Turbine Oil Varnish Explained: Why Bearings Lacquer While the Oil Tests Clean
Varnish is the visible end stage of oil oxidation, not contamination. That distinction explains turbine lubrication's most confusing failure: a unit whose oil report looks acceptable while bearing pads lacquer and servo valves stick. The deposit-forming fraction stays dissolved until it does not — so no filter removes it, and no routine report measures it.
Why a clean sample and varnished bearings coexist
A 60 MW extraction steam turbine, 14 months after an oil change: bearing drain temperatures 4–6 °C above baseline, the extraction servo valve hunting at low load, filter life halved. The report read viscosity on grade, 180 ppm water, acid number near new, ISO cleanliness 17/15/12 — and “continue monitoring”. At inspection the pads were lacquered amber-brown, the spool lands sticky, and the cooler outlet header carried a varnish ring.
Every number in that report was correct. None of them measured insoluble soft contaminants.
Q: Does varnish always come with a high acid number?
A: No. Acid number tracks acidic products in the bulk oil; the deposit-forming fraction is a different set of molecules. An oil can be near new-oil acid number and still be saturated with soft contaminants.
The mechanism: solubility falls as temperature falls
Oxidation products begin soluble and invisible. Solubility is finite, and it moves the way most operators do not expect: the oil holds more of these products when hot and less when cool. Fluid leaving a 60 °C reservoir and passing through a 45 °C bearing therefore arrives supersaturated, and deposits first on the coolest metal it meets and where flow is quiet — bearing housings, cooler tubes, valve bodies, dead legs. The tank, hottest and best mixed, stays superficially clean. This is why raising tank temperature to drive off moisture makes varnish worse, and why cold shutdowns accelerate it.
Q: Where should the sample be taken, then?
A: From a live, turbulent point downstream of the bearings — a drain line or the return header before the tank. Reservoir samples understate soft-contaminant load, so keep the point fixed.
Four routes to varnish
- Oxidation — the dominant route; rate roughly doubles per 10 °C, and RPVOT tracks the antioxidant reserve.
- Thermal degradation — bearing metal, immersion heaters, shaft seals and high-shear points crack oil locally.
- Contamination and incompatibility — mixed brands, phosphate ester from control fluid, process gas carry-over, fine metal particles acting as catalysts.
- Additive depletion — as antioxidants are consumed, the oil loses its ability to keep degradation products dissolved.
Q: Can two good oils form varnish together?
A: Yes, and it is a common root cause in multi-supplier plants. Each product is validated alone, never in combination. Qualify one product per system, and treat a supplier change as a controlled changeover with sampling.
What to monitor
| Indicator | Test | Watch | Act |
|---|---|---|---|
| Varnish potential (ΔE) | MPC, ASTM D7843 | 15–30 | above 30 |
| Oxidation reserve | RPVOT, ASTM D2272 | below 50% of new | below 25% of new |
| Water | ASTM D6304 | 200–300 ppm | above 500 ppm |
| Air release | ASTM D3427 | rising trend | 50% beyond spec |
| Foam | ASTM D892 | rising trend | exceeds spec |
ASTM D4378 is the reference practice. MPC thresholds are guidance, not universal limits, and RPVOT measures oxidation reserve rather than deposit tendency — judge trends, not single results.
Q: Why is MPC missing from our standard report?
A: Because it is a specialist test and many routine packages are built around cheaper bulk properties. Request it explicitly, and specify your condition-monitoring suite when tendering the laboratory contract.
Removing it: three routes, used together
- Correct the root cause — air ingress and suction sealing, tank level and residence time, cooler performance, one oil product in circulation.
- Mitigate in service — ion-exchange, electrostatic or adsorbent kidney loops remove soft contaminants over weeks; they slow new deposition rather than dissolve aged lacquer.
- Clean offline against a written criterion — full-system flush verified by MPC ΔE, membrane patch inspection and ISO cleanliness, with bearing temperatures back to baseline.
Q: Can varnish be removed without a shutdown?
A: Often yes, with a kidney loop run over weeks. It will not free an already-sticking spool valve or dissolve lacquer baked onto a pad for a year.
What to ask a supplier
- RPVOT of the new oil, and the MPC baseline — you cannot trend either without a start value
- Base oil group — group II or III generally outperform group I at equal viscosity
- Air release and foam at operating temperature, not only at 40 °C
- A written compatibility statement against the oil currently in the system
Q: Is a higher RPVOT always better?
A: Not in isolation. Very high values can come from packages with inferior air release or water separation. RPVOT is one number among several and says nothing about deposit mechanics alone.
Without a starting point there is no trend — and with no trend, the next discovery happens at the bearing.