Carbonation is an equilibrium, not a dose
Put a keg of beer under CO2 pressure and gas crosses into the liquid until the beer is saturated for that pressure and that temperature. Turn the pressure up and more dissolves; let the beer warm and gas comes back out. There is no point at which the beer is "finished" — it is always chasing the equilibrium set by the two conditions you control. That is why a carbonation figure alone means nothing: 12 psi is not a carbonation level, it is a pressure that produces 2.7 volumes at 38 °F and 1.9 volumes at 54 °F.
Henry's law is the underlying physics: the amount of gas dissolved in a liquid is proportional to the partial pressure of that gas above it, with a proportionality constant that falls as temperature rises. Beer is not a simple solution and CO2 is not a perfectly ideal gas, so brewers work from measured tables rather than from theory. The equation on this page is a polynomial fitted to those tables, giving pressure directly in psi from a Fahrenheit temperature and a carbonation target in volumes.
The practical consequence is worth stating plainly: the pressure you set and the temperature you keep are one setting, not two. If your kegerator drifts from 38 to 44 °F over a hot weekend and the regulator does not move, the beer sheds gas until it matches the new equilibrium — around 2.2 volumes instead of 2.5 — and the excess vents through the relief valve. Bring it back down and you have to wait days for it to reabsorb.
Why the beer line has to fight the pressure
The pressure that holds carbonation is almost always more than the pressure that gives a good pour. Beer arriving at the tap at 12 psi is moving fast and drops pressure instantly at the faucet, so the dissolved CO2 comes out of solution in the glass as foam. The fix is to make the beer line absorb the surplus through friction, so that beer reaches the faucet at roughly 1 psi and a flow of about one gallon every two minutes.
Three terms make up the balance. Line resistance is the big one and depends steeply on internal diameter: 3/16 in vinyl offers about 3 psi per foot, while 1/4 in vinyl offers only about 0.85 and 3/8 in barely 0.2. Going one size up in tubing roughly quadruples the length you need. Vertical rise costs about 0.5 psi per foot of lift from the keg outlet to the tap, which matters in a tower and not at all in a chest freezer. Tap loss is conventionally taken as 1 psi for a standard faucet.
So the balanced length is (P − 1 − 0.5h) / r. At 12 psi with a 2 ft rise on 3/16 in vinyl that is (12 − 1 − 1) ÷ 3 = 3.3 ft. Homebrewers routinely run 5 to 10 ft instead, which over-restricts slightly and gives a slow, quiet pour — a deliberate and sensible choice, since a line that is too long only costs you patience while one that is too short costs you a glass of foam.
If your pour is foamy and the line is already long, look elsewhere first: a warm faucet or a warm section of line, a partially closed shutoff, or beer that is simply over-carbonated because the regulator has been sitting at 25 psi since you burst-carbonated it.
Worked example: a 40 °F kegerator at 2.5 volumes
Corny keg in a converted fridge holding the beer at 40 °F, target 2.5 volumes, 3/16 in vinyl line, and 2 ft of rise from the keg outlet up to the shank.
- Temperature-only terms. −16.6999 − (0.0101059 × 40) + (0.00116512 × 1600) = −16.6999 − 0.40424 + 1.86419 = −15.23994.
- Temperature–volume term. 0.173354 × 40 × 2.5 = 17.33540.
- Volume terms. (4.24267 × 2.5) − (0.0684226 × 6.25) = 10.60668 − 0.42764 = 10.17904.
- Pressure. −15.23994 + 17.33540 + 10.17904 = 12.27 psi, which is 0.846 bar.
- Pressure the line must absorb. 12.27 − 1 (tap) − 1 (2 ft × 0.5 psi/ft) = 10.27 psi.
- Balanced line length. 10.27 ÷ 3 = 3.4 ft. Round up to 5 ft in practice for a slower, cleaner pour.
- Dissolved CO2. 2.5 × 1.96 = 4.9 g/L.
Now cool the same keg to 34 °F and repeat: the pressure drops to 9.6 psi for the identical carbonation, and the balanced line shortens to 2.5 ft. Colder beer holds gas more willingly, which is why commercial cellars run cold and why raising your fridge temperature to "serve it warmer" quietly changes your carbonation.
Setting the pressure and knowing when it has worked
Set and forget is the reliable method. Chill the keg to serving temperature, set the regulator to the calculated pressure, connect the gas and leave it for one to two weeks. Nothing can go wrong, because the equilibrium is the target — leave it a month and it is still correct. The cost is time.
Burst carbonating trades safety for speed: pressurise to 25–30 psi for 24 to 48 hours, then bleed the keg down and reset the regulator to the serving pressure. It works because the higher pressure drives gas in faster, but there is no equilibrium to protect you and 12 hours of inattention makes a keg you cannot pour for a week. If you burst carbonate, set a timer, and always finish by dropping back to the calculated serving pressure.
