A smartphone passes two mobile network sites and enters a house at the end of the route. This interactive learning model shows which cell serves the connection, how signal level and quality change, and when several frequency bands can be used together.
Compare frequency bands along the route
Interactive · Understanding mobile networks
Understanding frequencies
One journey, three frequency bands: compare the connection outdoors and inside a house.
The table shows the strongest sector for each site and band, plus the cells in use and any handover candidate.
| Cell / channel bandwidth | RSRP dBm | Model SINR dB | Trees dB loss | House dB loss | Status |
|---|
RSRP: reference-signal strength. Model SINR: signal relative to interference and noise. Tree and building losses are listed separately.
Controls, model assumptions and sources
Controls: The slider and station buttons select a position along the route. Start journey plays the route forwards and can pause at each explanation station. The switches compare radio conditions at the selected position. Reset restores the starting position at 0 percent and the original switch settings. Reset view changes only the camera. To rotate with a finger, enable touch rotation first. When the scene has keyboard focus, the arrow keys rotate it and Home resets the camera.
A fictitious single-operator 5G SA network with three fixed sectors per site. Site A has 700 and 1800 MHz in each sector; site B also has 3600 MHz with an active antenna in each sector. Each sector–frequency pair is a separate cell: 15 cells in total. LTE/5G NSA and Dual Connectivity are outside this model.
Carrier Aggregation (CA): The switch enables suitable secondary carriers (SCells) alongside the primary carrier (PCell) during an assumed ongoing download. The PCell remains the primary cell. Adding a secondary carrier leaves the serving PCell unchanged. This example combines bands only within the same site and sector, assuming that the phone and network support those combinations. Real CA is not generally restricted to a single sector. A new secondary carrier must exceed its base thresholds by 2 dB for 200 ms; an active secondary carrier is retained down to its base thresholds. When the PCell changes within the same sector, the example configuration retains suitable carriers already in use; the former PCell may become an SCell. Following a site or sector change, secondary carriers are selected again. Configuration, activation and signalling timings are simplified.
Additional radio resources: Here, 700 + 1800 + 3600 MHz provides 10 + 20 + 80 = 110 MHz of channel bandwidth across three carriers. The usable data rate also depends on signal quality, load, TDD allocation, MIMO and device capabilities. In this model, CA leaves each carrier's received level, SINR and the PCell selection rule unchanged. The display keeps dBm and SINR values separate: the summary describes the PCell and the table lists each carrier. Parallel coloured lines illustrate aggregation rather than separate spatial propagation paths.
Sectors: Each site has three antenna faces spaced 120 degrees apart with fixed radiation patterns. The illustrative 65-degree horizontal mask also applies to the passive antennas. Active sectors additionally use reference and data beams. The 120 degrees describe the separation of the orientations; 65 degrees is the example full half-power beamwidth. Reception varies continuously across a sector boundary. As the person passes a site, the network may hand over to another sector cell on the same mast. Each drawn connection starts at the antenna face actually in use. In an overlap, the comparison margin initially retains the current cell despite small fluctuations.
Cell and beam changes: A data-beam change within the same cell leaves the PCell unchanged. A handover changes the PCell, including changes between sectors or frequencies at the same site. Adding or removing an SCell is a separate action. The phone measures and the network controls these changes. The drawing shows carriers assumed to be in use during a download; it does not show every configured but inactive SCell or its activation signalling.
Example selection rules: When reception is adequate, the order of preference is 3600, then 1800, then 700 MHz. Each carrier has its own minimum thresholds, listed in the table status fields. A new candidate needs an additional 2 dB reserve; at equal priority, it needs at least a 3 dB RSRP advantage while the current cell still meets its base thresholds. The advantage must persist for 320 ms. This is the evaluation time before the decision, rather than a connection interruption. Signalling and measurement filtering are outside the simulation. These are illustrative learning rules rather than a universal 3GPP algorithm. Playback uses a separate educational timescale. Moving the slider backwards shows the same forward progression at the chosen position.
Reading the values: RSRP describes received reference-signal strength: −80 dBm is stronger than −100 dBm. Model SINR describes that signal relative to interference and noise. At 10 dB, signal power is ten times the combined interference and noise power; at 0 dB, the two are equal. A weaker signal can still have a favourable ratio. Both displays contain calculated example values. Very low calculated values may lie below a phone's measurement range.
Calculation: RSRP is based on an SS-RSRP-like quantity. Model SINR = reference-signal power / (interference power + noise power), converted to dB. All powers refer to the same individual reference resource rather than the whole channel bandwidth. Powers are summed in linear units. Noise assumes −174 dBm/Hz, a 7 dB receiver noise figure and a subcarrier spacing of 15 or 30 kHz. The 10 / 20 / 80 MHz channel bandwidths describe the carriers and are kept separate from reference-signal power.
