• Telecom & ICT Equipment Supplier
  • Multi-Vendor Sourcing
  • Worldwide Shipping
YT CIG Telecom & ICT Equipment
Get a Quote

Repeated RRU alarms: root causes and a triage sequence

Updated on 2026-11-19

Repeated RRU alarms in LTE base stations usually trace to four physical layers: the antenna system (VSWR), the CPRI optical link, the -48 V DC feed, and thermal management inside the radio. Most repeats are one marginal connector, module or cable crossing several alarm thresholds; unrelated simultaneous failures are rare.

What causes repeated RRU alarms in LTE base stations?

An RRU alarm is a threshold crossing, not a diagnosis. The radio measures forward power, reflected power, received optical power, frame sync, supply voltage and case temperature; when one crosses a configured limit, the management system raises an alarm with a severity drawn from ITU-T X.733 (critical, major, minor, warning, indeterminate, cleared) and a probable cause mapped in 3GPP TS 32.111-2. A repeat means the measured quantity is oscillating around the threshold, or the first repair changed the symptom without removing the cause. Outdoor oscillation drives many repeat tickets. A connector that is dry at noon and wet after rain produces a VSWR alarm that appears for hours and clears. A fan with worn bearings produces a temperature alarm every afternoon. The second case is misdiagnosis: one water-logged jumper can raise VSWR, push the radio into TX power reduction, and produce a TX power alarm that sends a technician to the wrong board.

Vendors name alarms differently; the measured quantity and reporting object carry across systems, so the first step groups alarms by scope: whole site, one sector, one RRU, one antenna path.

Which four alarm families account for most repeat tickets?

Four families cover the alarms that repeat in field logs; the names in the third column are representative.

Alarm family Measured quantity Alarm names seen in logs Usual physical cause First field check
VSWR / antenna Forward and reflected power per port, converted to VSWR or return loss Antenna port VSWR high, TX path mismatch, reflected power alarm Water ingress, loose 7/16 DIN or N connector, crushed feeder, damaged antenna element Return-loss sweep with distance-to-fault; connector torque and sealing
Optical / CPRI Received optical power, loss-of-signal and loss-of-frame state, coding errors LOS, LOF, RAI, optical RX power low, CPRI link down, SFP TX fault Contaminated LC end-face, bend radius violation, failed SFP, broken fiber, wrong rate or wavelength End-face inspection to IEC 61300-3-35, DDM receive power, patch-cord swap
DC power Input voltage and current at the RRU feed Input power fail, DC undervoltage, breaker open, battery discharging, rectifier fail Rectifier failure, aged battery, corroded lug, undersized cable, tripped breaker Voltage at the RRU input under load
Thermal Case temperature, fan speed, TX power derating state Temperature high, fan failure, TX power derated, thermal shutdown Fan bearing wear, dust-blocked heatsink, clogged filter, solar load, blocked cabinet airflow Case temperature against ambient, fan RPM, heatsink condition

A minor VSWR alarm that repeats forty times a week is more urgent than one critical RRU failure replaced under warranty: the mismatch is still present. The RRU family guide covers the hardware, including branch counts that change CPRI load.

Reading a VSWR alarm before the climb

The VSWR value in the alarm is a ratio derived from forward and reflected power. Return loss and reflected power follow from the reflection coefficient Γ = (VSWR – 1) / (VSWR + 1), with RL = -20 log10|Γ| and reflected power = 100 Γ²; the table is computed from those identities.

VSWR at the radio port Return loss Reflected power Reading
1.2:1 20.8 dB 0.8% Healthy jumper and antenna
1.5:1 14.0 dB 4.0% Acceptance point
2.0:1 9.5 dB 11.1% Alarm region; sweep before climbing
2.5:1 7.4 dB 18.4% Moisture or a damaged joint
3.0:1 6.0 dB 25.0% Often an open or short

A 2:1 reading means 11% of forward power returns, or 9.5 dB return loss; compare that with the site specification.

The measurement point matters. Return loss at the radio port includes the jumper and, where the RRU sits in a ground cabinet with a long feeder, the feeder. A one-way feeder loss of 3 dB adds 6 dB to the measured return loss, because the reflected wave crosses the loss twice. An antenna with 9.5 dB return loss then reads 15.5 dB at the radio. On an RRU site the jumper is short, so the reading tracks the antenna.