Judge the result by pouring, not by looking at the gauge. A keg mid-way to equilibrium pours flat at the top of the glass with visible bubbles clinging to the sides; a keg at target pours with a head that holds. If you have a keg you did not carbonate yourself, use the reverse lookup: enter the pressure the previous owner was running and it tells you where the beer has settled.
Pick the target from the style, the same numbers you would use for bottles. Most ales and lagers on tap sit at 2.4 to 2.6 volumes, British cask-style ales at 1.2 to 1.8, German wheat beers at 3.0 to 4.0. The priming sugar calculator uses the identical scale, so a recipe carbonated one way transfers directly to the other. Note that force carbonating a keg and bottling from it are different problems — a beer kegged at 2.6 volumes and then bottled from the tap will lose gas in transfer.
Carbonation pressure chart (psi)
| Beer temp | 2.0 vol | 2.2 vol | 2.4 vol | 2.6 vol | 2.8 vol | 3.0 vol |
|---|---|---|---|---|---|---|
| 34 °F / 1.1 °C | 4.3 | 6.3 | 8.2 | 10.2 | 12.1 | 14.1 |
| 38 °F / 3.3 °C | 6.0 | 8.1 | 10.2 | 12.3 | 14.4 | 16.5 |
| 42 °F / 5.6 °C | 7.7 | 10.0 | 12.2 | 14.4 | 16.7 | 18.9 |
| 46 °F / 7.8 °C | 9.5 | 11.8 | 14.2 | 16.6 | 19.0 | 21.3 |
| 50 °F / 10.0 °C | 11.3 | 13.8 | 16.3 | 18.8 | 21.3 | 23.8 |
| 54 °F / 12.2 °C | 13.1 | 15.7 | 18.4 | 21.1 | 23.7 | 26.3 |
| 58 °F / 14.4 °C | 15.0 | 17.8 | 20.6 | 23.3 | 26.1 | 28.9 |
Read across a row to see the cost of extra carbonation and down a column to see the cost of a warm fridge: holding 2.4 volumes needs 8.2 psi at 34 °F and 20.6 psi at 58 °F. These figures assume pure CO₂ in the headspace; a nitrogen blend changes them entirely.
A CO₂ and nitrogen blend does not follow this chart
Only the partial pressure of carbon dioxide sets carbonation. Nitrogen is nearly insoluble in beer, so a 70/30 nitrogen-CO2 blend run at 30 psi puts just 9 psi of CO2 partial pressure on the beer, which at 38 °F holds about 2.3 volumes — while the total 30 psi pushes the beer through a restrictor plate to make the tight cascading head that nitro dispense is used for. If you are running a blend, use this chart on the CO2 partial pressure only: multiply your gauge reading by the CO2 fraction of the mix.
Why a keg pours wrong
- The regulator is still at burst pressure. The most common cause of a foamy keg by a wide margin. Bleed the headspace and reset to the serving figure.
- The beer line is too short for the pressure. Three feet of 3/16 in vinyl balances about 10 psi; a 6 ft line handles a 19 psi system. Lengthening the line is cheaper than every other fix.
- The fridge is not as cold as the dial says. Measure the beer, not the air. A keg reading 46 °F when you thought it was 38 is carrying half a volume less than you calculated.
- Warm faucets and warm tower. Beer that warms in the last foot of line breaks out of solution at the faucet no matter how well the system is balanced. A tower fan or a cooled shank solves it.
- A partially closed shutoff or a kinked line. A restriction concentrated at one point causes turbulence and foam, unlike the even friction of a long line.
- Confusing carbonation with pressure. The gauge reading is meaningless without the temperature it belongs to — check what your setting actually delivers with the reverse lookup above.
Force carbonation next to the alternatives
Force carbonation is the most controllable of the three ways to carbonate beer, and the only reversible one. Over-carbonated? Vent and wait. Under-carbonated? Raise the pressure. Neither is possible once beer is in a bottle.
Bottle conditioning manufactures its gas from a measured sugar dose and takes three weeks, but needs no CO2 cylinder and produces the fine, dense carbonation many drinkers prefer. It is also the only method available for beers you want to age for years. The priming sugar calculator handles that side, and it corrects for exactly the same residual-CO2 effect that this page describes, just from the opposite direction: there, dissolved gas is a credit against the sugar you need.
Spunding captures the tail end of fermentation under a pressure-relief valve set to the figure this calculator gives, so the beer carbonates itself with its own CO2. It uses no cylinder gas at all and gives fully natural carbonation, but it needs a pressure-rated fermenter and gravity readings accurate enough to know when to seal — a beer sealed with ten points left will overshoot badly, which you can head off by tracking gravity with the ABV and attenuation calculator. Set the spunding valve to the pressure for your target at fermentation temperature, not at serving temperature; the difference between 68 °F and 38 °F is large.