Interference: All other sector cells on the same frequency contribute using an illustrative effective coupling factor of 0.12 each, including neighbouring sectors on the same mast. An assumed outdoor background from further transmitters adds −113 dBm per reference resource. Indoors, the building envelope also attenuates that external background while receiver noise stays constant. A wall therefore attenuates both the useful signal and external interference, so SINR can decline less than RSRP. The coupling factor represents neither measured load nor precise SSB overlap. Actual time and frequency resource use is outside the model, hence the label Model SINR rather than standards-compliant SS-SINR. The assumed activity of other cells stays fixed when comparing CA. Different bands do not interfere with each other in this model. These values describe the reference resource rather than the drawn data beam or a predicted data rate.
Beamforming: The 3.6 GHz antenna uses seven illustrative wider reference directions and 17 narrower data directions. The radio direction for the data beam is assumed to be ideally known, rather than inferred solely from the displayed SS-RSRP. The animation uses illustrative timing and a simplified beam codebook. The transparent shape represents the main lobe schematically rather than a coverage boundary.
Trees along the route: A separate group of trees attenuates paths crossing its assumed volume. Only the length of the straight radio path inside the three-dimensional obstacle contributes; a path above the volume receives no tree loss. A poor path towards B can therefore make the more distant site A more usable. The maximum additional losses of 17 / 27 / 36 dB are illustrative assumptions rather than general vegetation-loss values. Diffraction, scattering, seasonal changes and individual branches are outside the model. The unknown background from other directions receives no blanket tree loss. The Trees in the radio path switch compares the same location. Technical context: ITU-R P.833, attenuation in vegetation.
Inside the house: The house is across the route. It adds no extra outdoor shadowing in this scene. After entry, the example loss rises to at most 17 / 27 / 36 dB at 700 / 1800 / 3600 MHz. The increase is smoothed for explanation: after 1.3 metres, about 63 percent of the respective maximum loss is reached. These are assumed total losses rather than universal material properties or a full 3GPP building simulation. The same frequency-dependent envelope affects external useful signals and interference; incidence angles and individual wall crossings are omitted. The chosen signal levels and network rules determine the frequency change indoors. With building attenuation switched off, the person stays at the same indoor position while the additional loss is removed.
Cutaway view: Roof and external walls open or become transparent to reveal the interior. Building attenuation remains active. A dashed indoor radio path indicates the attenuated connection rather than a calculated spatial field-strength distribution.
Scope: This learning model does not predict reception at a real location. Multipath, actual channel occupancy, radio-link failure and uplink behaviour are outside its scope. Buildings, antennas and people use schematic scales for readability. The assumed 10 / 20 / 80 MHz channel bandwidths are separate from carrier frequency. Context for building attenuation: TR 38.901, section 7.4.3.
Technical references: TS 38.300, mobility and Carrier Aggregation · TS 38.215, measurement quantities · TS 38.214, measurements and reporting. This simulation is an educational illustration rather than a protocol-conformance test.
Using the model
Start journey follows the route. With pausing enabled, the animation stops at explanation stations. Next station jumps to the next handover or building entry. The position slider selects a location; the switches compare tree attenuation, building attenuation, the additional 3.6 GHz band and Carrier Aggregation at that location.
Reset restores the position to 0 percent and all original settings. Reset view changes only the camera. On a phone, swiping over the scene scrolls the page until touch rotation is enabled. With keyboard focus on the scene, the arrow keys rotate it and Home resets the camera.
Reading frequency, reception and channel bandwidth
The first display identifies the primary carrier and its cell. RSRP describes received reference-signal strength. Model SINR compares that signal with interference and noise. Channel bandwidth describes the spectrum provided by the carriers in use. This example assumes 10, 20 and 80 MHz of channel bandwidth at 700, 1800 and 3600 MHz respectively.
With Carrier Aggregation enabled, suitable carriers at the same site and sector can be used simultaneously. Three carriers provide 110 MHz of total channel bandwidth in this model. Each carrier keeps its own received level and SINR. Actual throughput also depends on signal quality, load, TDD allocation and device capabilities.
Handovers along the route and reception indoors
Each combination of site, sector and frequency is a separate cell. As the person passes a site, the connection can move to another frequency, antenna face or site. The trees affect only radio paths crossing their assumed volume. Building attenuation begins after entry into the house. The cutaway view opens the building for visibility while its attenuation remains active.
Model assumptions and scope
The landscape and 5G SA network are fictitious. Selection rules, channel bandwidths and additional losses are assumptions for this learning model. Distances and losses support comparisons within the scene. Playback uses an educational timescale. Reception at a real address and specific download rates are outside its scope.
The expandable model notes provide the full selection rules, loss assumptions, timing values and technical references. The measurement table lists cells individually, including tree and building losses.
Explore the technology
The article on 5G frequencies and bands explains band numbers, channel bandwidth and spectrum allocations. The guide to 5G range at different frequencies covers propagation. The interactive antenna view explores antenna direction and tilt, while beamforming and Massive MIMO explains spatial signal shaping.