Waveform Pattern
In-Cylinder Pressure Analysis
Read an in-cylinder pressure capture on a lab scope: peak pressure, the two pockets, exhaust and intake ramps, intake closing and the exhaust plateau. Read sealing, valve events, exhaust flow and intake flow from one trace, then confirm the cause with a second test before you condemn a part.
What the trace actually measures
A pressure transducer in the spark plug hole records cylinder pressure through all four strokes. The plug is out, so the test cylinder doesn't fire: you're looking at a motored cylinder, a sealed volume compressed and expanded by the crank, with the valves opening and closing on top of it.
That makes the trace a direct readout of four physical things:
- Sealing — peak pressure, tower symmetry, and expansion pocket depth.
- Valve events — when the valves let gas move, in crank degrees. That is the effective timing, not the cam card's seat timing.
- Exhaust flow — the pressure the piston pushes against on the exhaust stroke.
- Intake flow — how closely cylinder pressure follows manifold pressure on the intake stroke.
A compression gauge gives one number. This trace gives all four, without removing the head.
What you'll need
- Lab scope, at least 2 channels: the pressure transducer, and the test cylinder's ignition signal if you want the spark angle. A crank sensor channel helps you read angles on your own scope; the app does not read a crank channel on this test yet.
- In-cylinder pressure transducer with a DC (static) output.
- Gasoline engines: a range of about 16 bar (232 psi) or more. Open-throttle and snap-throttle peaks on high-compression engines can approach that ceiling; a flat-topped peak is clipping, not a plateau.
- Diesels: you can use an in-cylinder pressure transducer on a diesel too. Make sure the transducer and setup are rated for the pressures you'll see.
- Spark plug adapter hose with the Schrader (check) valve removed. A check valve blocks everything below atmosphere and destroys the pocket, the exhaust ramp, and the intake readings.
- The same hose on every cylinder you compare. Hoses differ in volume and stretch as they heat; a different hose alone can move the peak 5 to 20 psi. Before you condemn a cylinder, swap the hose and capture it again.
- An angle reference for reading the scope yourself: the crank sensor signal (best), or an ignition signal. The app measures angles from the compression peak of each cycle.
- The engine's bore, stroke, rod length, and compression ratio. If you have the cam card, note the lift it was measured at. A pressure trace shows when the gas moves, not when the valve leaves its seat, so the numbers will not match the card.
Sample rate: degrees per second = RPM × 6, so 0.1° resolution needs 60 × RPM samples per second.
| RPM | °/s | For 0.1° per sample | One 720° cycle |
|---|---|---|---|
| 700 | 4,200 | 42 kS/s | 171.4 ms |
| 2,000 | 12,000 | 120 kS/s | 60.0 ms |
| 2,500 | 15,000 | 150 kS/s | 48.0 ms |
The app needs at least 10 kS/s and recommends 100 kS/s. Record seconds, not milliseconds: every capture must hold several whole engine cycles. A short trigger-centred buffer holds too few.
Sensor and hose delay. A transducer with a 1 ms (10–90%) rise time lags the true pressure by about 0.45 ms on its own, and the hose adds to that: a millisecond or two for the whole chain. Angles read from the peak of the same trace mostly cancel that out, but spark or crank marks on another channel do not. Keep the same sensor, the same hose and the same RPM when you compare.
Step-by-step capture
- Warm the engine to operating temperature.
- Disable the test cylinder.
- Disable fuel to the test cylinder: unplug its injector or turn it off with a scan tool, so no fuel enters the cylinder or the catalyst.
- If you want the spark on the screen, put a spare plug in that cylinder's coil and ground it to the block, away from the transducer. Otherwise unplug the coil.
- Plug-wire systems: connect the lead to a spark tester so the coil still has a path to ground.
- Waste-spark (DIS) coils fire two plugs at once: handle both wires of the pair, or the companion cylinder can still light the fuel.
- Ground the spark away from the transducer. A spark that jumps to the transducer can destroy its electronics.
- Install the adapter and transducer. Make sure the adapter's O-ring or washer seals: an adapter leak looks exactly like a valve or ring leak. Leak-check the hose and its fittings before you trust a low reading.
