What a CFM rating actually describes
A carburettor's CFM figure is a flow measurement taken on a test bench at a stated pressure drop. Pull air through the carburettor until the depression across it reaches the reference value, measure the volume flow, and that is its rating. For a four-barrel the reference is 1.5 inches of mercury; for a two-barrel it is 3.0 inches. Nothing about the number describes the engine.
Your job is to work out how much air the engine wants at peak power and then choose a carburettor that can pass it without needlessly exceeding it. The engine side of the calculation is pure geometry plus one efficiency term. A four-stroke fills each cylinder once every two crank revolutions, so it swallows half its displacement per revolution. At 6,000 rpm a 350 cubic inch engine therefore moves 350 ÷ 2 × 6,000 = 1,050,000 cubic inches of air per minute if it filled perfectly, which is 1,050,000 ÷ 1,728 = 607.6 cubic feet per minute.
No engine fills perfectly at every speed. Volumetric efficiency is the fraction of the theoretical volume actually drawn in, and it is what turns 607.6 CFM into a realistic 516 CFM at 85%. VE is not a constant: it peaks near the torque peak and falls off either side, which is why you size at the power peak, where the product of speed and VE is largest.
The engine's displacement comes from bore, stroke and cylinder count — check yours with the displacement calculator if the block has been bored or stroked, since a nominal size can be several cubic inches off.
Why the divisor is 3456, and what the depression means
Two conversions are folded into a single constant. There are 1,728 cubic inches in a cubic foot, and a four-stroke engine completes an intake stroke in each cylinder once per two revolutions. Multiply those and you get 3,456, so CFM = CID × RPM × VE ÷ 3456. A two-stroke draws a fresh charge every revolution, so it drops the factor of two and uses 1,728 — which is why a 100 cubic inch two-stroke wants as much carburettor as a 200 cubic inch four-stroke.
Volumetric efficiency is the honest part of the calculation. A stock engine with a mild cam and restrictive exhaust typically runs 75 to 85% at its power peak. A well-developed street engine with good heads and headers reaches 85 to 95%. A race engine with a tuned intake and exhaust can exceed 100%, because intake ram effects and exhaust scavenging push more mass in than the piston alone would draw. Those bands are practitioner rules of thumb — if you have dyno airflow data, use the measured value instead.
The test depression matters when you compare ratings. Flow through an orifice varies with the square root of the pressure difference across it, so a carburettor tested at 3.0 inHg reads √2 = 1.414 times the flow it would show at 1.5 inHg. A 500 CFM two-barrel and a 500 CFM four-barrel are therefore not the same carburettor: the two-barrel flows 500 at twice the depression, which corresponds to about 354 CFM on the four-barrel scale. This calculator reports the two-barrel equivalent so you can compare like with like.
The same square-root law explains why a carburettor is never a hard flow limit. Ask a 750 CFM four-barrel for more than 750 and it will deliver it — at a higher depression, which costs manifold pressure and therefore power. The rating is the flow at a defined restriction, not a wall.
Worked example: 350 in³ street engine at 6,000 rpm
A 350 cubic inch V8 with aluminium heads, a hydraulic roller cam and headers, making peak power at 6,000 rpm, with an estimated 85% volumetric efficiency.
- Theoretical volume per minute. 350 × 6,000 = 2,100,000 in3 per two revolutions’ worth of cylinders, which is 2,100,000 ÷ 3,456 = 607.6 CFM at 100% VE.
- Apply volumetric efficiency. 607.6 × 0.85 = 516.5 CFM.
- Convert to air mass. Standard air is about 0.0765 lb/ft3, so 516.5 × 0.0765 = 39.5 lb/min. That mass figure is the one you carry to a compressor map or to fuel system sizing.
- Two-barrel equivalent. 516.5 × √2 = 730.4 CFM on the 3.0 inHg scale.
So a 600 CFM four-barrel is a sensible street choice and a 650 gives a little headroom. Check the 750 you were considering: 750 × 3,456 ÷ (350 × 0.85) = 8,713 rpm before it becomes the restriction. That is far beyond where this engine makes power, so the 750 buys nothing at the top and costs signal at the bottom.
Now change one thing: raise VE to 100% with a race intake and a bigger cam, and move the power peak to 7,000 rpm. 350 × 7,000 × 1.00 ÷ 3,456 = 708.9 CFM, and now the 750 is exactly right. The carburettor did not change; the engine's ability to use it did.
Choosing between the sizes on the shelf
Size to the calculated figure and round up to the next available carburettor, but be far more cautious about going up than the top-end numbers suggest. Oversizing costs real driveability.
