Stadium 5G uplink: frame change versus coverage fix
A 60,000-seat stadium is covered by a Nokia AirScale macro layer: three LTE carriers on 1800 and 2100 MHz (20 MHz each, 4×4 MIMO) and 12 rooflop 3.5 GHz massive-MIMO AAUs (64T64R, 100 MHz TDD). At the last match downlink peaked at 4.2 Gbit/s venue-wide, but uplink drove complaints: 22% of attendees uploaded video, cell-edge users saw 0.8 Mbit/s with a 65 ms 99th-percentile latency. The TDD pattern is 2.5 ms DDDSU with a strong downlink bias. Two lower-tier blocks have poor line-of-sight to the roof AAUs. The operator has no new spectrum, an LTE anchor at 1800 MHz, and can add small cells only inside those two blocks. The SLA for next season is 5 Mbit/s uplink for 90% of users and 40 ms latency.
1. Would moving to a more uplink-heavy TDD pattern fix the 0.8 Mbit/s cell-edge uplink? Reason quantitatively.
Read the worked answer
A frame change buys capacity, not coverage, and the complaint is a coverage complaint.
Changing from DDDSU (roughly 4:1 downlink-biased) to a pattern such as DSUUU can roughly double the uplink slots. That helps aggregate uplink capacity for the 22% of users uploading in good radio conditions. It does not help a user at the edge, whose throughput is limited by UE transmit power, path loss, body loss and interference - none of which change when you add uplink slots.
Quantify the gap: free-space loss difference between 3.5 GHz and 1.8 GHz is 20·log10(3.5/1.8) ≈ 5.8 dB, and the NLOS blocks add far more than that in blockage and penetration loss. The edge user needs roughly 10–20 dB more link budget, not more symbols.
The right use of a frame change is a dedicated uplink-heavy carrier for capacity in the bowl, combined with a coverage fix for the two blocks. Latency also matters: check SR periodicity and scheduling delay before choosing an uplink-heavy pattern, because more uplink slots can lengthen the round trip and work against the 40 ms target.
2. Compare supplementary uplink (SUL) on 1800 MHz, small cells in the two blocks, and more AAUs with altered downtilt. Which combination meets the SLA?
Read the worked answer
SUL on 1800 MHz. Reuses the LTE anchor band and gives roughly 5.8 dB of free-space advantage plus better penetration and body-loss behaviour - call it 8–10 dB effective. It is the cheapest way to lift outdoor edge users, but it depends on UE support (Rel-15 SUL-capable devices) and on scheduling that actually steers edge users to the supplementary carrier.
Small cells in the two blocks. The blocks have 20 dB or more of NLOS blockage, which SUL cannot recover. Four small cells per block on a licensed band, cabled or microwave-fed, is the only option that meets 5 Mbit/s for those users.
More AAUs / downtilt. Adds capacity and can improve interference shaping, but it does not add transmit power at the UE and therefore does not fix the edge.
Recommended combination: SUL on 1800 MHz for the bowl, one uplink-heavy 3.5 GHz carrier for capacity, and small cells in the two lower blocks. Expect edge throughput to move from 0.8 to roughly 3 Mbit/s with SUL plus extra uplink slots, with the blocks lifted above 5 Mbit/s only by the small cells. Verify with a drive/walk test against the 90%/5 Mbit/s target before committing to the second season.
3. Check the capacity plan: do 12 AAUs at 100 MHz carry the observed downlink peak, and what is the uplink headroom?
Read the worked answer
Downlink. A 64T64R 100 MHz cell at a planning figure of about 30 bit/s/Hz gives roughly 3 Gbit/s per cell; 12 cells gives about 36 Gbit/s venue peak. The observed 4.2 Gbit/s is around 12% of that, so downlink is not the constraint and the 4.2 Gbit/s figure is more a measure of demand than of capacity.
Uplink. With a 2:2-like pattern and about 5 bit/s/Hz, a 100 MHz cell gives roughly 500 Mbit/s, so about 6 Gbit/s venue-wide. If 22% of 60,000 attendees (13,200) upload and 10% are concurrently active at the 5 Mbit/s SLA rate, demand is 13,200 × 0.10 × 5 Mbit/s = 6.6 Gbit/s - above the uplink capacity. That is marginal even before interference.
The conclusion is that uplink, not downlink, sizes the venue. The uplink-heavy carrier plus SUL moves the capacity and coverage in the right direction; without them the 90% SLA cannot be claimed. Re-check the numbers with realistic concurrent-uploader statistics from the previous match rather than a flat 22%.