Reading Compression, Leak-Down, and Vacuum Results
Interpret compression, leak-down, vacuum and running compression results together, and decide what they prove before anyone opens an engine.
Table of contents
Anyone can screw a gauge into a spark plug hole. The hard part is deciding what the numbers mean, and whether they justify telling a customer that their engine needs to come apart. A wrong call in either direction is expensive. Condemn a good engine and you cost someone thousands. Miss a failing one and a new part goes on a bad foundation.
You know how to run these tests. This is about reading the results together, spotting when a test misleads you, and being able to defend the conclusion to the customer and to the next technician who looks at the car.
What You Will Be Able to Do
- Judge whether a test result can be trusted before you act on it.
- Combine compression, wet test and leak-down results into one conclusion.
- Read a vacuum gauge pattern, and know where the old rules stop working.
- Use running compression to find breathing faults that a cranking test misses.
- Use a relative compression test to screen an engine quickly.
Warning: Disable fuel and ignition as the service information directs before cranking. Keep clear of belts and fans. For any test with the engine running, keep hoses and leads away from moving parts, and run the exhaust extraction.
Can You Trust the Reading?
A bad setup produces numbers that look like an engine fault. Before you draw a conclusion, check these.
| Condition | What it does to the result |
|---|---|
| Weak battery or slow starter | All cylinders read low, and later cylinders read lower |
| Throttle not held open | All cylinders read low |
| Engine fully cold | Readings can run low and uneven |
| Cranking too few strokes | The gauge does not reach its peak |
| Leaking gauge valve or hose fitting | One or more cylinders read low |
| Worn or oily test hose seal | Leak-down shows air that is really escaping at the plug hole |
Cranking speed is the one to watch. Compression depends on how fast the piston moves. If cranking speed falls during the test, the last cylinder tested gets the worst number. If that cylinder looks low, recharge the battery and test it first next time.
Some engines read lower than you expect, even when healthy. Variable valve timing can sit in a position that closes the intake valve late during cranking. Some engines have a compression release to ease starting. Use the published figure for the engine, and compare the cylinders with each other.
Reading Compression Patterns
The percentage rule tells you whether a cylinder is out of line. The pattern of low cylinders tells you where to look next.
| Pattern | What to suspect | Next test |
|---|---|---|
| All cylinders even and normal | Sealing is fine, so the fault is elsewhere | Fuel, ignition, breathing |
| All cylinders even but low | Late cam timing, or an engine worn all over | Check timing before you blame rings |
| Two neighbors low | Head gasket between them | Leak-down, check for air at the next plug hole |
| Low in pairs on a timing belt engine | Valves hit by a jumped belt | Timing check, then leak-down |
| One cylinder near zero | Burned valve, holed piston, broken valve spring | Leak-down, borescope |
| One cylinder much higher than the rest | Carbon build-up, or a cam lobe that is not opening a valve | Borescope, valve cover off |
Never skip the "all even but low" pattern. An engine that has jumped a tooth makes all cylinders look worn. If you quote rings, and the real fault is timing, you just sold a rebuild for a timing belt job.
Worked Example: Reading Three Tests Together
A four-cylinder engine has a rough idle and a misfire on cylinder 3. The engine is warm and the battery is fully charged.
| Test | Cyl 1 | Cyl 2 | Cyl 3 | Cyl 4 |
|---|---|---|---|---|
| Dry compression | 172 psi | 168 psi | 112 psi | 170 psi |
| Wet compression | Not needed | Not needed | 118 psi | Not needed |
| Leak-down | 8 percent | 9 percent | 42 percent | 8 percent |
- Dry test: Cylinder 3 is 35 percent below the highest cylinder: (172 - 112) / 172 = 0.35. That is far outside the usual 10 percent.
- Wet test: It rose by only 6 psi. Oil did little, so rings are not the main leak.
- Leak-down: 42 percent. You hear air at the tailpipe, not at the oil fill cap or throttle body. There are no bubbles in the coolant.