A carburettor meters fuel using the pressure signal generated by air accelerating through the venturi, and that signal grows with the square of velocity. Fit a carburettor twice as large as the engine needs and the air moves at half the velocity, so the signal falls to a quarter of what the metering circuits were designed around. The result is lazy throttle response, poor part-throttle metering, and a bog on tip-in — not a top-end gain. Vacuum secondaries mitigate this by keeping the secondary barrels shut until the primaries generate enough signal, which is why a vacuum-secondary carburettor tolerates being slightly oversized far better than a mechanical-secondary one does.
Undersizing is more forgiving than most people expect, because the carburettor's rating is a flow at a reference depression rather than a hard limit. A slightly small carburettor costs a few percent of peak power and nothing else. That is why the traditional advice — when in doubt, go one size smaller — survives.
Throttle bodies on fuel-injected engines are sized by the same airflow requirement but with an important difference: there is no venturi, no fuel metering signal, and therefore no low-speed penalty for being generous. A throttle body is a valve, so oversizing it costs only throttle-pedal resolution.
Two engine-side factors move the requirement. Forced induction raises the mass of air the engine consumes, so a blow-through carburettor or a boosted throttle body must handle the density from your boost calculation. And whatever the induction, the fuel system has to match the air: work out injector flow from the power the airflow supports rather than from displacement.
Required CFM at 85% volumetric efficiency
| Displacement (in³) | 4,500 rpm | 5,500 rpm | 6,500 rpm | 7,500 rpm |
|---|---|---|---|---|
| 302 | 334 | 409 | 483 | 557 |
| 350 | 387 | 473 | 560 | 646 |
| 383 | 424 | 518 | 612 | 707 |
| 454 | 502 | 614 | 726 | 837 |
Figures are four-barrel CFM at the 1.5 inHg reference. Multiply by 1.414 for the equivalent two-barrel rating at 3.0 inHg.
Mistakes and limits
- Sizing at the redline instead of the power peak. Airflow demand falls once VE drops away above peak power, so sizing at redline oversizes the carburettor.
- Guessing volumetric efficiency optimistically. Entering 100% on an engine that actually makes 82% oversizes by more than 20%, which is a whole carburettor size.
- Comparing two-barrel and four-barrel ratings directly. They are measured at different depressions and differ by a factor of 1.414.
- Assuming CFM is a hard ceiling. A carburettor will pass more than its rating; it just does so at a higher pressure drop, which costs manifold pressure.
- Applying carburettor logic to a throttle body. A throttle body has no metering signal, so the penalty for being generous is negligible.
- Ignoring what is behind the carburettor. Intake manifold, port and valve flow are usually the real restriction. A larger carburettor cannot fix a head that will not flow.
Key terms
- Volumetric efficiency
- The fraction of the theoretical swept volume the engine actually inducts per cycle, measured at a given rpm.
- Test depression
- The pressure drop across the carburettor during flow testing — 1.5 inHg for four-barrels, 3.0 inHg for two-barrels.
- Venturi signal
- The pressure drop the venturi creates, which the metering circuits use to pull fuel. It varies with the square of air velocity.
- Vacuum secondaries
- Secondary throttle plates opened by manifold and venturi signal rather than mechanically, so they stay shut until the engine can use them.
Air mass, and where this number goes next
The most useful output here is often not the CFM at all — it is the air mass flow in pounds per minute. Volume flow depends on air density and therefore on temperature and altitude; mass flow is what actually determines power and fuel requirement. A dyno pull on a cold morning at sea level and one on a hot day in the mountains can show the same CFM through the carburettor and very different power, because the mass in each cubic foot is different.
Mass flow also connects this page to everything downstream. Roughly speaking, a naturally aspirated gasoline engine makes on the order of 9 to 10 horsepower per pound per minute of air, since burning stoichiometrically at 14.7:1 fixes the fuel mass that air can support. The 39.5 lb/min in the worked example therefore corresponds to a plausible 360 to 400 hp for that engine, which is a useful reality check on both the VE you assumed and the power you were hoping for. Convert that power to torque at the peak with the torque and rpm calculator.
For forced induction, the mass flow figure is the horizontal axis of a compressor map, and the vertical axis is the pressure ratio. Compute the naturally aspirated mass flow here, multiply by the density ratio from the boost calculator, and you have the operating point to plot.
One historical note that still matters: the 1.5 inHg four-barrel and 3.0 inHg two-barrel conventions come from the era when carburettor manufacturers needed a common basis for comparison, and they persist in every catalogue today. When you see a carburettor advertised only as “600 CFM”, check the barrel count before you compare it with anything else. The number without its depression is incomplete.