3 Mistakes I Made Specifying Generators (And the Checklist That Fixed It)

I’ve been handling generator procurement for mission-critical facilities for about eight years now. In that time, I’ve personally made and documented three significant mistakes that totaled roughly $47,000 in wasted budget and avoidable delays.

This isn’t a theoretical guide. This is the checklist I wish I’d had when I started. It’s built around the specific errors I made, so you can skip the tuition I paid.

There are three steps. Follow them in order.

Step 1: Lock Down the Load Profile Before You Call a Supplier

My first major mistake happened in January 2019. I specified a Caterpillar 3412 diesel generator for a new data center wing. Rated at 500 kW standby. Looked perfect on paper. The problem? The load wasn't what I assumed.

The inrush current from the UPS system and the simultaneous startup sequence of three HVAC units created a transient load that the 3412 couldn't handle. Voltage dipped below acceptable thresholds for 400 milliseconds. The critical load switched to battery, and the generator spent its first six months in nuisance trip mode.

The fix required adding a soft-start kit and re-tuning the governor. That cost $3,200 in parts and labor, plus a 1-week delay in commissioning. (Should mention: the equipment was already on-site. The alternative specification—a 3512 with a larger alternator—would have cost $4,500 more upfront. But it would have worked on day one.)

The lesson: You need to know peak starting kVA, not just running kW. Get the actual load study from the electrical engineer. Don’t guess.

Checkpoint for Step 1:

• Do you have the starting sequence of all motor loads documented?
• Have you calculated the voltage dip at the generator terminals during the worst-case step load?
• Is your load factor above 30%? (Running a caterpillar gas generator under 20% load for extended periods will cause bore glazing.)

Step 2: Understand What 'Emergency' Actually Means for Your Facility

In September 2022, I ordered a caterpillar generator 6500-class unit for a remote oilfield communications shack. The spec called for emergency backup power. I chose a standard diesel unit with a 10-second automatic transfer switch (ATS). Simple, right?

Wrong. The shack didn’t need backup power—it needed continuous power. The grid was unreliable. Outages averaged 4 hours per week. The generator would be running 60 hours a month, not 10 hours a year. That's a completely different use case.

The diesel unit started wet-stacking within three months. The service interval was every 250 hours—meaning I was changing oil every month. Logistics to that site cost $450 per trip. The whole thing was a disaster.

Looking back, I should have specified a caterpillar gas generator for that application. Natural gas is cleaner for continuous runtime, and the maintenance interval is longer. Or, I should have at least specified a larger fuel tank and a load-bank cycle to handle the light-load running.

If I could redo that decision, I’d match the generator type to the actual duty cycle, not just the label. But given what I knew then—standard industry practice dictates diesel for backup—my choice wasn’t crazy. It was just wrong for the real-world load profile.

Checkpoint for Step 2:

• What is the expected annual runtime? (Standby < 200 hours / Prime < 8,760 hours / Continuous = 8,760 hours)
• Is the fuel type selected for the duty cycle? (Diesel for standby; natural gas or diesel for prime/continuous with load management)
• Have you budgeted for scheduled maintenance based on runtime, not calendar days?

Step 3: Budget for the 'After' — Not Just the 'Before'

My third mistake was about total cost of ownership. In early 2023, I was comparing a Predator 3500 inverter generator vs Honda for a mobile workshop application. The Predator was $800. The Honda was $2,200. I know the conventional wisdom, but the budget was tight. I went with the Predator.

The Predator worked fine for three months. Then the inverter board failed. No warranty coverage because the unit was used commercially (despite what the box said). A replacement board cost $350 plus shipping. Then the carburetor gummed up. I replaced it with a generic for $40. Then the recoil starter broke.

By month nine, I had spent $1,200 on repairs on top of the initial $800. The unit was still unreliable. I bought the Honda. That was two years ago. It has needed only the scheduled oil changes.

Saved $1,400 by buying cheap. Ended up spending about $2,000 on the cheap unit plus the eventual Honda. Lost work time in between. Net loss: significantly more than just the money. (Should mention: this isn't a knock on all parts. We stock Kobalt 40V battery charger units for light tools—they're fine. But for a generator running critical loads, reliability has a price.)

Checkpoint for Step 3:

• What is the expected service life? (Consumer < 1,000 hours / Commercial > 10,000 hours)
• Are parts and service available locally for that brand?
• What is the total cost of ownership over 5 years, not just the purchase price?

Final Note: The Rush Fee Trade-Off

In March 2024, we paid $400 extra for rush delivery on a caterpillar generator control panel. The standard lead time would have been three weeks. We needed it in five days to avoid delaying a factory startup. The alternative was missing a $15,000 per day production schedule.

The electric charger for the battery bank was also expedited. We paid a 35% premium on that item. Total cost for the two items: about $1,200 extra.

Was it worth it? Yes, because the cost of missing the deadline was orders of magnitude larger. That’s the time certainty premium in action. In an emergency, spend the money for the guaranteed delivery.

But—and here’s the caveat—you can only rush if you’ve done Steps 1 and 2 correctly. If you rush the wrong spec, you just get the wrong part faster.