-
Who This Checklist Is For
-
The 7-Step Operational Checklist
-
Step 1: Build Your TCO Baseline Before You Send a Single RFQ
-
Step 2: Calculate Real Load — Don't Trust "Probably Enough"
-
Step 3: Put Eaton 93PM in the Comparison — But Beyond the Spec Sheet
-
Step 4: Read the Eaton UPS Manual — Focus on Three Sections
-
Step 5: The 150kW Diesel Generator Match — The Step Most People Skip
-
Step 6: AC Contactor Replacement — Price It Before It Surprises You
-
Step 7: Battery Charger Wiring — The Project Isn't Done Until This Is Signed
-
Step 1: Build Your TCO Baseline Before You Send a Single RFQ
-
Common Mistakes and Things to Watch
Who This Checklist Is For
This checklist is for anyone planning a power system replacement in a data center or industrial facility — specifically, the person who approves the budget, compares quotes, and signs off on the technical agreement. If you've got a UPS replacement or expansion project in the 100-300kW range coming down the pipe, these 7 steps will help you sidestep some genuinely expensive mistakes.
My situation: I manage the power and infrastructure procurement budget at a roughly 500-person company. That's about $340,000 annually, and I've been handling it for six years. I've compared quotes from 40+ vendors, built more than a few TCO spreadsheets, and lived through the classic "lowest bidder takes two months longer than promised" scenario. The biggest lesson I've learned? A low unit price is often just the down payment on a more expensive project. For context: when I shifted our procurement policy from "lowest bid wins" to "mandatory TCO comparison" in 2023, our hidden costs on power maintenance dropped by about 27% that year.
The 7-Step Operational Checklist
Step 1: Build Your TCO Baseline Before You Send a Single RFQ
Most people blast out three RFQs, get three quotes, and pick the cheapest one. My approach is different — spend half a day building a TCO table first. Rows are cost categories over the next five years. Columns are candidate vendors. At minimum: unit price, installation and commissioning, old equipment removal, battery replacement cycle (typically 3–5 years), spare parts kit, annual maintenance contract, and those sneaky "hidden" items like live-transfer testing labor.
I went back and forth between the incumbent brand and a newer, cheaper option for close to three weeks on one project. The incumbent quoted about 18% higher, but bundled parallel commissioning and a 5-year battery replacement cycle. The new entrant looked better on paper — until I noticed the $2,300 line item for parallel commissioning, and a 3-year battery cycle spec. Finance kept pushing for the savings. I held out until I ran the five-year TCO: the incumbent came out 11% cheaper over the equipment's life. Bottom line — once the TCO is on paper, that decision becomes a no-brainer.
Checkpoint: Your TCO table should have at least 12 line items and cover a 5-year horizon before you send any RFQ.
Step 2: Calculate Real Load — Don't Trust "Probably Enough"
Get this wrong and everything downstream is wasted. You need actual load curve data — at least 12 months of daily peak and evening peak readings. The three numbers that matter: peak load (kW), power factor, and expected growth over the next 2–3 years.
Here's a thing nobody tells you upfront: the design engineer's standard estimate and the actual load from the operations log can differ by 20% or more (I learned this the hard way on a 2022 project). We sized for 200kVA based on design values, then discovered during acceptance testing that actual peak load was around 130kW. Oversized. The annual efficiency loss from light-load operation plus extra maintenance ran about $3,500. Conversely, older data centers adding new equipment tend to underestimate load growth, not overestimate it.
Checkpoint: Get a signed load calculation sheet from your electrical engineer. It must include three numbers: peak load, power factor, and 3-year growth projection.
Step 3: Put Eaton 93PM in the Comparison — But Beyond the Spec Sheet
For data center applications in the 100–300kW range, the Eaton 93PM UPS will almost certainly land on your shortlist. In the projects I've run, its strengths are modular architecture (you can scale capacity as load grows), a stable efficiency curve across the 30%–70% load band, and a fairly mature parallel configuration option. Those are real advantages for data center use.
But here's my caution: spec sheet numbers and on-site reality are two different things. You still need to cross-check input voltage tolerance, bypass requirements, and upstream breaker/cable sizing against your electrical drawings. I ran a side-by-side comparison in 2023 — same 200kW capacity, the 93PM and a competitor were within a rounding error of each other on full-load efficiency, but at 40% light load they diverged by about 1.5 percentage points. At our electricity rates, that was roughly $1,100 per year. Peak efficiency numbers are basically marketing — the efficiency curve is what matters.
Checkpoint: Add the efficiency curve (not the single peak number) for each candidate model to your TCO table, annotated with the efficiency value at your facility's typical load factor.
Step 4: Read the Eaton UPS Manual — Focus on Three Sections
A procurement manager doesn't need to study circuit diagrams. You're reading the manual for two reasons: to confirm technical parameter definitions and to write acceptance terms into the contract. I focus on three sections: the input/output specification table, battery and charger requirements, and the fault code table.