Intermittent VSWR has a weather signature. An alarm during rain points to water ingress at the connector or a cracked jumper jacket. An alarm on hot afternoons points to thermal expansion in a connector that was not torqued to specification.

PIM is a nonlinear product measured to IEC 62037. A PIM fault can pass a return-loss sweep because reflected power at the carrier stays normal; when uplink noise rises but VSWR is clean, test PIM at the ports.

Distance-to-fault turns a sweep into a location. Resolution is Δd = vf·c / (2·BW). With a velocity factor of 0.88 and a 60 MHz sweep, resolution is about 2.2 m, enough to separate the radio connector from a jumper joint.

CPRI fault versus radio fault

CPRI carries digitized IQ samples between the BBU and the RRU over a fixed-rate serial link, and CPRI V7.0 fixes the line rates.

CPRI option Line bit rate Typical deployment
1 614.4 Mbit/s Legacy
2 1228.8 Mbit/s 20 MHz LTE, 2T2R
3 2457.6 Mbit/s Multi-carrier or 4T4R
4 3072.0 Mbit/s Vendor-specific
5 4915.2 Mbit/s Multi-carrier
6 6144.0 Mbit/s Multi-carrier
7 9830.4 Mbit/s 8T8R or high capacity
8 10137.6 Mbit/s 10G SFP+

Options 1 to 6 use 8b/10b coding; options 7 and 8 use 64b/66b. A 10G SFP+ carries option 7; a 1.25G SFP cannot. The 10G SFP+ guide and the 4T4R versus 8T8R comparison cover modules and branch counts.

LOS means no light reaches the receiver. LOF means light arrives but frame alignment fails, pointing to a rate mismatch, bit errors, or a closed eye. RAI is sent toward the far end when the remote detects a defect. An RRU reporting LOS while the BBU reports RAI places the break between them. Both ends reporting LOS points to a fiber cut or a powered-down far end.

Read SFP DDM receive power at both ends against the datasheet sensitivity and overload point. A receiver a few dB above sensitivity has margin; one at sensitivity produces LOF before a clean LOS. ITU-T G.652.D fiber is specified at 0.35 dB/km or better at 1310 nm and 0.20 dB/km or better at 1550 nm. A 10 km span consumes at least 3.5 dB at 1310 nm; 40 km consumes at least 8 dB at 1550 nm.

IEC 61300-3-35 defines connector end-face acceptance zones. A failed connector is cleaned and re-inspected; a scratched ferrule is replaced. A zip-tied patch cord adds bend loss.

A CPRI outage removes the baseband control channel; the radio may report TX or VSWR conditions that clear when the link returns. Re-read the RF alarms against a live link.

Why do power and temperature alarms arrive together?

Power and temperature share one output variable: TX power. An RRU reduces output when the rail sags or the case gets hot, and the network notices that reduction first.

ETSI EN 300 132-2 puts the normal service range of a -48 V DC interface at -40.5 V to -57 V. A 16 mm² copper pair, 50 m long, carrying 20 A, has a loop resistance of 2 × 0.0172 × 50 / 16 = 0.108 Ω (copper resistivity to IEC 60028), so it drops 2.15 V. At 100 m and 25 A the same cable drops 5.38 V, leaving a -48 V plant at -42.6 V at the RRU, close enough to the lower rail that a battery discharge takes it out of tolerance. Measure at the RRU input under load, with a meter that holds a minimum value. A normal reading at the rectifier terminals is compatible with a low voltage at the RRU; the minimum at the RRU input decides. The minimum-hold function matters because the alarm often appears during a transient load, while a steady-state reading looks normal.

Battery strings fail under discharge load; float voltage stays normal even when the string has lost capacity. IEEE 1188 covers VRLA maintenance and testing; a string that holds float voltage can collapse in the first seconds of a mains failure, when every RRU draws current. A site whose alarms cluster around grid outages is pointing at the battery or rectifier.

Fan bearings wear, heatsink fins clog, and a cabinet ventilated for 30 °C ambient gets a 45 °C afternoon. The RRU derates TX power at a vendor-set case temperature, so the first symptom can be a coverage complaint or a TX power alarm. The temperature alarm follows once derating can no longer hold the case temperature at the vendor limit.

