Here's the uncomfortable truth I wish someone had told me in 2018: a Perkins generator is not automatically the right answer just because the nameplate says Perkins. I know that sounds like a cop-out in a market that wants a clean 'yes' or 'no,' but after seven years, roughly 340 generator purchase orders, and about $18,000 in documented mistakes, I've become convinced that most buyers get burned not by bad equipment — they get burned by assuming 'Perkins' guarantees a specific outcome regardless of what they bolted onto it.
I want to walk you through the two orders that taught me this, because the lessons apply far beyond just Perkins. If you're sourcing a 40 kW Perkins generator for a small industrial site, evaluating a 125 kW Perkins industrial generator for a commercial building, or just trying to figure out whether to go propane vs gas generator on a whole house diesel generator install — the failures below will save you money.
What I got wrong: the 'Perkins means reliable' assumption
In my first year on this job (2017), I made the classic assume the brand does the work mistake. We had a client — a small cold-storage warehouse — that needed backup power for a 40 kW load. Easy, I thought. Sourced a 40 kW Perkins generator, spec sheet matched the load, ordered it, scheduled the install. Done, right?
Two weeks after commissioning, the building lost power for about 90 minutes during a grid event. The generator started. Then it didn't hold. The load was right at the edge of the unit's continuous rating, and the inrush current from the compressor bank tripped it. I had matched steady-state kW but completely ignored motor starting surge. That error cost $890 in emergency callout, plus a week of downtime and a very unhappy facilities manager who then spent three months bad-mouthing us to other tenants in the park.
What I should have done: verified the surge profile with the client's electrician before spec'ing the unit. Perkins diesel engines are reliable — that part is genuinely true and I'll defend it — but a 40 kW generator sized to a nameplate that doesn't account for a 3x inrush is a 40 kW generator that's going to trip every time that compressor cycles on. Assumed the spec matched the load. Didn't verify the surge profile. Turned out the compressor alone pulled 82 kW for about a second on startup.
Where the remote start quiet generator promise broke down
The second, more expensive lesson came in September 2022. A commercial client insisted on a remote start quiet generator configuration — their words, not mine. They wanted the unit to be unheard from the adjacent office building and to be startable from a phone from anywhere. Fair enough as a request. The problem is that 'quiet' and 'remote start' are both solutions that interact with the specifics of a site in ways the brochures don't show.
The unit we supplied was a 125 kW Perkins industrial generator with an aftermarket acoustic enclosure and a cellular remote-start module. On paper, everything checked out. In practice, the acoustic enclosure was rated for 75 dB(A) at 7 meters — which sounded great until we realized the adjacent building was 4 meters away and the client had an employee office directly above the enclosure pad. And the remote start module's cellular dependency dropped out during a regional carrier outage in November, at exactly the moment the client needed it.
That particular disaster ran us about $3,200 in re-work — enclosure re-baffling, a hardwired dual-SIM failover, additional pad isolation — plus roughly two weeks of client patience. Not the biggest loss I've ever recorded, but the one that changed how I spec quiet generators forever: dB(A) at a distance is a label; perceived noise at a listener's location is the spec. If your 'quiet' generator has a bedroom or office within 5 meters of the enclosure, the number on the datasheet is not the number that matters.
“In my experience, roughly one in five generator specs that look right on paper fall apart on site — and the acoustic and remote-start categories are the two biggest offenders.”
The propane vs gas generator question buyers keep getting wrong
I want to address this directly because it comes up on nearly every whole-house diesel generator project I've touched, and buyers frequently confuse it with the fuel question. Propane vs gas generator is not a fuel-neutral comparison. Propane burns cleaner, stores indefinitely, and doesn't degrade like gasoline — those are real advantages. But per unit of energy, propane has lower energy density than diesel, which is why a propane genset usually needs a larger displacement engine to hit the same kW output as a diesel unit like a Perkins-based system. That translates to a bigger footprint, more fuel storage volume, and — critically for industrial buyers — a bigger monthly refueling commitment if you're on a supply contract.
Here's what I tell people now: if you're sizing a standby unit for a building that might sit untouched for 11 months and run for 40 hours a year, propane or natural gas is often the smarter, lower-maintenance choice. If you're running an industrial facility with a 125 kW characteristic load and you might need 200+ hours of runtime in a bad grid year, the diesel efficiency and the depth of the Perkins service network probably outweigh the convenience. That's an honest limitation on both sides — neither fuel type is universally 'better.'
A practical way to know which side you're on
Before you get a quote, run two numbers: annual expected run-hours, and peak kW plus surge. If your annual run-hours are under 100 and your peak load fits comfortably within a mid-size propane or NG unit's continuous rating, don't let anyone talk you into a diesel because it's 'what the pros use.' If your run-hours are 150+ and your surge profile is heavy (motors, compressors, welding gear), the diesel path — including a properly sized Perkins — is going to be cheaper to own and easier to service over 10 years. That's the heuristic that's held up across the last ~150 orders I've tracked.
So why do I still recommend Perkins when it fits?
Because the failures above weren't Perkins failures. They were sizing failures and spec-interpretation failures. The Perkins diesel engines in those units — including the ones on the two bad jobs — did exactly what they were rated to do, and the service network answered the phone when I called at 6 a.m. on a Saturday. That's not nothing. I've watched buyers save $1,800 upfront by going with a cheaper engine brand and spend $4,000 in three years on parts-availability delays. The total cost conversation almost always favors the units with the deep parts network.
But here's the honest limitation I'll put in writing, because I think it's the most useful thing I can say: a Perkins generator is a poor choice for buyers who aren't going to do the sizing homework. If you're going to buy based purely on nameplate kW and price, and you're not going to verify surge load, site noise conditions, fuel infrastructure, or the service interval plan, then the brand on the engine isn't going to save you. You'll get a unit that starts fine in week one and fails you in the specific way you didn't think to check in month fourteen. I've been that buyer. It's expensive.
If you don't have the time or the internal expertise to run a proper load and site study — and plenty of small operations don't — the right move isn't to skip the study. It's to hire someone to run it, or to buy from a supplier who will run it before quoting. That's the 20% of buyers I'd tell to pause: if the sizing step feels like a formality, you're probably about to become my next $3,200 mistake.
What this means if you're in the 80%
For everyone else — the buyer who has a load calculation, a noise-sensitive site that's been surveyed, a fuel plan, and a service schedule — the multi-power-band coverage (roughly 20 kW to 1250 kVA) and the depth of Perkins' service network are the reasons I keep shortlisting the brand after a decade of doing this. The trick isn't finding the best generator. It's being honest about which category you're in before you spend the money. I didn't have that honesty in 2017, and it cost me $4,800 across two jobs to learn it. If this article saves you that, I'll consider the writeup paid for.