Preventive UPS Maintenance: The Annual Checklist That 80% of DCs in Mexico Skip

The UPS in a Mexican data center usually goes one of two ways: either the battery gets swapped after it has already died and triggered an alarm, or an annual manufacturer visit delivers a 14-page report that nobody reads. Between those two extremes there is a middle ground where reliability is actually won or lost. The middle ground is not a contract term: it is a checklist.

What Real Preventive Maintenance Includes (Not the ‘Glance’)

A preventive maintenance visit that takes two hours and produces a “system OK” line is not preventive maintenance: it is a signature for a check. Real preventive maintenance of an industrial UPS (10–500 kVA) requires at least one full working day per unit and covers eight blocks. Skipping any of them means leaving a known failure vector for the next power event.

Visual and Thermographic Inspection

The thermographic inspection identifies hot spots that will become failures within 6–18 months. It is done under load (not at idle), with a calibrated thermal camera (resolution ≤ 0.1°C, emissivity adjusted for the surface), and covers:

• Power connections (input/output terminals, breakers, busbars).
• Capacitors and DC bus filters.
• Magnetics (transformer and inductor windings).
• Cable lugs and crimping points.

A 5°C rise above the baseline on a lug means the connection is degrading. A 10°C rise means replace it now and trace the root cause before the next event. The IR report belongs in the maintenance file with the photograph and the timestamp.

Electrical Measurements Under Load

The UPS is not tested at idle: it is tested under load, with the building running. Three-phase voltage balance, current per phase, harmonic distortion (THDi on input, THDu on output), neutral-to-ground voltage, and frequency stability. Values outside IEEE 1159 or IEEE 519 limits indicate either upstream power problems or UPS internal degradation. Either case requires follow-up, not a checkbox.

Real Autonomy Test (Not the Theoretical One)

The autonomy label on the UPS data plate is at 25°C, with new batteries, at 0.8 power factor, and at rated load. In the real world, batteries are 2–4 years old, the load is rarely at the rated value, and the ambient is rarely 25°C. The autonomy test consists of simulating a real outage (load bank or building load transfer to UPS), measuring actual runtime at the real load, and comparing it with the calculated theoretical value. If the actual is less than 80% of the theoretical, the battery bank is end-of-life regardless of what the monitoring system says about internal impedance.

Cleaning and Re-Torquing of Connections

Torque on power connections is not a one-time spec: it relaxes with thermal cycling. A re-torque pass on all power terminals (with the equipment de-energized, not in bypass), and a cleaning of dust on filters, fans, and boards, is the difference between a UPS that lasts 12 years and one that fails in year 7. Torque values must come from the manufacturer manual, not from “a feel”.

Static and Maintenance Bypass Verification

The bypass is the path that keeps the load alive while the UPS is in maintenance. If the bypass does not transfer properly, a preventive maintenance visit can turn into an unplanned outage. Verify:

• Static switching (SCR or IGBT, depending on the design) responds within the specified time.
• The maintenance bypass breaker is mechanically interlocked and operationally labeled.
• The return transfer from bypass to UPS does not produce transients outside IEEE 141.

Firmware Update and Log Review

The UPS logs the events: outages, transfers to battery, alarms, internal warnings. A trained technician reads those logs and detects patterns that are invisible to the facilities team. Firmware updates, when the manufacturer has released stability fixes, are applied with a documented rollback plan. Updating firmware without reviewing logs is malpractice.

What is NOT Included (and Why They Skip It)

The items that are most often left out of preventive maintenance contracts are also the ones that cause the most incidents:

• Battery autonomy test under real load: Requires load bank or downtime coordination. Manufacturers prefer the quick impedance test because it is faster.
• Thermography under load: Requires the building to be in normal operation, not in maintenance window. Many providers avoid the coordination cost.
• Capacitor measurement (ESR, capacitance): Requires removing capacitors and bench testing. Often replaced by “they look fine”, which is not a measurement.
• Harmonics and power quality: Requires a power quality analyzer for at least one week. Not included in the “annual visit” because it changes the scope.

When the contract does not include these four items, the visit is not preventive maintenance: it is a visual inspection with a report.

Real Cadence (Not the One in the Contract)

The annual cadence is what the contract says, not what the equipment needs. Real cadence depends on the environment:

• Hot or dusty environments (Monterrey, Guadalajara, northern border): Quarterly visual inspection + thermography. Annual autonomy test.
• Standard environments (CDMX, Querétaro): Semiannual visual inspection + thermography. Annual autonomy test.
• Critical environments (financial, government, hospital DCs): Monthly visual inspection. Quarterly thermography. Semiannual autonomy test.
• Battery banks older than 3 years: Quarterly impedance or autonomy test, regardless of environment.

Practical Recommendation

If the current provider delivers a “system OK” report in less than 6 hours per UPS, change provider. If the contract does not specify autonomy test under load and thermography under load, expand the scope before signing the renewal. Preventive maintenance is not a line item in the budget: it is the difference between an incident that lasts 4 minutes and one that lasts 4 days. The cost of the full preventive maintenance visit (one working day per UPS) is approximately 8–15% of the cost of replacing the battery bank, and approximately 0.5–1% of the cost of an unplanned critical load outage.

Sources

1. IEEE Std 1188 — Recommended Practice for Maintenance, Testing, and Replacement of Valve-Regulated Lead-Acid (VRLA) Batteries. https://standards.ieee.org/ieee/1188/5677/
2. IEEE Std 1159 — Recommended Practice for Monitoring Electric Power Quality. https://standards.ieee.org/ieee/1159/5698/
3. Vertiv — Electrical safety and compliance services. https://www.vertiv.com/en-us/products/services/electrical-safety-and-compliance/
4. Tripp Lite — UPS preventive maintenance program. https://www.tripplite.com/products/ups-preventive-maintenance
5. Eaton — UPS products and services. https://www.eaton.com/us/en-us/products/backup-power-ups-surge-it-power-distribution.html