Generator low load problems begin below 40% of the rating, where the cylinders stay under 300 °C, the fuel does not finish burning, and the bore oil thins with diesel. Tesla Power sets out the measurements that show the damage, the loading rule a site should calculate, and the records a supplier quote should promise.
- Read the load log as a share of the rating, not as a kW figure on its own.
- Sample the oil at 250 hours and test for fuel dilution, because the exhaust tells you later than the sump does.
- Hold 60% of the rating for an hour a week, which burns off the deposit a low load leaves.
- Compare the frame against the 100kVA set on the WP4 engine before buying a larger machine for a load that does not fill it.
What counts as low load on a diesel generator?
Low load is a share of the rating, so the same kW figure is a problem on one frame and normal work on another. A 200kVA set at 40kW is running at 25% of its plate and sits in the wet stacking zone, while a 100kVA set at 40kW is at 50% and runs clean. The figure that matters is the ratio, and the ratio comes from the plate rather than from the meter.
The band a site should aim for is 60% to 80% of the rating, with the engine held there for at least an hour at a time. Below 40% the exhaust temperature falls under the point where the deposit burns off, and above 90% the set is working without reserve for the next motor start. We write the target band into the commissioning record of every set Tesla Power ships to a site with a load that varies, and the band is read against the ISO 8528-1 rating limits rather than invented for the site.
Is a lightly loaded set cheaper to run?
It burns less fuel per hour and costs more per kWh, which is the opposite of what a site expects. A set at 25% of rating uses about 6 L/h to make 50kW, so the fuel figure is close to 0.12 L/kWh only on paper; the real cost arrives later as oil changes, injector work and a shorter engine life.
Why does wet stacking start below 40% of the rating?
Combustion needs three things at once: enough air, enough heat in the cylinder wall, and a spray pattern fine enough to burn in the time available. At low load the turbocharger delivers little boost, so the air charge is thin. The cylinder wall cools because less fuel burns against it. The result is a flame that quenches against the metal before it finishes, and the unburnt fraction leaves as liquid.
That liquid goes two places, and both cost money. Some leaves through the exhaust port and shows as a dark wet film at the manifold joint. Some runs past the rings into the sump, where it thins the oil and lowers its film strength. Fuel dilution above 5% in an oil sample is the point where bearing wear begins to accelerate on a set that logs long hours. Tesla Power reads that film at the flange before it reads the sample, because the film appears weeks before the dilution figure moves.
- Exhaust temperature below 300 °C at the manifold.
- Wet, sooty film at the exhaust joint and at the turbo outlet.
- Oil level rising between changes, which is diesel entering the sump.
Which symptoms show first on a low load set?
The exhaust port tells a technician more than the panel does, and it tells it earlier. A set that has been stacking for a month shows a dark, oily film at the manifold flange. A set that has been stacking for a year shows a visible plume on load pickup, because the deposit that collected in the exhaust system burns off when the engine finally reaches temperature. Tesla Power photographs the flange at commissioning, so a later claim has a cleaned surface to compare against.
Season changes the picture more than most operators expect. On a generator low load in winter the film at the flange is thicker, because the intake air is colder and the flame quenches sooner, while the same set in summer may show nothing at all. A generator for the night shift runs long hours at that condition, which is why hotels and hospitals see the film before a factory does. A site that inspects only in warm months can carry a stacking problem for a year without seeing it.
| Symptom | What it shows | Reading to check |
|---|---|---|
| Oily film at the manifold | Unburnt fuel leaving the port | Exhaust temperature under 300 °C |
| Blue plume on load pickup | Deposit burning off the exhaust | Turbo boost slower than the sheet |
| Oil level above the mark | Diesel entering the sump | Fuel dilution above 5% |
| Rough idle and misfire | Injector nozzle coked | Balance test across cylinders |
How do you measure the load a site actually holds?
A clamp meter read once at the panel is not a measurement, because it catches whatever the site was doing at that minute. The figure that matters is the distribution over a week: how many hours sat below 40%, how many sat in the band, and what the highest half hour looked like. A load logger on the incoming cables answers all three questions for the price of a service visit, and it answers them in a form a claim can use.
Read the log against the plate rather than against the kW figure on its own. A hotel that draws 80kW for six hours and 20kW for eighteen hours holds 40% for a quarter of the day and sits in the wet stacking zone for the rest. The same log on a 250kVA frame tells the opposite story, and the difference decides whether the fix is a loading rule or a smaller set.
What does a week of logging cost?
A three phase logger with a week of memory is a small item against the price of an engine rebuild, and most suppliers will lend one with a service visit. Tesla Power Equipment sends the logger with the commissioning engineer on sites where the load is described as variable, because the log settles an argument that a single reading cannot.
What does a load bank run prove on a partly loaded set?
It proves the set can still reach its rating after months of light work, which is the question a site cannot answer from its own load. A resistive bank at 60% of the plate, held for an hour, raises the exhaust temperature into the cleaning band and shows whether the turbo still builds boost on demand. It also re-seats the rings, which is why a set that has been idle runs better after the test.
The reading to watch is the exhaust temperature at the end of the hour. A set that reaches the figure on its data sheet has a clean exhaust path; a set that stops short has a blocked intercooler, a coked turbo or a restricted silencer. That single figure decides whether the site needs a service or a strip-down, and it costs an hour to obtain.
- Bank at 60% of rating for one hour, no steps.
- Exhaust temperature recorded every 15 minutes.
- Boost pressure compared against the data sheet.
