What a flow number means and what it does not
A flow bench holds a fixed pressure difference across a port and measures the air that passes. The result — cubic feet per minute at a stated depression and a stated valve lift — is a measure of the port's resistance, not of the engine's output. It becomes a horsepower figure only through an empirical correlation, and only for an engine whose cam, induction, exhaust and displacement can actually use the flow.
Two things stop raw flow numbers from being comparable between shops. The first is the test depression. A bench set to 10 in-H₂O reads far lower than the same port on a bench set to 28 in-H₂O, and the ratio is not linear. The second is valve lift: a sheet is a curve, and quoting the peak number without the lift it occurred at hides whether the engine's camshaft ever reaches that lift. This calculator solves the first problem exactly and gives you the tools to reason about the second.
Use it to compare castings honestly, to sanity-check a supplier's claim, and to decide whether the head is the limit in a build before you spend money porting one that already flows more than the target needs.
Why flow scales with the square root of depression
Air passing through a restriction behaves like flow through an orifice. For an incompressible fluid, Bernoulli's relation gives a velocity proportional to the square root of the pressure drop, and since volumetric flow is velocity times area, the flow through a fixed port follows the same rule:
CFM₂ = CFM₁ × √(Δp₂ / Δp₁)
That is the whole conversion. Going from 28 in-H₂O down to 10 in-H₂O multiplies the reading by √(10/28) = 0.5976, so a 250 CFM port reads 149.4 CFM. Going the other way multiplies by √(28/10) = 1.6733. Notice the consequence: quadrupling the test depression only doubles the reading, which is why a shop quoting big numbers at 28 in and a shop quoting modest numbers at 10 in may be measuring identical castings.
The incompressible assumption is reasonable here. 28 in-H₂O is about 1.01 psi, roughly 7% of atmospheric pressure, so density changes across the port are small and the square-root law holds closely. It degrades at very high depressions and where a port chokes, which is one reason benches settled on a common reference in the first place.
The horsepower step is different in kind. It is a correlation drawn from dynamometer results, popularised by David Vizard, that a well-matched naturally aspirated engine makes roughly 0.257 hp for every CFM of intake flow at 28 in-H₂O, per cylinder. Equivalently, each horsepower needs about 3.9 CFM of intake port flow at that depression. It is an upper bound on what the head permits, not a promise about what the engine will make.
Worked example: a 250 CFM head on a V8, and a sheet from another shop
You have a flow sheet showing 250 CFM at 0.600 in lift, taken at 28 in-H₂O, for a V8.
- Already at the reference. The scale factor is √(28/28) = 1, so CFM at 28 in-H₂O is 250.0.
- Horsepower per cylinder. 250 × 0.257 = 64.25 hp.
- Whole engine. 64.25 × 8 = 514 hp of potential.
- Restated at 10 in-H₂O. 250 × √(10/28) = 250 × 0.597614 = 149.4 CFM.
Now a competing shop quotes a head at 160 CFM, tested at 10 in-H₂O. Is it better or worse?
- Normalise it. 160 × √(28/10) = 160 × 1.673320 = 267.7 CFM at 28 in-H₂O.
- Its potential. 267.7 × 0.257 × 8 = 550 hp.
The head that looked like it flowed 160 against your 250 in fact flows 7.1% more — 267.7 ÷ 250 − 1 = 7.1% — and supports about 36 hp more. Comparing the raw figures would have led you to the wrong casting.
Finally, work backwards from a target. For 500 hp on eight cylinders at 0.257: 500 ÷ (0.257 × 8) = 500 ÷ 2.056 = 243.2 CFM per port. Your 250 CFM head has 6.8 CFM in hand, so the head is not what stands between the engine and 500 hp.
How to read the horsepower potential figure
Treat it as a ceiling the cylinder head imposes, not a prediction. Reaching it requires everything else to be matched: a camshaft with enough duration and lift to use the port, an intake manifold and throttle that do not become the new restriction, an exhaust that lets the cylinder blow down, and enough displacement and rpm to move that much air in the first place.
