3GPP TR 38.821 Reference Scenarios, Simulated
Set 1 · LEO-600 · S-band · Transparent payload
TR 38.821 is the 3GPP study that defines how NR supports non-terrestrial networks. This page explains its six reference scenarios and walks through a NetSim system-level simulation of the Set-1 LEO-600 scenario: SINR and throughput distributions, satellite capacity, area traffic capacity, and spectral efficiency.
The six reference scenarios in TR 38.821
Clause 4.2 of TR 38.821 defines six reference scenarios, combining orbit, payload type, and beam behaviour (Table 4.2-1 of the report).
| Transparent satellite | Regenerative satellite | |
|---|---|---|
| GEO based non-terrestrial access network | Scenario A | Scenario B |
| LEO based non-terrestrial access network: steerable beams | Scenario C1 | Scenario D1 |
| LEO based non-terrestrial access network: the beams move with the satellite | Scenario C2 | Scenario D2 |
A transparent payload repeats the waveform unchanged (RF filtering, frequency conversion and amplification). A regenerative payload also demodulates, decodes and switches, placing all or part of the gNB on board the satellite.
| For calibration, the report also fixes | Where |
|---|---|
| Set-1 and Set-2 satellite parameters: EIRP density, antenna gain, beamwidth and beam diameter per orbit (GEO, LEO-1200, LEO-600) and band (S, Ka) | Tables 6.1.1.1-1 and 6.1.1.1-2 |
| UE characteristics: handheld in S-band (omnidirectional element, 23 dBm, 7 dB noise figure) and VSAT in Ka-band | Table 6.1.1.1-3 |
The scenario simulated in NetSim
A system-level simulation of the Set-1 LEO-600 S-band scenario with a transparent payload.
The satellite acts as a relay, passing signals between the ground gateway and the User Equipment (UE). 19 spot beams are arranged in a hexagonal layout, with 10 handheld UEs per beam, 190 UEs in total, all receiving full-buffer downlink traffic.
Objective
Measure the distributions and percentiles of SINR and throughput across the spot beams, and derive satellite capacity, area traffic capacity, and average spectral efficiency from the same run.
▶
2:42
3GPP 38.821 Set-1 Reference Scenario
Network scenario
Figure 1: 19 beams arranged in a hexagonal layout; S-band, LEO orbit; total 190 UEs (10 UEs per beam) connected via a transparent satellite relay with full buffer DL traffic.
Parameter configuration
| Evaluation parameters | |
|---|---|
| Satellite Orbit | LEO 600 |
| Satellite Altitude | 600 km |
| EIRP (dBW/MHz) | 34 |
| Noise Figure (dB) | 7 |
| Antenna Aperture (m) | 1 |
| Band | S |
| Frequency | 2 GHz (S Band) |
| Bandwidth (MHz) | 30 per beam |
| Scheduling Type | Round robin |
| Traffic | Full buffer DL |
| RU% | 100% |
| Elevation Angle | Beam centres are at elevation angle 90°. The UE’s elevation angle would depend on its location within the beam. |
| Antenna Pattern | Bessel function per section 6.4.1 of TR 38.811. All UEs are not at the Nadir point and hence antenna gains need to be computed. |
| Additional Loss (dB) | 0 |
| Clutter Loss (dB) | 0 |
| UE Density | 10 UEs per spot beam |
| UE Mobility | No Mobility |
| Antenna Temperature (K) | 290 |
| UE TX Power (dBm) | 23 |
| UE RX Antenna Gain (dB) | 0 |
| Traffic Model | Full buffer |
Table 1: System simulation parameters.
Results and discussion
The results are shown as CDF plots for UE throughput, SINR, and per-beam throughput, with their 5th, 50th and 95th percentiles.
| Throughput percentile metrics | |
|---|---|
| 5th percentile | 1.62 Mbps |
| 50th percentile | 6.49 Mbps |
| 95th percentile | 10.85 Mbps |
Figure 2: CDF of downlink throughput per UE; LEO satellite network using S-Band with 19 beams and 10 UEs per beam.
| SINR percentile metrics | |
|---|---|
| 5th percentile | -0.40 dB |
| 50th percentile | 9.50 dB |
| 95th percentile | 14.66 dB |
Figure 3: CDF of SINR; LEO; S-Band; 10 UEs per beam; 19 beams.
| Per-beam throughput percentile metrics | |
|---|---|
| 5th percentile | 45.39 Mbps |
| 50th percentile | 57.27 Mbps |
| 95th percentile | 69.97 Mbps |
Figure 4: CDF of per-beam throughput.
SINR follows beam geometry
The SINR CDF shows how signal quality changes with the UE’s location within a beam. UEs near the beam centre usually have better SINR, while edge UEs see lower values due to reduced antenna gain.
Throughput tracks SINR
Scheduling is round robin and all UEs carry full-buffer traffic, so the throughput CDF mirrors the SINR distribution: UEs with higher SINR achieve slightly better throughput.
Beams differ by UE placement
Each beam has the same bandwidth and number of UEs, but per-beam sum throughput depends on where the UEs fall within the beam, so randomness in UE positions produces a distribution across beams.
Capacity and efficiency metrics
Three figures of merit derived from the same run.
Satellite capacity
Sum throughput of all 190 UEs:
1127.54 Mbps
Area traffic capacity
- Number of beams = 19
- Beam radius, R = 55.13 km
- Area per beam:
\[ A = \frac{3\sqrt{3}}{2} \cdot R^2 = 7887.1 \text{ km}^2 \]
- Total coverage area:
\[ 7887.1 \times 19 = 149854.9 \text{ km}^2 \]
- Area traffic capacity:
\[ \frac{1127.54 \text{ Mbps}}{149854.9 \text{ km}^2} = 7.526 \text{ kbps/km}^2 \]
Average spectral efficiency
- Channel bandwidth = 30 MHz
- Number of TRxPs = 19
-
\[ \frac{1127.54 \times 10^6}{30 \times 10^6 \times 19} \] \[ = 1.98 \text{ bits/s/Hz/TRxP} \]
Take it further
Reproduce this scenario, change the constellation, or move from calibration to your own study.