When to repair vs replace your UPS: the 4 signs a maintenance provider hides from you

A 10-year-old UPS can fail you in the month you need it most. The repair-or-replace decision is not technical: it is financial with downtime risk. These are the four signals that a serious maintenance provider should flag (and that many keep quiet about until it gets expensive).

Signal #1: the batteries are past their second life

VRLA batteries have a useful life of 3-5 years in normal operation. If you have already replaced them once and the UPS is over 8 years old, you are in the second generation of batteries. When you start seeing 'battery aging' codes or backup runtime dropping more than 20% versus spec, the cost of a new battery bank (USD 4-12k depending on capacity) can approach 30-50% of the cost of a new UPS. At that point, replacement beats new batteries.

Signal #2: the DC link electrolytic capacitors

Rectifier and inverter capacitors degrade silently. When your maintenance provider reports 'capacitor ESR drift' on more than 30% of the bank, you are 12-24 months from a catastrophic rectifier or inverter failure. Replacing them costs USD 8-18k in parts and specialized labor (not every technician can desolder banks at that height). If your UPS brand has discontinued parts, replacing the UPS is cheaper than waiting for the failure.

Signal #3: the firmware no longer receives security updates

Modern UPS units have Linux-embedded controllers and SNMP, Modbus, or REST connections to the DCIM. If the manufacturer discontinued firmware support (3+ years without patches), you have an unpatched network vulnerability likely. In a DC with serious segmentation this is a minor issue; in one where the UPS shares a VLAN with the rest, it is a real attack vector. Replacing to get updated firmware is valid in some compliance regimes (PCI-DSS 4.0, ISO 27001).

Signal #4: your maintenance provider's mean time between failures

This signal is not about the UPS: it is about your provider. If in the last 24 months they had more than two incidents where the UPS was out of service for more than 4 hours on scheduled maintenance (not failures), the question is not about the UPS but about the provider. Switching provider costs one audit and a month of transition; switching UPS costs more and solves the same problem only if the root cause was the equipment, not the service.

When it IS worth repairing (not replacing)

The decision flips when:

• The UPS is less than 5 years old, the brand has active spare parts, and repair cost is less than 25% of a new unit.
• The root cause is a single redundant module (fan, communication card) and the UPS kept running in N+1 redundancy while being repaired.
• You have a maintenance contract with 4h response and a local spare-parts bank. Even if the gear is old, operational risk is mitigated by the service.

When it IS worth replacing

• The UPS is standby (offline) topology and critical loads exceed 5 kVA. A double-conversion online unit lowers operating cost 8-15% via efficiency and eliminates harmful transients.
• The load grew more than 30% over the original sizing and the UPS already runs at 75-85% capacity. Buying another of the same in parallel costs more than one correctly sized.
• The manufacturer went out of business or discontinued the line. Buying spares on the secondary market at 2x their normal price is not sustainable.

Before signing a replacement, request a real load study (not nameplate). A UPS running at 40% load is an undersized UPS: switch to a smaller one and the savings pay back in 18-30 months through efficiency and cheaper batteries.

Sources

1. Vertiv — News and Insights (recursos técnicos oficiales sobre UPS y ciclo de vida). https://www.vertiv.com/en-us/about/news-and-insights/
2. ANSI/TIA-942-C — Telecommunications Infrastructure Standard for Data Centers (estándar oficial, mayo 2024). https://tiaonline.org/resource/tia-942-c-data-center-infrastructure-standard/