- Zero the transducer in open air with the engine off and check it reads 0 psi; vacuum reads negative. Use DC coupling. Never use a zoom or AC mode for this test: it removes the pressure levels the test reads.
- Connect the angle reference, and record the operating state: RPM, manifold pressure, coolant temperature, and the commanded cam position on variable-timing engines. Note your altitude, the symptom and any codes before you start.
- Make one continuous recording of about 20 seconds:
- A steady idle
- A throttle snap, held about a second and a half
- Release it and let it coast down
- Back to a steady idle
- Make a separate cranking capture with the engine disabled so it won't start and the throttle closed, to read the pockets. Open the throttle only for a sealing check on the peak.
- If you record the spark, read it beside the pressure, not as a judge of it. Spark after the peak while cranking is normal on many engines; the app grades spark only at a steady idle.
- Keep each run short — a few minutes at most — so the transducer doesn't heat-soak.
The 720° map
Angles are measured from the compression peak, which the app treats as 0°/720°. That peak is a reference, not a true TDC measurement.
| Crank angle | Stroke | What you should see |
|---|---|---|
| 0° | Compression peak | The compression tower. Its peak is the 0° reference |
| 0–~130° | Expansion | Pressure falls; at idle it drops below atmosphere into the expansion pocket |
| ~130–200° | Exhaust valve opening | Pressure ramps up as exhaust gas flows back into the cylinder: the exhaust ramp |
| 180–360° | Exhaust | A plateau near atmosphere: the exhaust plateau |
| ~360–420° | Overlap, intake opening | Pressure ramps down from exhaust level to manifold level: the intake ramp |
| 420–540° | Intake | Pressure at or slightly below manifold pressure |
| ~570–600° | Intake valve closing | At a warm idle the intake's effect ends about 30–60° after BDC; compression begins. Cranking reads earlier |
| 600–720° | Compression | Pressure rises into the next tower; most of the rise comes in the last 30° |
The physics baseline
Here is the physics you can work by hand. The app uses the same reasoning for its comparisons, but it does not print a predicted peak.
Cylinder volume vs crank angle (a = stroke / 2, L = rod length, B = bore, CR = compression ratio):
Vd = (π·B²/4) · stroke Vc = Vd / (CR − 1) V(θ) = Vc + (π·B²/4) · [ L + a − ( a·cosθ + √(L² − a²·sin²θ) ) ]
Compression after the intake valve closes (polytropic):
P(θ) = P_IVC · ( V_IVC / V(θ) )ⁿ
- P_IVC is about manifold absolute pressure at the moment the intake valve closes.
- n is the polytropic exponent. With no heat loss, a compressed mixture at a 10:1 compression ratio reaches about 20–21 times its starting pressure, an equivalent exponent of about 1.30–1.32. Air alone runs a little higher. Wall heat loss lowers the measured n, and leakage lowers it further.
Measure n from your trace with two points on the compression side after the intake valve closes:
n = ln(P₂ / P₁) / ln(V₁ / V₂)
Use absolute pressure (add the barometer reading), include the hose in the clearance volume, and know where TDC really is. Otherwise the n you compute is off.
Plotting ln P against ln V turns the compression stroke into a straight line of slope −n. The point where it starts being straight is where the intake valve has effectively closed.
The adapter adds clearance volume
The adapter hose and transducer add dead volume to the combustion chamber, which lowers the measured peak. On a 6.2 L V8 (103.25 mm bore, 92.0 mm stroke, 10.41:1), an 8 cc hose drops the effective compression ratio from 10.41:1 to about 9.57:1 and costs roughly 10–13% of peak pressure. The app does not correct for hose volume, so use the same hose on every cylinder you compare, and never compare captures taken with different hoses.
Worked example: what the physics predicts for a 6.2 L V8 at idle
These are calculated values, not measurements. They show the range a healthy cylinder should land in, and how much each input matters.