The three tests agree. Cylinder 3 has an exhaust valve that is not sealing. A leak like this usually means a burned valve, a bent valve or a valve held open. Now you can quote a cylinder head job with confidence.
If the tests disagree, stop and find out why. A cylinder with low compression but a good leak-down result often has a mechanical fault that does not show up with the piston at top dead center (TDC). Examples include a worn cam lobe that barely lifts the valve, or a timing problem.
Getting More From a Leak-Down Test
Leak-down testers are not all built the same. There is no common standard for the size of the restriction inside tools sold to shops, so two testers can give different percentages on the same cylinder. Compare cylinders using one tester. Treat percentage bands as a guide and not a law.
- Test at more than one position. Rings wear unevenly along the bore, and wear is usually worst near the top. A cylinder can read differently at TDC than lower in its stroke, if you can hold the crank there with both valves closed.
- Look for scored walls. A leak at the rings on one cylinder only may mean a scored cylinder. Check with a borescope.
- Check the other plug holes. Remove all the plugs. Air hissing from a neighboring hole means a head gasket leak between the cylinders.
- Keep the crank from moving. Air pressure can turn the engine with force if the piston is off TDC. Hold the crank with the tool the service information calls for, not your hand.
Some air always passes the rings, so a soft hiss at the oil fill cap is normal. Judge how loud it is against the other cylinders.
Reading a Vacuum Gauge
An engine is an air pump. Intake manifold vacuum shows how well it pumps, cycle by cycle, without taking anything apart. Connect the gauge to a manifold source below the throttle, with the engine warm and idling.
A healthy engine at sea level usually holds a steady 17 to 21 inches of mercury (inHg) at idle. Air gets thinner as you climb, so expect about 1 inHg less for each 1,000 feet of elevation. At 5,000 feet, a steady 14 inHg can be normal.
| Needle at idle | Likely fault | What to do next |
|---|---|---|
| Steady, but low | Late ignition or cam timing, or an intake leak | Check timing, then smoke test |
| Regular drop at the same point | One valve not sealing, such as a burned valve | Compression and leak-down |
| Random, occasional drop | Sticking valve | Compression test, then inspect the valves |
| Fast flutter at idle that steadies with rpm | Worn valve guides | Oil use check, teardown inspection |
| Slow swing up and down | Mixture or idle control fault, intake leak | Fuel trim and live data |
The regular drop is the most useful pattern. It happens once for each two crank turns, every time the same cylinder comes around, and the drop is often several inches. Compare the fault with a misfire reading on the scan tool to confirm which cylinder it is.
Testing for a Restricted Exhaust
- Warm the engine, and note the idle vacuum.
- Raise and hold the engine at about 2,500 rpm.
- Watch the needle for at least 30 seconds.
Vacuum should drop briefly as the throttle opens, then come back to about the idle reading or higher. If it slowly falls and stays low, the exhaust cannot get out. A plugged catalytic converter is the usual cause. Confirm by measuring exhaust back pressure at an oxygen sensor port, and compare with the service information.
Where the Old Rules Stop Working
These patterns came from older engines. Treat them as clues on a modern engine.
- Variable valve timing can hold the intake valve open longer at idle, which lowers vacuum on a healthy engine.
- Turbocharged engines make boost under load and may run little vacuum at higher throttle openings.
- Diesels often have no throttle, so they run little or no manifold vacuum, and this test does not apply.
Check what the maker publishes, or compare with a known-good vehicle of the same kind.
Running Compression
A cranking test shows that a cylinder can seal. It does not show that it can breathe. A cylinder with a worn cam lobe or a clogged intake port may crank fine and still make no power. A running compression test measures pressure with the engine actually running.
Take out the plug on one cylinder, and fit a gauge with its check valve removed, so the needle follows the live pressure. Disable fuel and ignition to that cylinder only, as the service information directs, so raw fuel does not wash the cylinder. Start the engine and read pressure at idle. Then snap the throttle open and shut quickly, and read the peak.
