The Problem
A facility running critical communications and monitoring equipment was experiencing random power drops. Not full blackouts - just enough instability to crash systems, corrupt data, and trigger false alarms. The existing UPS and rectifier setup had been installed years earlier with no documentation and no load planning.
Every time something went down, the response was to restart it and hope. No one had investigated why the power was unstable in the first place.
The Approach
We treated this as an engineering problem, not a hardware swap.
- Audited the full power chain: mains input, distribution board, UPS, -48VDC rectifiers, and load distribution
- Found the rectifier bank was undersized for the actual load - running at 95%+ capacity with no headroom
- Identified poor earthing and loose terminations introducing voltage ripple
- Redesigned the power layer with correctly rated rectifiers, proper load balancing, and clean cabling
- Added monitoring so the client can see power health before problems surface
The Outcome
Zero unexpected power drops since the redesign. Systems stay up through mains fluctuations. The client has visibility into power load and battery health for the first time. No more guessing, no more restarts.
Lessons
Power problems are almost never about the battery or the UPS alone. They're about the entire chain - sizing, termination quality, earthing, and load management. If you only look at one piece, you'll keep replacing parts without fixing anything.