Tesla Power Equipment runs the same hour on sets that arrive for service with a stacking complaint, and the temperature curve it records becomes the baseline for the next visit. A hospital in the United States files that curve beside the NFPA 110 load test intervals, because the standard asks for the same evidence in a different format and a surveyor reads both.
Worked example: a hotel set held under 30%
A 250kVA set in a hotel ran the kitchen and the corridors at night and carried the whole building only during a grid outage. The log showed 16kW for 14 hours a day, which is 16kW ÷ 0.8 = 20kVA, or 8% of the plate. Over eight months the oil level rose 4 mm on the dipstick between changes, and the sample came back with 7% fuel dilution.
The fix cost less than a new frame. The hotel moved the laundry and the pool pump onto the set for two hours each morning, which lifted the load to 150kVA, or 60% of rating, and burned the deposit off within three weeks. The oil change interval returned to 500 hours, and the dilution figure fell below 2% by the second sample.
How does low load shorten engine life?
Three mechanisms work at once, and each one is slow enough to be missed between services. Diesel in the oil lowers the film strength, so the bearing surfaces run closer together. The deposit on the exhaust side restricts flow, so the turbo works against a higher back pressure. The injector nozzles coke, so the spray pattern degrades and the next start is harder than the last.
The combined effect shows as a set that starts poorly at 4,000 hours and needs injectors at 6,000 hours, where a loaded set reaches both figures later. A site that logs its hours against its load can see the curve before the failures arrive, and the load factor figures show what a normal curve looks like for a hotel, a mine and a data hall. Tesla Power traces the same three mechanisms in the oil sample before it recommends injector work.
Does idling the set make it worse?
Idling is the extreme case of the same problem, because the engine makes no useful power while it wears. A set that idles for twenty minutes before it takes load spends that time at the worst possible condition, which is why a load acceptance test starts the set and applies load inside a minute.
Which fixes work without replacing the frame?
Four fixes work on the frame a site already owns, and they are ranked by what they cost. Move existing load onto the set so it holds 60% for an hour. Fit a load bank or a resistive heater bank that the controller switches in when the site load falls. Add a second set and run the two in parallel so one carries the base load at 70%. Replace the frame with one sized to the real load.
- Move laundry, pumps and compressors onto the set for a daily hour.
- Fit a controlled load bank that switches in below 40%.
- A generator in parallel with a second set holds 70% of the rating.
- Replace the frame when the real load stays under 30%.
When should the frame be replaced instead?
Replacement makes sense when the real load stays under 30% of the plate for most of the year and the site cannot move any more load onto the set. A 500kVA frame carrying 100kVA is a machine working at a fifth of its design point, and no loading rule will change that. The arithmetic is simple: the fuel, the service hours and the wear all follow the frame, not the load. A generator low load problems case for a hotel usually ends with the kitchen moved onto the set rather than with a new frame.
A smaller frame also starts better on a cold morning, because the same battery and starter turn a smaller mass. The camp or hotel that replaces a large idle frame with a properly sized one usually reports a shorter start, a cleaner exhaust and a lower service bill in the first year, and the prime and standby ratings explain why the replacement should be chosen on the duty rather than on the peak. Tesla Power compares the two options on fuel and service hours before a site commits to either.
Which records prove a low load fault at claim time?
Three records decide a claim on an engine damaged by light load: the load log, the oil sample history and the service book. The log shows the duty, the samples show the dilution curve, and the service book shows whether the site followed the interval for the duty it was running. A site that keeps all three argues from data rather than from memory, and the first question a warranty desk asks is answered by the log.
- Load log covering the 30 days before the fault.
- Oil samples with the dilution figure for each change.
- Service book with the interval and the technician's name.
Chinese generator suppliers often issue a schedule written for continuous duty and leave the low load clause out, so the engineering notes on service planning are worth reading before the schedule is signed. Most generator manufacturers will add the loading rule on request, because the rule protects the engine rather than the seller, and Yangzhou Tesla Power writes that rule as a line with a number on it rather than as a paragraph of advice. The generator low load problems documentation for a claim then carries the same figure the site was asked to hold, and the conformity assessment for the site installation starts from a schedule both sides signed.
Frequently asked questions about generator low load problems
Between 60% and 80% of the rating is the band a site can hold for weeks without deposit building in the exhaust. Below 50% the engine still runs, but the exhaust temperature falls under the cleaning point and the oil begins to collect diesel between changes.
Deposit builds within a month of running under 30%, and the measurable damage to bearings begins when fuel dilution passes 5%. A set that runs at that condition for a year without an oil sample usually needs injector work before it needs a rebuild.
It can, when the deposit is soft. Two hours at 80% of rating raises the exhaust temperature enough to burn the film off, and the plume that appears during the run is the deposit leaving. A set with hard carbon on the valves needs a strip-down rather than a bank.
Less than the engine, but it does affect it. A lightly loaded alternator runs cooler and its winding insulation absorbs moisture in a humid climate, which is why a set held at low load in a coastal yard should be run loaded once a month rather than idled.
It is cheaper per kWh, because the engine works closer to its design point and the fuel burn per unit falls. The saving is large enough that a site running under 30% for most of the year usually pays for a replacement frame inside three years of fuel and service costs.
At 20% of rating the burn is close to 4 L/h on a 250kVA frame, which looks cheap against 40 L/h at full load and costs more per kWh. Ask a supplier for the curve rather than the single figure, and check the point on the curve that matches the site's real duty.
The rule for a lightly loaded set fits on one line: hold 60% of the rating for an hour a week and sample the oil every 250 hours. Send the load log to our export desk and Tesla Power will read it against the plate, then name the fix that costs least. We keep the oil sample history against the serial number, so a low load claim is answered from the file rather than from a conversation.