Correlate alarm time with weather and load. Alarms peaking between 14:00 and 17:00 and clearing after sunset are thermal. Alarms during storms point to lightning damage on the AC side and a failed SPD; IEC 62305 covers the protection design, IEC 61643-11 the SPD.

ITU-T K.27 covers bonding inside a telecom building. A loose bond at the RRU or cabinet raises fault-path impedance and produces an intermittent alarm.

A triage order that discriminates fast

Order the checks by discrimination per minute, with site infrastructure first. A DC sag or CPRI outage can create RF alarms that look like antenna faults, so an RF sweep first produces hard-to-interpret evidence.

  1. Export 24 hours of alarms with timestamps and severities. A site-wide burst is a shared cause; one path alarming forty times is local.
  2. Measure DC at the RRU input under load against the ETSI window. Ten minutes removes the layer that imitates the rest.
  3. Read CPRI at both ends: LOS, LOF, RAI, DDM receive power. No light or no frame alignment makes downstream RF readings unreliable.
  4. Read case temperature, fan RPM and the derating flag. If the radio is derating, treat thermal management as the first fault.
  5. Sweep RF only after steps 2 to 4 are clean. One bad path points to a jumper, connector or element; every path on one sector points to a shared feeder or antenna.
  6. Change one variable per visit and record before and after values. A reset clears the latched state and destroys the timestamp evidence, so record both values before resetting.
  7. Observe one full trigger cycle before closing: a rain event, a heat peak or a grid outage.
What the log shows Likely layer Check before dispatch What confirms it
Every RRU alarms in one polling interval DC plant, BBU or transport Rectifier voltage, breaker state, CPRI at the BBU All RRUs recover when the rail or link returns
One sector, all paths Shared feeder, antenna or RRU VSWR per path, receive power per SFP Sweep shows one shared feeder or connector fault
One RRU, one path Jumper, connector or antenna element Return loss and distance-to-fault DTF null at the jumper or connector
Alarm every afternoon, clears at night Thermal Case temperature, fan RPM, solar exposure Alarm stops after fan, filter or airflow repair
Alarm only during rain Moisture Connector sealing, VSWR trend Dry-out clears; resealing prevents return
Alarm only during grid outage Battery or rectifier Battery discharge test, rectifier load sharing Voltage stays inside the ETSI window on battery
Alarm after a lightning event SPD or bonding SPD status indicators, bond resistance SPD and bond repair stop the repeats

Remote steps are cheap; the climb comes last. Carry a jumper, an SFP, a cleaning kit and a power meter. The refurbished equipment inspection checklist covers bench checks before an RRU enters site stock; the Huawei RRU5901w, RRU5904 and Ericsson RRU6630 listings carry the band and power fields.

Which 3GPP and site standards set the pass/fail line?

Conformance standards define the RF performance a healthy radio must meet; site standards define the installation quality that keeps it there. Neither sets the alarm threshold, a vendor or operator parameter.

Document What it fixes Value or rule used in triage
3GPP TS 36.104 E-UTRA base station RF Frequency error within ±0.05 ppm; ACLR 45 dB
3GPP TS 36.141 Base station conformance testing RF and environmental test methods
3GPP TS 37.105 Active antenna system base station Over-the-air requirements
3GPP TS 32.111-2 Fault management alarm IRP Alarm structure and probable cause
ITU-T X.733 Alarm reporting critical, major, minor, warning, indeterminate, cleared
ETSI EN 300 132-2 DC power interface -40.5 V to -57 V normal service
ITU-T G.652.D Single-mode fiber 0.35 dB/km at 1310 nm; 0.20 dB/km at 1550 nm
IEC 61300-3-35 Connector end-face Visual acceptance zones
IEC 61754-20 LC connector interface LC dimensions
ITU-T K.27 Bonding and earthing Bonding inside a building
IEC 62305 Lightning protection Protection design
IEEE 1188 VRLA battery Maintenance and testing

TS 36.104 is the reference when a frequency or output-power alarm is suspected. A conforming radio holds ±0.05 ppm and 45 dB ACLR, so a frequency alarm points to the reference clock or sync chain.

A 2.0:1 threshold is an operator setting, chosen to catch a broken jumper while ignoring normal impedance variation across the band. The same antenna can pass on another network.