The figures below are shop rules of thumb, not published specs. Compare with the other cylinders on the same engine.
| Reading | What it often means |
|---|---|
| Idle pressure about half the cranking figure | Normal |
| Snap reading about 80 percent of cranking | Normal |
| Snap reading well under 80 percent | Air cannot get in: worn intake lobe, heavy carbon on the valve |
| Snap reading well over 80 percent | Air cannot get out: worn exhaust lobe, or a plugged exhaust if all cylinders do it |
Worked example: Cranking compression on all six cylinders is about 170 psi. On cylinder 5, the snap reading is 95 psi. 95 / 170 = 0.56, or 56 percent, and the other cylinders read about 135 psi, near 80 percent. That points to a breathing problem on the intake side of cylinder 5. A borescope shows heavy carbon on the intake valve, so you have your cause.
Warning: A running compression test puts tools near a running engine, and the gauge hose can get very hot. Do this with a second person, and keep all hoses and leads clear of the fan and belts.
Relative Compression
A relative compression test is a fast screen. With a lab scope, you clamp a current probe around the starter cable and crank the engine with fuel disabled. Each compression stroke makes the starter pull more current. A healthy engine shows even humps. A weak cylinder shows a lower hump.
It takes minutes and removes no spark plugs. That makes it good for engines where plug access is poor, or for a pre-purchase check. Two limits apply. It only compares cylinders with each other, so if all cylinders are equally weak, the trace still looks even. And you need a sync signal, such as a cam or ignition trigger, to say which hump belongs to which cylinder.
Putting It Into Words
A senior technician writes a finding that another technician could check. Say what you tested, what you measured and what you concluded. For example: "Cylinder 3 dry 112 psi, wet 118 psi, leak-down 42 percent with air at tailpipe. Other cylinders 168 to 172 psi and 8 to 9 percent. Exhaust valve not sealing on cylinder 3. Recommend head removal."
Numbers on the repair order protect the customer and the shop, and they let an advisor explain the work without guessing.
Common Mistakes
| Mistake | What it costs |
|---|---|
| Testing on a weak battery | False low numbers and a wrong quote |
| Calling even low compression "worn out" | Selling a rebuild for a timing fault |
| Comparing leak-down figures between testers | A wrong pass or fail |
| Ignoring altitude on a vacuum reading | A healthy engine gets condemned |
| Treating old vacuum patterns as law on a modern engine | Chasing a fault that is not there |
| Stopping when one test looks bad | Missing the real leak path |
Check Your Understanding
- All six cylinders read about 110 psi, within 5 psi of each other, on an engine rated around 170 psi. What do you check first? Cam timing, because an even low reading can mean a jumped belt or chain, and that is much cheaper than a rebuild.
- Dry compression on cylinder 2 is 105 psi, and it rises to 150 psi wet. Leak-down shows air at the oil fill cap. Do these agree? Yes. Both point to ring sealing. Next, check the bore for scoring and look for why the rings failed.
- Vacuum at idle is a steady 15 inHg in a town at 4,000 feet. Is that a fault? Probably not. Expect about 4 inHg less than at sea level, so 13 to 17 inHg is reasonable. Confirm with other tests.
- Vacuum at 2,500 rpm sinks slowly from 20 to 9 inHg. What do you suspect, and how do you confirm it? A restricted exhaust. Measure back pressure at an oxygen sensor port.
- Cylinder 4 cranks to the same pressure as the others, but its snap reading is far below theirs. What does that suggest? It seals but does not breathe. Look at the intake side, such as a worn lobe or heavy carbon.
Key Takeaways
- Trust a result only after you have checked the battery, throttle, temperature and tools.
- Read the pattern, not just the lowest number. Even low readings suggest timing first.
- Dry, wet and leak-down results should agree. If they do not, keep testing.
- Compare leak-down results using one tester.
- Vacuum patterns are clues, and they need correcting for altitude and engine design.
- Running compression finds breathing faults that a cranking test cannot see.