (less negative, closer to zero) means the reference signal makes up a large proportion of what is being received. A low RSRQ (more negative) means the reference signal is a small proportion of the total – which means interference and noise are contributing significantly.
RSRQ is the metric that reveals interference. When RSRP looks reasonable but RSRQ is poor, you are seeing a situation where the serving cell’s signal is present but competing with substantial interference from other sources. That interference could be other LTE carriers on the same or adjacent bands, it could be industrial RF noise, it could be simply being in a dense urban environment with many competing cells all audible at similar signal levels.
RSRQ reference scale
| RSRQ Range | Typical Assessment |
|---|---|
| Above -10 dB | Excellent |
| -10 to -15 dB | Good |
| -15 to -20 dB | Poor |
| Below -20 dB | Very Poor |
The diagnostic value of RSRQ
The combination of RSRP and RSRQ tells you something neither metric tells you alone.
If both RSRP and RSRQ are poor: your device is at the edge of the cell’s coverage. Path loss is the problem. You need more received power – better antenna gain, better antenna placement, possibly a different carrier or frequency band.
If RSRP is reasonable but RSRQ is poor: you have an interference problem. The signal is getting to you but so is a lot of other stuff. Adding antenna gain alone will not solve this because a higher-gain antenna will also pick up more of the interfering signals. In this scenario, you need a more directional antenna to improve the signal-to-interference ratio, or you need to address the interference source.
If RSRP is good and RSRQ is good: your radio environment is clean. If throughput is still poor, the issue is elsewhere – network congestion, device configuration, application-layer problems.
What people get wrong about RSRQ and antennas
This is one of the most expensive mistakes I see in IoT deployments. A site has poor RSRQ – say, -18 dB. Someone looks at that number, concludes that signal quality is poor, and specifies a high-gain omnidirectional antenna to boost everything. They fit it. RSRQ barely improves because an omnidirectional antenna picks up interference from all directions, not just from the serving cell. The fundamental problem – too much interference relative to useful signal – is unchanged or even worsened.
The correct response to poor RSRQ is usually a directional antenna oriented toward the serving cell. By narrowing the antenna’s field of view, you reduce the amount of interference entering the receive chain while keeping the gain on the serving cell’s signal. RSRQ improves because you have changed the ratio, not just the absolute power.
SINR: The Number That Actually Predicts Your Throughput
What people say
“What does SINR mean again?”
What they think it means
SINR is probably the least understood of the four metrics, despite being arguably the most directly useful for predicting real-world performance. People who know it at all often conflate it with RSRQ, treating them as roughly equivalent quality metrics.
They are not the same thing. They measure in different ways and tell you different things.
What it actually measures
SINR – Signal to Interference plus Noise Ratio – expresses how much stronger the wanted signal is compared to the combined effect of interference and thermal noise. It is expressed in decibels, and unlike RSRP and RSRQ, higher is unambiguously better.
SINR = Signal Power / (Interference Power + Noise Power)
A positive SINR means your signal is stronger than the combined interference and noise floor. A negative SINR means the noise and interference are overwhelming your signal – and in that situation, reliable communication becomes very difficult regardless of what RSRP looks like.
SINR is what the network’s modulation and coding scheme (MCS) selection is based on. The higher your SINR, the higher the modulation order the network can use – moving from QPSK through 16-QAM, 64-QAM, up to 256-QAM in good LTE conditions. Higher modulation means more bits per symbol, which means more throughput. SINR is the metric that connects your radio environment directly to your achievable data rate.
SINR reference scale
| SINR Range | Typical Assessment | Typical Implication |
|---|---|---|
| Above 20 dB | Excellent | High-order modulation, maximum throughput |
| 13 to 20 dB | Good | Strong performance |
| 0 to 13 dB | Fair | Reduced throughput, lower MCS |
| -3 to 0 dB | Poor | Very limited throughput |
| Below -3 dB | Very Poor | Connection marginal or failing |
Why SINR is the metric that tells you whether your antenna is actually working
When you fit a new external antenna to an IoT router and want to know whether it has made a meaningful difference, SINR is the most telling number to watch – alongside actual throughput measurements.
A scenario where RSRP improves by 5 dB but SINR barely moves is a scenario where you have added gain but not meaningfully changed the interference picture. Your device is hearing the serving cell more clearly, but it is also hearing everything else more clearly, and the ratio between them has not shifted enough to change the network’s modulation decisions.
A scenario where SINR improves significantly after fitting a directional external antenna is a scenario where the antenna is doing exactly what it should – increasing the proportion of useful signal relative to noise and interference. That improvement will show in throughput.