Installation standards carry what crews control: connector torque, weatherproofing, bend radius and bonding. IEC 61754-20 fixes LC dimensions; the torque for a 7/16 DIN connector comes from the connector datasheet and the operator’s installation specification. A loose connector returns every time the wind picks up.

When is a repeat alarm a design problem?

A repeat alarm can be a design fault when the hardware works inside its ratings but outside the installation’s margins.

The DC cable is the clearest case. The 100 m, 25 A example drops 5.38 V, leaving a -48 V plant at -42.6 V. The radio may pass at commissioning, then alarm on the first battery discharge. Replacing the RRU leaves the voltage drop unchanged, so the cable is the fault.

An antenna specified at 1.5:1 across the band can reach 2.0:1 at the band edges, and a threshold set at 1.8:1 will alarm on a healthy antenna at the highest or lowest channel. The pattern is a VSWR alarm on one edge channel with clean values across the rest of the band. A module rated for 10 km on a 15 km span may come up cold and fail later, because the link never had margin. The DDM reading at commissioning is the baseline.

A cabinet with blocked airflow, or a heatsink in direct sun, can keep an RRU derating for weeks each summer. The repair is shade, airflow or a larger cabinet.

These cases share a signature: the alarm returns at the same load, weather or channel condition, and the replaced part tests clean. A bench test exercises the radio under lab conditions; the site conditions that produced the alarm remain untested.

How do you prove the repeat is gone?

A cleared alarm records that the measured value crossed back over the threshold; the cause can still be present. Proof of repair needs an observation window long enough to contain the original trigger at least once.

Pick the window from the repeat pattern. A rain-triggered alarm needs a significant rain event; a thermal alarm needs several hot days; a battery alarm needs a discharge test. Seven days is a workable minimum, but the trigger matters more. Record five numbers before and after: alarm count per object, return loss per path, DDM receive power per SFP, minimum DC voltage at the RRU input, and maximum case temperature. A repair that clears VSWR but leaves one path 3 dB low has moved the problem.

Close the ticket when the trigger recurs and the alarm does not. If the same object alarms again, the cause was not removed. If a different object on the same sector alarms, the work may have introduced a fault: a disturbed jumper, a dirty connector, a loosened bond.

Keep the traces with the site record so the next crew can compare a new event with the commissioned values.

Frequently asked questions

What VSWR level justifies a tower climb?

No single level does. A 1.5:1 reading is 14.0 dB return loss and 4.0% reflected power; 2.0:1 is 9.5 dB and 11.1%. A single path in the alarm band while the others sit below 1.5:1 points to a jumper or connector, and a distance-to-fault sweep can locate it. Climb when the sweep places the fault above the radio jumper, where the antenna system begins.

Can a dirty fiber connector cause a VSWR alarm?

No direct path exists. VSWR is measured on the RF port; a dirty LC end-face affects the CPRI link. The indirect path is real: CPRI LOS or LOF removes the baseband control channel, and the radio may report TX power or VSWR conditions. Check the optical link first when a VSWR alarm and LOS arrive in the same polling interval.

Why does an RRU alarm clear after a reset and return hours later?

A reset clears latched alarm state and restarts the measurement cycle; it does not change the connector, fan or battery that produced the crossing. An alarm that returns at the same ambient temperature is thermal. One that returns after rain is moisture. One that returns on battery is voltage.

Is a repeated CPRI LOS alarm always a broken fiber?

No. LOS means no light reaches the receiver. A failed transmitter SFP, a pulled patch cord, a tight bend, a dirty end-face and a powered-down far end all produce it. Swap the patch cord and SFP before ordering a fiber repair; an OTDR trace confirms a break.

How much reflected power does a 2:1 VSWR represent?

11.1% of forward power, equivalent to 9.5 dB return loss. The reflection coefficient is 1/3 at 2:1, and (1/3)² is 0.111. At 3:1 the reflected share rises to 25% and return loss falls to 6.0 dB. 11% of 40 W is 4.4 W returned.

Which alarm should be cleared first, temperature or VSWR?

Temperature, when the radio is derating. Derating changes the RF operating point, and VSWR measured at reduced power has less dynamic range. Restore cooling or the DC feed, confirm full output, then re-read VSWR. A VSWR alarm that survives full power is a real antenna fault.

Need a specific part?

Send the part number and quantity — we confirm condition, availability and price before you commit.