LTE Speed Test Troubleshooting: A Field Engineer's Checklist

Isolating Layer 1 RF Metrics from Network Congestion

When conducting field verification or response to customer throughput complaints, low throughput during speed tests is a frequent observation. The first requirement in LTE speed test troubleshooting is determining whether the performance bottleneck exists in the physical layer (Layer 1), radio access network (RAN) configuration, backhaul bandwidth, or core network congestion.

3GPP TS 36.214 defines the physical layer measurements for E-UTRA networks. Reference Signal Received Power (RSRP) measures the linear average over the power contributions of the resource elements that carry cell-specific reference signals within the considered measurement frequency bandwidth. While RSRP indicates signal strength, it does not guarantee high throughput on its own. A drive test scenario showing an RSRP of -80 dBm can still produce low throughput if co-channel interference or high cell utilization degrades the Signal-to-Interference-plus-Noise Ratio (SINR).

To isolate Layer 1 issues from backhaul or server-side bottlenecks, compare the calculated physical layer throughput against the theoretical maximum for the assigned Modulation and Coding Scheme (MCS) and Physical Resource Block (PRB) count. If the eNodeB allocates 100 percent of available PRBs and utilizes 64QAM or 256QAM modulation, but the end-to-end speed test remains constrained, the bottleneck typically resides above the radio interface, such as limited transmission backhaul or server throttling.

Evaluating RSRP, RSRQ, and SINR Thresholds

RF engineers rely on specific RF performance windows to evaluate whether the radio environment can support high-order modulation schemes. Downlink throughput depends on maintaining low Block Error Rates (BLER) while using maximum resource allocation.

• RSRP (Reference Signal Received Power): Values stronger than -85 dBm indicate excellent signal coverage, typically found near cell sites. Coverage between -85 dBm and -95 dBm represents good mid-cell coverage. Values from -95 dBm to -105 dBm reflect edge coverage, while values lower than -110 dBm lead to degraded channel quality and potential handovers or dropped connections.
• RSRQ (Reference Signal Received Quality): Defined by 3GPP TS 36.214 as the ratio N x RSRP / E-UTRA carrier RSSI, where N is the number of PRBs. RSRQ accounts for both signal power and structural noise/interference. Values above -10 dB indicate clean RF conditions. Values below -15 dB signal significant co-channel interference or high traffic loading on neighboring sectors.
• SINR (Signal-to-Interference-plus-Noise Ratio): SINR directly influences the Channel Quality Indicator (CQI) reported by the user equipment (UE) back to the eNodeB. To achieve 64QAM modulation, an SINR above 15 dB is generally required. Achieving 256QAM downlink modulation, as defined in 3GPP TS 36.101 requirements, requires stable SINR values exceeding 20 dB to 22 dB. When SINR drops below 0 dB, control channel decoding errors increase and Hybrid Automatic Repeat Request (HARQ) retransmissions rise, drastically reducing net throughput.

Diagnostic Logic for Downlink and Uplink Throughput Drops

When executing an LTE speed test troubleshooting routine, separate the analysis into downlink and uplink paths, as their limiting factors differ significantly in field conditions.

Downlink Throughput Troubleshooting

1. Check CQI and MCS Distributions: If the UE reports low CQI values despite strong RSRP, verify physical sector orientation. RF interference from over-shooting sectors or incorrect mechanical tilt often degrades SINR. Cross-check sector alignment against baseline engineering documents using tools like GSM Azimuth Checker.
2. Verify Rank Indicator (RI) and MIMO Mode: Downlink throughput drops substantially if spatial multiplexing degrades. If the Rank Indicator reported by the UE drops from Rank 2 (2x2 MIMO) or Rank 4 (4x4 MIMO) down to Rank 1, the system falls back to transmit diversity or single-antenna transmission. Rank drops are caused by poor channel decorrelation or low SINR.
3. Inspect Sector PRB Utilization: High PRB allocation per user is required for peak speeds. On a 20 MHz channel (100 PRBs), a single user under zero-load conditions should receive close to 100 PRBs. If the eNodeB limits allocation to 20 or 30 PRBs during a speed test, check for active multi-user scheduling or eNodeB rate-shaping parameters.

Uplink Throughput Troubleshooting

Uplink speed test drops are frequently caused by power limitations at the mobile terminal or elevated uplink interference at the cell site:

• UE Transmit Power Limits: According to 3GPP TS 36.101, Class 3 LTE UEs have a maximum transmit power of 23 dBm (plus or minus 2 dB tolerance). When a UE reaches maximum output power (Power Headroom = 0 dB), it can no longer maintain high-order uplink modulation (such as 16QAM or 64QAM) or wide PRB allocations, causing uplink throughput to drop rapidly at the cell edge.
• Uplink Interference on eNodeB: External interference sources, Passive Intermodulation (PIM), or high user density on adjacent cells elevate the Received Total Wideband Power (RTWP) at the eNodeB receiver. High uplink noise forces the eNodeB scheduler to assign lower uplink MCS indices.

Verifying Carrier Aggregation and System Configurations

Carrier Aggregation (CA) combines multiple component carriers (CCs) to increase bandwidth across same or different frequency bands, as governed by 3GPP TS 36.331 specifications. When an LTE speed test fails to achieve expected multi-carrier data rates, evaluate carrier management mechanisms.

Component carrier activation is dynamic. The Primary Cell (PCell) manages RRC connection and control signaling, while Secondary Cells (SCells) are added or deactivated based on data volume and RF quality thresholds. If SCell RSRP or SINR drops below configured thresholds, the eNodeB deactivates the secondary carrier, dropping overall aggregate throughput.

Different frequency bands exhibit different path loss characteristics. High-band carriers (such as 2.6 GHz Band 7 or 3.5 GHz Band 42) attenuate faster than low-band carriers (such as 800 MHz Band 20 or 700 MHz Band 12). When performing drive tests across coverage boundaries, model link budgets and expected path loss using GSM Coverage Planner Web to verify if SCell drop-outs align with expected propagation limits.

Field logging tools, such as GSM Drive Test, allow field personnel to record real-time cell parameters, active component carriers, allocated PRBs, and CQI metrics during active speed test sequences, helping pinpoint the exact moment an SCell is deconfigured.

Field Engineer Troubleshooting Protocol

When arriving at a site or drive test route exhibiting low LTE speed test results, follow this structured troubleshooting sequence:

Frequently asked questions