Specifically, in the Eaton UPS manual, check the input voltage and frequency tolerance ranges, the battery charger input voltage and charging current range (this matters for Step 7), and the minimum/maximum operating temperature in the installation environment chapter. Why temperature? Because it directly determines whether you're looking at a facility air conditioning upgrade — a hidden cost that can run into tens of thousands of dollars (and one that gets left out of budgets more often than you'd think, unfortunately).
According to the battery charger section of the Eaton 93PM user manual (2023 revision), the charger's input voltage range and charging current settings must match the battery bank configuration, and the manual explicitly requires a full-power charge/discharge test after installation. Those details should become acceptance criteria in your contract.
Checkpoint: Screenshot the relevant manual pages and file them in your procurement record as the technical acceptance basis.
Step 5: The 150kW Diesel Generator Match — The Step Most People Skip
Most power replacement projects treat the UPS itself as the boundary. The thing that actually trips you up when the grid goes down is upstream backup power. If you've got a 150kW diesel generator on site, you need to calculate: generator capacity, UPS input power, charger power draw, startup current, and the generator's ability to handle nonlinear loads.
Common misconception: people assume "generator bigger than UPS equals fine," but the UPS charger pulls extra current when the battery is depleted. A 150kW generator running a 200kVA UPS looks sufficient — until grid power returns and the battery charger kicks in at full power, stacked on top of downstream IT load, pushing the generator toward its limit. My approach: model the "grid disconnected + battery recharge" condition as a separate line item. In Q2 2024, we ran a live transfer test and caught this exact scenario — generator frequency dipped 0.8Hz, the battery charger tripped on protection, and we paid about $6,000 for a new startup module. Ugh.
Checkpoint: Require a signed "generator-UPS matching calculation" from your electrical engineer. It must include the worst-case condition of "depleted battery + full load."
Step 6: AC Contactor Replacement — Price It Before It Surprises You
Power system replacements are rarely isolated projects. Often, when the UPS gets swapped, the AC contactors in the upstream distribution panel are also due (typically an 8–10 year cycle) and should be replaced. This is the kind of scope item that quietly becomes a change order at acceptance.
My current practice: make ac contactor replacement a separate line item in the TCO table during project scoping. Include the contactor bodies, arc chutes, auxiliary contacts, labor, and the cost of coordinating an outage window. If coordinating a single outage is expensive (e.g., you need business units to shift workloads to off-peak), then bundle the contactor replacement and UPS replacement into the same outage window — do both jobs in one power-down. I've run the numbers: combined versus separate outages, the labor and business-impact delta is usually $2,000–$4,000 depending on site size.
Checkpoint: Confirm on your outage window schedule that "contactor replacement" is stacked into the same window as the UPS work.
Step 7: Battery Charger Wiring — The Project Isn't Done Until This Is Signed
A lot of projects get signed off once "the UPS powers on and runs." But if the battery charger wiring (the how to connect battery charger part of the scope) isn't properly verified, you're setting yourself up for a miserable troubleshooting session later.
Even if you outsource this to an electrical contractor, as a procurement person you should know what the acceptance points are. I require three deliverables before acceptance: the as-built charger wiring diagram, actual measured charging current records, and battery equalize/float transition test logs. All three, or it's not done. Why the measured records? Because correct drawings and correct wiring don't guarantee correct parameters — we had a 2022 project where the drawing was perfectly accurate, but the charger DIP switches were set wrong, causing chronic undercharging. The battery bank swelled up in under two years, and we ate about $4,000 in replacement costs. After that, "measured records" became a standard contract clause.
Checkpoint: Three documents, no exceptions. The live test records must be signed by both parties.
Common Mistakes and Things to Watch
- Comparing unit prices instead of TCO. I've said this repeatedly, but I still see people rubber-stamping the lowest total on three quote sheets.
- Treating lead time as a verbal promise. If your project has a hard go-live date (tenant move-in, audit, or aging equipment past its service life), the rush fee and time-certainty clause deserve to be written into the contract separately. The premium for guaranteed delivery is usually a rounding error compared to the cost of a missed deadline. Trust me on this one.
- Ignoring the hidden cost of outage windows. One unschedulable outage window can drag a project out 2–3 weeks. Price that into the plan upfront.
- Accepting "it powers on" as acceptance. Battery charger, contactors, transfer logic — each one needs measured test records.
- Comparing old models against new ones. If your existing system is 8+ years old, recalculate load and TCO from scratch. Don't carry forward legacy assumptions.
One last thing that took me a while to internalize: on power systems, paying a bit more for a solution with high delivery certainty tends to be cheaper over the long run. Especially when there's a hard deadline involved — the premium for time certainty is almost always minor compared to the cost of delay. Sounds like common sense, but when the moment comes, most people still get pulled in by the quote that's 10% cheaper. Bottom line? The money you save upfront tends to come back through a different door later.