The correlation also assumes the flow figure is one the engine can reach. A sheet peaking at 0.700 in lift is irrelevant to a cam that only opens the valve 0.520 in. Read the flow at your actual net valve lift — work that out with the rocker arm ratio and valve lift calculator if you are changing rockers — and use the flow at that lift, not the peak on the sheet.
When the calculated potential comes out well above what the engine makes on the dyno, the head is not the limit and porting it further will not help. Look at cam timing, intake runner length, exhaust and the induction system instead. When it comes out below the dyno figure, either the sheet is optimistic, the coefficient is being applied to a lift the engine does not reach, or the engine is not naturally aspirated — boosted engines break this correlation entirely, because the compressor supplies pressure the port never had to earn.
Airflow demand also has a cross-check from the other direction. Work out the mass flow the engine needs at your target rpm with the compressor airflow calculator or the induction demand with the carburettor CFM calculator, and see whether the head's ceiling and the engine's demand land in the same place.
Horsepower potential from intake port flow at 28 in-H₂O
| CFM per port at 28 in-H₂O | hp per cylinder | 4 cylinders | 6 cylinders | 8 cylinders |
|---|---|---|---|---|
| 200 | 51.4 | 206 | 308 | 411 |
| 225 | 57.8 | 231 | 347 | 463 |
| 250 | 64.3 | 257 | 386 | 514 |
| 275 | 70.7 | 283 | 424 | 565 |
| 300 | 77.1 | 308 | 463 | 617 |
| 350 | 90.0 | 360 | 540 | 720 |
| 400 | 102.8 | 411 | 617 | 822 |
Rounded to the nearest horsepower. The figures are a ceiling set by the cylinder head, achievable only when cam, induction, exhaust and displacement are matched to it.
Depression conversion factors from 28 in-H₂O
| Target depression (in-H₂O) | Factor from 28 in | 250 CFM becomes |
|---|---|---|
| 10 | 0.5976 | 149.4 |
| 15 | 0.7319 | 183.0 |
| 20 | 0.8452 | 211.3 |
| 25 | 0.9449 | 236.2 |
| 28 | 1.0000 | 250.0 |
| 35 | 1.1180 | 279.5 |
Each factor is the square root of the depression ratio: √(10/28) = 0.5976, √(35/28) = 1.1180.
Mistakes that make a flow comparison worthless
- Comparing depressions without converting. The most common error on forums, and it is worth 67% between a 10 in and a 28 in sheet on the same port.
- Quoting peak flow instead of flow at your lift. A number recorded at 0.700 in lift is meaningless for a cam that opens the valve 0.520 in. Use the row of the sheet that matches your net lift.
- Ignoring exhaust flow. The intake port sets the ceiling, but a poor exhaust-to-intake ratio limits what the engine can reach in practice. Most builders want the exhaust port flowing somewhere around 70 to 80% of the intake, and a head far below that will not deliver its intake-side potential.
- Applying the correlation to a boosted engine. The 0.257 figure comes from naturally aspirated dynamometer results. A compressor supplies the pressure ratio the port would otherwise have to earn, so a boosted engine routinely exceeds the ceiling this page calculates.
- Forgetting that flow is per port, not per engine. Enter the flow for one intake port and let the cylinder count do the multiplication. A total-engine flow figure divided incorrectly is a factor-of-eight error.
Where head flow sits among the other limits
A naturally aspirated engine's power is set by how much air it can pass per unit time, and the cylinder head is only one of four things standing in the way. The other three are induction (throttle, manifold and, on a carburetted engine, venturi area), camshaft timing, and the exhaust's ability to empty the cylinder before the intake opens.
Work in that order. Size the head's flow ceiling here, then check whether the cam has the duration and lift to reach it — the camshaft duration and overlap calculator converts a cam card into the actual valve events — and finally confirm the induction can pass the same mass of air. Any one of the four can cap the engine, and spending on the other three while one remains the limit buys nothing.
If you are pricing a porting job, the calculation to make first is the one at the end of the worked example: divide your target power by the coefficient and the cylinder count to get the CFM each port must supply, and compare that with what the casting already flows at your valve lift. Heads with 20 CFM in hand do not need porting; heads 40 CFM short may not get there at all without a different casting.