Inputs:
- Bore 103.25 mm, stroke 92.0 mm, compression ratio 10.41:1 (published specification)
- Rod length 154.89 mm (family value; not on the published specification sheet)
- Idle manifold pressure 28–35 kPa absolute
- Effective intake closing assumed between 40° and 60° after BDC (this engine's valve timing is not published)
- n = 1.30, and an 8 cc adapter
| Calculated at idle | Result |
|---|---|
| Swept volume per cylinder | 770.3 cc |
| Peak compression pressure | 45–72 psig |
| Half the pressure drop from peak is reached at | 27.5–27.8° after TDC |
| Pressure at 90° before TDC | −6 to −9 psig |
| Share of the compression rise in the last 30° before TDC | 54–55% |
| Peak at full manifold pressure (100 kPa, e.g. snap throttle) | 199–232 psig |
What this tells you:
- The peak range is wide because idle manifold pressure and intake closing both vary. Measure MAP at capture time and the uncertainty shrinks.
- Measure the half point from the pocket floor. Take halfway between the top of the tower and the pocket floor, not half of the gauge reading. Measured that way, the half point sits close to 30° either side of the peak; the model lands at about 28°. Half of the gauge reading lands at 23–25° on the same model.
- Pressure at 90° before TDC is still well below atmosphere at idle. A rule that says the cylinder should be near atmospheric there only holds at higher manifold pressure, so don't use it at idle.
- The snap peak climbs well above idle, often three to four times, because the cylinder fills closer to atmosphere. The ratio depends on the idle vacuum, so it is a comparison, not a pass mark.
Reading the pattern
Healthy pattern
- Peak repeating within a few percent from cycle to cycle, and in line with the manifold pressure and the altitude you are working at.
- Tower close to symmetric: at half height, the rising side and the falling side are close to the same width, the falling side a touch narrower from heat loss. A tower that clearly leans says the cylinder is losing something, but not whether it's rings or valves.
- The two pockets at about the same level at a steady idle: the expansion pocket and the intake pocket.
- Exhaust ramp centered near BDC (180°).
- Intake ramp centered roughly 20° after exhaust TDC (about 380°); later on engines with variable intake timing, depending on the commanded position.
- Intake closing: at a warm idle the intake valve's effect usually ends 30 to 60 degrees after bottom centre. Cranking reads earlier, so compare like with like.
- Exhaust plateau close to atmosphere at idle.
The ramp and closing windows above are field screening windows used in the trade. When the engine's own effective valve timing is known, measure against that instead.
Low peak, pockets out of line
At a steady idle, an expansion pocket that sits clearly deeper than the intake pocket means charge escaped while the valves were shut. One that sits a lot shallower usually means air is getting in, and the first suspect is the hose or adapter.
A low peak with a pocket out of line is a sealing problem, but the pocket's direction does not tell you where the leak is.
- Leak-check the hose and adapter first, re-seat it and capture again.
- Then confirm with a second witness, such as a cranking capture, the coast-down or a leakdown test, before you name a part.
- Leakage has more time to act at low piston speed, so a marginal leak shows more clearly on a cranking capture than on a running one.
Low peak, but the pockets are level
The cylinder seals but traps less air. Look at the intake closing point. At low speed, cylinder pressure stays near manifold pressure until the valve closes, so a retarded intake cam closes the valve late and the rising piston pushes charge back out. A worn intake lobe lets less air in, and the coast-down capture shows it best. Check the intake ramp too.
Exhaust and intake ramps shifted
| Exhaust ramp | Intake ramp | Most probable cause |
|---|---|---|
| Late | Late by the same amount | Whole-cam retard: stretched chain, jumped tooth, or a phaser parked retarded. On a single-cam engine the phaser moves both events together |
| Late | Normal | Exhaust side only: exhaust phaser, or a worn exhaust lobe (which also rounds the ramp) |
| Normal | Late | Intake side only: intake phaser or intake cam. A late close also lowers the peak |
| Early | Early by the same amount | Whole-cam advance, or a phaser stuck advanced |
Compare valve events only within one state and one RPM. A shift with RPM on a fixed cam is gas flow, not a fault. On variable timing, check what was commanded. Set valve lash first if the engine has adjustable lash.
Shifts are in crank degrees; one cam degree equals two crank degrees.
Engine computers only flag a cam/crank error past a set limit, so a small error can hide from the computer and still show on this trace.
