Somebody at your shop has a stack of oil analysis reports in a drawer. They get filed because nobody is quite sure what to do with a page of elements measured in parts per million, and the machine is running fine, so.
That stack is the most useful maintenance data you own and it is being wasted. Compressor oil analysis costs about the same as a case of oil and it will tell you an airend is failing months before it makes a noise. Here is how to read one.
The one principle that makes the whole report make sense
Trend beats absolute value. Every time.
A single report showing 18 ppm iron means very little on its own. Three reports showing 6, then 9, then 18 ppm iron is a machine telling you something is coming apart. Iron doubling between samples matters more than the number itself, because you do not know what normal looks like for your machine until you have a few samples of it.
So the first thing to do is not interpret one report. It is to sample on a schedule, from the same port, the same way, and build a baseline. A single sample pulled in a panic is worth a fraction of a trend.
The oil condition numbers
These tell you whether the lubricant is still a lubricant.
Viscosity
The headline property. If it has moved off spec, something else is wrong and the viscosity is the symptom rather than the disease.
- Viscosity up generally means oxidation, or contamination with a heavier fluid, or the oil is cooking. A hot machine is the usual culprit.
- Viscosity down generally means contamination with something lighter: water, solvent, or somebody topped it off with the wrong grade.
That last one is common and avoidable. Mixing grades or brands is a real way to ruin a fluid. See our compressor oil guide for what belongs in your machine.
Total acid number
TAN measures acidity, and rising TAN is the classic oxidation signature. Oil oxidizes when it gets hot, and heat is the thing that kills compressor lubricant. A climbing TAN alongside rising viscosity is a machine running hotter than it should, and the fix is usually upstream of the oil: a fouled cooler, a bad thermostatic valve, a hot room, a plugged inlet.
Chase the temperature, not the oil. Our guide to overheating causes and fixes covers where to look.
Water content
Water is the most damaging contaminant on the report and it takes very little of it. It attacks additives, promotes corrosion and destroys film strength.
In a compressor, water in the oil usually means the machine is running too cool, which sounds backwards until you think about it. A rotary screw needs to run hot enough to keep condensate from settling out in the sump. Short cycling, a thermostatic valve stuck open, a cold room, or a machine grossly oversized for its load will all put water in the oil. Our note on short cycling covers that failure mode.
The wear metals, and what each one points at
This is the part that looks intimidating and is actually a lookup table. Each element traces back to a material in the machine.
| Element | Where it usually comes from | What to think |
|---|---|---|
| Iron | Rotors, housings, shafts, gears | The main wear indicator. Watch the trend hard. |
| Copper | Bearings, bushings, bronze parts, cooler tubes | Bearing wear, or a cooler starting to leak |
| Lead and tin | Bearing overlay materials | Bearing distress, often with copper |
| Aluminum | Housings, some bearing cages, and dirt | With silicon, it is dirt ingestion. Without, it is component wear. |
| Chromium | Plated surfaces, some bearing races | Usually accompanies iron |
| Silicon | Dirt and dust, or silicone antifoam additive | The single most actionable number. See below. |
| Sodium and potassium | Coolant, on water cooled machines | Coolant leak into the oil |
| Zinc, phosphorus, calcium | Additive package | Falling values mean the additives are depleting |
Why silicon is the one to look at first
Silicon is dirt. Silicon rising above roughly 25 to 30 ppm, especially paired with rising aluminum, generally means you are ingesting dust, and the usual doors are a failed or bypassing inlet filter, a bad seal, or a breather.
Dirt is not just a contaminant sitting in the oil. It is an abrasive circulating through your bearings and across your rotors, and it is what turns into the iron numbers on your next report. So silicon is a leading indicator: fix it and you prevent the wear, rather than measuring it later.
The fix is almost always cheap. Check the inlet filter, check where the intake is drawing from, and check the seals. A compressor sitting in a blast shop or a woodshop breathing unfiltered room air will show you this pattern every time.