SINR and 5G NR
In 5G NR, SINR becomes even more important because of the use of massive MIMO and beamforming. The network is trying to direct energy toward your specific device. The quality of that beam – how well-formed it is, how much it concentrates signal energy in your direction while suppressing interference – is reflected in your SINR. Deployments where 5G is available but SINR is poor often point to the device being in a challenging position relative to the antenna array on the base station, or to significant multipath and interference in the environment.
Reading All Four Together: A Diagnostic Framework
Individual metrics in isolation are limited. The pattern across all four metrics is where the real diagnostic power lies.
Scenario 1: Low RSRP, Low RSRQ, Low SINR
This is a straightforward range problem. The device is too far from the serving cell, or there is significant physical attenuation. The solution is to improve received signal power. A high-gain external antenna with good elevation toward the serving cell, combined with a quality low-loss cable run, will typically help. If RSRP is below -110 dBm and you are already using a reasonable external antenna, you may be looking at a coverage gap that requires a different network operator, a different frequency band, or a signal repeater solution.
Scenario 2: Reasonable RSRP, Poor RSRQ, Poor SINR
This is an interference-dominated scenario. The serving cell’s signal is present but heavily competed by interference. Fitting a directional antenna oriented toward the serving cell will often make a significant difference. In dense urban environments or industrial sites with significant RF activity, this type of scenario is extremely common. It is also the scenario where the wrong antenna choice is most expensive.
Scenario 3: Good RSRP, Good RSRQ, Good SINR, Poor Throughput
If the radio metrics are healthy and throughput is still disappointing, the issue is not in the radio environment. Possible causes include: network congestion on the serving cell, incorrect APN configuration on the device, QoS policies applied by the SIM operator, device processing limitations, or application-layer inefficiencies. Adding a better antenna will not help here.
Scenario 4: Good RSRP, Decent RSRQ, Good SINR, Inconsistent Throughput
Inconsistency with otherwise reasonable metrics often points to cell congestion during peak periods, or to the device frequently camping on a marginal cell and periodically switching. Check whether the device is locking to a specific cell, what frequency band it is operating on, and whether the inconsistency correlates with time of day.
Scenario 5: Good RSRP and RSRQ, But Poor SINR
This typically indicates significant interference that is elevating both total received power and reference signal power, but in a way that RSRQ’s ratio does not fully expose. Very dense cell environments can produce this pattern. Check whether the device is being served by an appropriate cell for its location, and whether a closer or less congested cell might offer better SINR.
What This Means for Antenna Selection and Deployment
The first question is not “how much gain?”
The first question should be: what is my radio environment, and what is the dominant problem?
- If the problem is range (low RSRP): omnidirectional gain helps by receiving more energy from the serving cell
- If the problem is interference (good RSRP, poor RSRQ/SINR): a directional antenna oriented toward the serving cell will typically outperform an omnidirectional even at lower peak gain, because it reduces interference entering the system
- If the problem is multipath: cross-polarised MIMO antennas help because they resolve multipath components more effectively
Cable loss is the enemy of everything
Every decibel lost in the cable run between the antenna and the router’s antenna ports is a decibel of antenna gain that never reached the device. LMR-195 equivalent cable runs approximately 0.2 dB/m loss at 700 MHz and significantly more at 2.6 GHz and above. A 3-metre run on standard RG58-type cable at 2.6 GHz can represent 2 dB or more of loss. Use the shortest run you can achieve and use quality low-loss coaxial where runs exceed 1 metre.
MIMO requires independent paths
For MIMO to deliver its throughput benefits, the antenna paths need to be sufficiently decorrelated. The two or four antennas serving the MIMO paths need to be physically separated, differently oriented, or both, such that the signals arriving on each path are substantially independent. Properly specified MIMO antennas provide ports designed to maintain the decorrelation needed for MIMO gains to materialise in real deployments.
Quick Reference
| Metric | What it Measures | Unit | Better Direction | Primary Use |
|---|---|---|---|---|
| RSSI | Total received power (signal + noise + interference) | dBm | Higher (less negative) | Quick orientation, coarse comparison |
| RSRP | Power of LTE/5G reference signals specifically | dBm | Higher (less negative) | Coverage assessment, cell selection |
| RSRQ | Ratio of RSRP to total received power | dB | Higher (less negative) | Interference diagnosis |
| SINR | Signal power relative to interference plus noise | dB | Higher | Throughput prediction, MCS estimation |
Related: Before you act on any of these readings – or before you spend money on an external antenna – read Before You Buy a Cellular Antenna, Do This First. It walks through a practical baseline test that takes an afternoon and will tell you whether an antenna is actually going to help your specific situation.
Nick Appleby has worked in cellular IoT and M2M connectivity for over 25 years, including founding and operating businesses in the UK cellular data sector. He writes about connectivity, antennas, network architecture, and IoT deployment practice on this site. Views expressed here are his own.