No exhaust ramp; a second tower at 360°
At a steady idle, this means the exhaust valve never opened. The cylinder compresses and expands its gas like a spring, so a tower appears every revolution instead of every other one. Causes: a collapsed lifter, a broken rocker or pushrod, or a flat lobe. On engines with cylinder deactivation, confirm deactivation wasn't commanded during the capture first. In a coast-down, a second tower is normal.
Exhaust plateau high
The piston is pushing against a restriction: a plugged catalyst, a crushed pipe, or a failed muffler.
- At idle the exhaust plateau should sit close to atmospheric. A plugged converter can still look nearly normal at idle, so idle alone won't clear the exhaust.
- A blocked exhaust shows best on a throttle snap. Capture idle and the snap in one recording, and let the plateau's growth from idle to the snap decide. Exhaust-stroke pressure rises normally with flow, so never apply an idle limit at high RPM. Confirm with an exhaust backpressure test before you condemn a converter.
- A motored cylinder reads slightly high. Its exhaust gas is cooler and denser than a firing cylinder's, so the pressure difference across the exhaust valve is larger late in the stroke.
Intake plateau well below manifold pressure
On the intake stroke, cylinder pressure normally runs a little below manifold pressure because of the pressure drop across the valve. A cylinder that sits well below the others is starving: deposits on the intake valve, low valve lift, or a restricted runner.
Cycle-to-cycle variation
Peaks or ramps that move from cycle to cycle at a steady idle point to a sticking valve, a lifter pumping up and bleeding down, or a timing drive that is moving. Compare consecutive cycles in the overlap region around 360°.
Why the reference channel matters
Without a crank signal, angle has to be estimated from time:
θ = 720 · (t − t_peak) / T_cycle RPM = 120 / T_cycle (T_cycle in seconds)
This assumes constant crank speed, but the crank slows on every compression stroke and speeds up after it. At a steady idle that costs well under a degree; at cranking or on a snap it can cost ten degrees or more. A crank sensor channel removes that error, because every tooth edge is a true angle mark. Use it on your scope whenever valve timing is the question. The app does not read a crank channel on this test yet: it measures from time, one cycle at a time, so read valve timing at a steady idle, never off a snap or a decel cycle.
Common mistakes
- Leaving the Schrader valve in the adapter.
- Comparing an idle running peak against a cranking compression spec: the idle peak is low by design.
- Comparing captures taken with different hoses.
- Leaving the injector connected.
- Judging valve timing to a degree or two without a crank reference, or across different RPMs.
- Condemning a valve before leak-checking the hose and adapter.
- Naming a leak path from the pocket's direction alone.
- Splitting idle and the snap into separate clips.
- Judging a cranking capture by idle numbers.
- Calling a deactivated cylinder a failed lifter without knowing the commanded state.
- Applying an idle exhaust-backpressure limit at high RPM.
Frequently asked questions
Why is my idle peak so much lower than the compression spec?
At idle the cylinder only fills to manifold pressure, around a third of an atmosphere. Compression multiplies whatever is trapped, so the peak is correspondingly low: about 45–72 psig on the 6.2 L V8 example, depending on manifold pressure and valve timing. Cranking with the throttle open fills the cylinder closer to atmospheric and reads much higher.
Why does the trace go below zero?
At idle, the trapped charge starts below atmosphere, so both the expansion stroke and the intake stroke pull vacuum. That vacuum section, the expansion pocket, is one of the best leak indicators on the trace. The adapter must not contain a check valve, or you won't see it.
How do I know where TDC is?
The app puts 0° on the compression peak; that is a reference, not a TDC measurement. Heat loss and leakage pull the real peak a little early. The hose and sensor delay push the recorded peak late, and that delay grows with rpm. Only a crank reference and the engine's spec give true TDC.
Can I do this test cranking only?
Yes, for sealing. Crank with the throttle closed to read the pockets; open it only to check the peak. Cranking has its own limits, so never judge a cranking capture by idle numbers. For valve timing, use a steady idle: at cranking speed the crank speed is uneven and the events read earlier.
How is this different from the relative compression test?
Relative compression compares all cylinders against each other through starter current. In-cylinder pressure looks inside one cylinder and adds valve events and exhaust flow. Use relative compression to find the weak position, this test to see what the cylinder is doing, and a leakdown or second witness to confirm the cause.