A worked example
Three consecutive samples from a 50 HP rotary screw, taken 1,000 hours apart:
| Parameter | Sample 1 | Sample 2 | Sample 3 |
|---|---|---|---|
| Viscosity | In spec | In spec | Slightly high |
| TAN | Low | Rising | Rising |
| Water | Trace | Trace | Trace |
| Silicon | 8 ppm | 22 ppm | 41 ppm |
| Aluminum | 2 ppm | 5 ppm | 11 ppm |
| Iron | 5 ppm | 7 ppm | 16 ppm |
Read it in order. Silicon and aluminum climbing together is dirt ingestion, and it started between samples 1 and 2. Iron begins climbing one sample later, which is the dirt doing its work on the machine. TAN rising and viscosity creeping up says the machine is also running hot, which is consistent with a restricted inlet.
The whole story is one plugged or bypassing inlet filter, and the report found it before anything failed. Change the filter, find out why it plugged, change the oil because it is now full of abrasive, and sample again at a shortened interval to confirm iron settles back down.
That is a hundred dollar fix instead of an airend rebuild.
Sampling so the numbers mean something
- Sample the same way every time. Same port, same procedure, same lab. Consistency is what makes a trend real.
- Sample warm, with the machine at operating temperature and running or just shut down. Cold oil in a sump has let things settle out, which is exactly what you do not want in the bottle.
- Do not sample from the drain. The bottom of the sump is where the sludge lives and it will overstate everything.
- Keep the bottle and the port clean. Contaminating your own sample produces a scary silicon number and a wasted afternoon.
- Record hours on the oil and hours on the machine. The lab cannot interpret a trend without knowing the interval.
- Sample at a sensible interval. Every 1,000 to 2,000 hours suits most rotary screws, and every oil change at minimum. Tighten it up when something looks off.
What to do with what you find
- Silicon up: inlet filter and seals, now. Then change the oil.
- Iron trending up on its own: shorten the sampling interval and watch. If it keeps climbing, plan the rebuild rather than waiting for the failure.
- Copper up sharply on a water cooled machine: suspect the cooler.
- TAN and viscosity both up: the machine is hot. Fix the cooling before you fix the oil.
- Water present: the machine is running too cool or short cycling. Look at load pattern and the thermostatic valve.
- Additives depleting with everything else normal: the oil is simply used up. Change it, and consider whether your interval is too long.
And the general point: oil analysis is how you change oil on condition instead of on the calendar. On a machine with clean air and a good duty cycle you may safely extend intervals. On a dirty one you will find out you should be changing sooner. Either way you are working from data instead of a number in a manual written for average conditions. Our maintenance schedule covers the rest of the program.
Frequently Asked Questions
How often should I sample compressor oil?
Every 1,000 to 2,000 hours suits most rotary screws, and at minimum pull a sample at every oil change. Shorten the interval when something looks off, because the value is in the trend and you need more points to see one developing.
What does high silicon in compressor oil mean?
Almost always dirt ingestion. Silicon above roughly 25 to 30 ppm, particularly alongside rising aluminum, points to a failed or bypassing inlet filter, a bad seal, or a breather letting dust in. It is worth acting on immediately, because that dirt is abrasive and it becomes your iron reading on the next report.
Why is there water in my compressor oil?
Usually because the machine runs too cool to drive condensate out of the sump. Short cycling, a thermostatic valve stuck open, a cold compressor room, or a machine badly oversized for its load will all do it. Water attacks additives and film strength, so it is worth chasing rather than tolerating.
Is oil analysis worth it on a small compressor?
On a small piston machine, probably not. The oil charge is small and cheap enough that you just change it. On a rotary screw with a meaningful oil charge and an expensive airend, it pays for itself the first time it catches something, and it lets you set oil change intervals from evidence instead of a calendar.
What single number should I look at first?
Silicon, because it is the leading indicator. Iron tells you damage already happened. Silicon tells you the mechanism causing that damage is active right now, and the fix is usually an inexpensive filter rather than a rebuild.
The short version
Sample consistently, build a trend, and read silicon first, then iron, then the oil condition numbers. Chase the cause rather than the symptom, because a hot machine and a dirty inlet explain most reports you will ever see. Done properly, compressor oil analysis is the cheapest early warning system available for the most expensive part of the machine.
Browse our compressor oil and lubricants, or send us a report and your machine model and we will help you read it.
