What Perkins generator should I buy? That is the question I hear most, and my honest answer starts with it depends, which frustrates anyone who wants a model number. I understand that. Nobody calls an applications engineer because their day was too easy.
But it depends is not a dodge. A generator only makes sense when you know what failure you are protecting against. A 175 kW unit that is perfect for one commercial building would be undersized for a plant full of motor loads. A 1250 kVA unit that keeps a factory alive would be a waste of fuel and floor space in a small office.
I've been an applications engineer in the emergency power business for 11 years, and I've personally made and documented enough specification mistakes to fill a binder. The biggest one happened in 2017, when I sized a generator from a list of nameplate loads and added a comfortable 25% margin. It looked correct until the client's chiller tried to start. The set tripped, the site went dark, and the retrofit cost far more than the better-sized unit would have cost in the first place. That experience became the first entry in our team's pre-purchase checklist.
The oversimplified advice I stopped trusting
Most generator advice says the same thing: add up your loads, add 20%, and buy that size. It's tempting because it turns a complicated decision into a calculator exercise. But it ignores motor-starting current, load sequencing, transfer switch behavior, fuel system design, and what happens when the generator actually has to carry the building at 2 a.m.
It's tempting to think that generator selection is just a comparison of kW numbers. In practice, identical loads can behave completely differently depending on how they start and how they are sequenced.
The machines that fail aren't always the ones that are too small on paper. Sometimes they are the ones that were never tested under real load. That is why I now start every conversation with a different question: which scenario are you in?
Three scenarios, three different answers
After more than a decade of emergency power work, nearly every serious request falls into one of three categories:
- Scenario A: You need permanent backup power and there is no reliable unit on site.
- Scenario B: You already have a generator, but it has become unreliable and you need to decide whether to repair it or replace it.
- Scenario C: You need power before any new equipment can be delivered, so speed and certainty matter more than price.
Scenario A: You are buying permanent protection
If losing power creates a safety issue, spoils product, or stops revenue, you are in new-generator territory. Buy the smallest unit that can handle the actual worst case, then verify that worst case with a load study.
For commercial facilities with elevators, pumps, lighting, telecom, and a modest block of critical loads, a 175 kW Perkins electric generator is often the right class. It has enough muscle for a real building, but what makes it reliable is the transfer switch sequence and the starting capacity for motor loads. I have seen 175 kW sets trip at sites where two large motors started at the same time, even though the steady-state load looked fine.
For larger industrial operations with process loads or big motor starts, a 1250 kVA Perkins generator is a more realistic building block. At 0.8 power factor, that is roughly 1000 kW of output. It is not an upgrade from a 175 kW unit. If the facility does not need that kind of power, a 1250 kVA set is just a larger fuel bill and a heavier maintenance schedule.
Use the rating definitions in ISO 8528 as a reference. A standby-rated set is not designed to carry a facility indefinitely after a disaster. If you expect weeks of outages, you need to talk about prime power, fuel storage, oil sampling, and a maintenance plan. And use NFPA 110 as your acceptance testing checklist. A generator that has never started under full load is a theory, not a backup system.
Scenario B: You already have a generator that scares you
These calls sound different. The caller usually says it starts most of the time, or it ran fine during last month's test. The word most is a red flag. For emergency power, probably is not a strategy.
The first mistake people make here is applying car-engine logic to a stationary diesel. I once worked with a client who was convinced his set had a fuel pump problem. He had spent the weekend reading about the fuel pump chrysler 200 owners replace in their cars. In his mind, an engine is an engine. But a gasoline car's electric in-tank pump and a diesel generator's lift pump and injection system are completely different systems. In his case, the real problem was a clogged fuel filter and an air leak on the suction side. No pump replacement was needed.
The same confusion appears with the phrase oil in spark plug well. That is a real symptom on spark-ignited gasoline or natural gas engines; it usually means a leaking valve cover gasket or bad spark plug tube seals. Left alone, oil can cause a misfire at the worst possible moment. But a Perkins diesel has no spark plugs, so that specific symptom does not apply to a typical diesel generator. If oil is appearing around a diesel engine, look at the valve cover, injector seals, or crankcase breather before assuming the engine is dying.
How to check for power with a multimeter
Before you replace expensive parts, do a basic electrical check. The question I get most is how to check for power with a multimeter without making a dangerous mistake. Here is the short version:
- Prove your multimeter on a known live source first. A meter that lies to you is worse than no meter at all.
- Set the meter to AC voltage, not resistance and not continuity.
- Measure at the generator output terminals, line to line and line to neutral. A 208V or 480V class set should show approximately that voltage at rated speed.
- If voltage is present at the generator terminals but missing at the load side of the transfer switch, the problem is in the switching or control logic, not the engine.
Use a CAT III rated meter around panels, and never change the range setting while the probes are connected to live terminals.
Once you have a diagnosis, the repair-versus-replace decision gets clearer. If the engine itself has low compression, high oil consumption, or serious mechanical wear, replacement is usually the honest answer. If the problem is a fuel filter, a lift pump, a sensor, or a tired automatic transfer switch, repair can be completely rational. The key is to test the unit under load after the repair. If you cannot verify that it carries the building, you have not actually fixed the uncertainty.
Scenario C: You need power before the lead time allows
This is the scenario that changes everything. When a deadline is fixed, waiting for a cheaper quote or a longer lead time is not a cost decision anymore. It is a risk decision.
In September 2023, I documented a rush order where the client paid a meaningful premium for guaranteed delivery and a commissioning engineer to be on site before a scheduled plant startup. The alternative was waiting for a lower-priced delivery that might arrive in time. The plant's downtime cost was roughly $35,000 per day, so the premium was not an expense. It was insurance.
Something I've learned the hard way: an urgent situation is not the time for probably. If a supplier says probably by Friday, they are asking you to carry the risk. A firm commitment, a written delivery date, and a defined commissioning plan are what you are actually paying for. The premium buys certainty, not just speed.
How to tell which scenario you are in
If you are unsure, do not ask a salesperson to guess. Ask yourself two questions.
First, what does one hour without power cost? Include lost production, spoiled materials, idle staff, safety risks, and contractual penalties. If you cannot calculate that number, you are not ready to buy a generator yet.
Second, how much warning do you have before the power is needed? If you have months, work through Scenario A properly. If you already own a generator that is misbehaving, start with Scenario B and get a real diagnosis. If your deadline is fixed and close, treat Scenario C as temporary insurance while the permanent solution is being built.
There is no cheapest generator in the abstract. A low quote is worthless if the unit arrives late, fails under load, or cannot be serviced when you need it. The buyers who do best are the ones who understand what failure they are protecting against and pay for the certainty that matters most.
