Vapour Temperature at Altitude: Measured Against a Commercial Still

Nearly every vapour temperature chart in the hobby assumes sea level. Almost none of them say so. That is a quiet problem for anyone distilling on high ground, because lower atmospheric pressure lowers every boiling point, and a reading that means one thing in Rotterdam means something quite different in Denver.

This page tests that claim against real numbers rather than asserting it. A commercial distillery at roughly 5,800 feet supplied two logged spirit runs, sample by sample, with hydrometer readings, applied corrections and column temperatures. The data was run through the calculators on this site. The runs were logged by Mitchell T. Abate, Head Distiller at Downslope Distilling in Centennial, Colorado, and are published here with permission. The underlying log is not reproduced, only what came out of the comparison.

The Data

Two runs were available. Between them they cover thirty eight individual samples taken through the length of a run, from the first spirit off the column down to the point where collection was stopped. Each sample records the hydrometer reading in proof, the correction the distiller applied, the corrected proof, and the column temperature at the moment of sampling. Strength falls from the high 180s of proof to the high 60s across a run. Column temperatures rise from around 162 F to 196 F.

Two questions were asked of it. Does the temperature correction used on this site agree with what a working distillery does in practice, and does the vapour temperature model reproduce a real column when local pressure is accounted for.

The Correction Cross Check

The sample temperature was never written down. Only the correction was. That turns out to be more useful than it sounds, because the correction can be inverted to recover the temperature it implies.

On one of the two runs a flat correction of 3 proof was applied to every single sample, from 156 proof at the start down to 67 at the end. A constant correction across a range that wide looks like a shortcut. Inverted through the correction slopes used by our alcoholmeter temperature correction calculator, it implies a sample temperature of 68.1 F on the first sample and 67.6 F on the last.

That is the cleanest confirmation of those slopes we have had, because it comes from outside. Had the slope curve been materially wrong, a constant correction would have back solved to a temperature wandering by ten degrees or more across the run. It moves by half of one.

The second run alternates between 3 and 4 proof rather than holding constant. Inverted, it gives a mean of 69.1 F with a spread of under 3 F, and the alternation is explained by rounding to whole proof points. One proof point is worth roughly 2.6 F at these slopes, so a true temperature sitting between two integers will produce exactly that pattern. The first sample of that run backs out to 76 F, warmer than the rest, which is what the first spirit off a hot still should be.

Finding the Altitude

The second test was harder. The column temperatures were fed into the Van Laar vapour liquid equilibrium model behind our vapour temperature calculator, and the model was asked, in effect, what elevation would make these numbers consistent.

It was given no location. Only temperatures and strengths.

The two runs independently best fit at about 4,600 and 6,100 feet. The distillery sits at roughly 5,800. The model bracketed the real elevation from run data alone, which is a stronger result than simply matching numbers it was handed.

What a Sea Level Table Costs

The practical finding is in the first row of that table. A distiller using a sea level chart on these runs would have been wrong by more than 20 proof on one run and by about 35 on the other. Those are not rounding errors. They are the difference between believing you are collecting hearts and actually collecting tails.

The gap widens as the run progresses. Early on, while the column is still producing strongly, a sea level table is merely optimistic. By the tail end it has collapsed entirely, reporting near zero where the still is in fact still delivering usable spirit in the thirties of ABV. A distiller following it would cut far too early and leave product in the boiler.

This page previously stated that a sea level table at altitude introduces an error of 5 to 15 percent ABV. Against measured commercial data that estimate was low. The figures here are 11 and 17.

Where the Model Stops

Honesty about the ceiling matters more than the headline. On the strong end of one run, five samples were taken with the column down around 162 F. There the model flattens out and simply reports the azeotrope, while the distillery measured strengths from 164 to 186 proof.

This is not a defect in the physics. It is the ceiling of a single stage equilibrium assumption. A real column with plates and reflux produces vapour that is not in equilibrium with the liquid in the boiler, and early foreshots are further depressed by low boiling congeners that a two component ethanol and water model does not represent.

One further piece of structure is worth reporting. The two runs drift in opposite directions through the tails. One reads consistently higher than the model predicts, the other consistently lower. Opposite signs rule out an error in the model and point instead at thermometer placement and reflux behaviour differing between the two setups. It is a reminder that a vapour reading describes the point in the column where the probe sits, not the still as a whole.

What To Do About It

Three things follow from this for anyone distilling above sea level.

Enter your actual elevation. Not your nearest city, your site. A few hundred metres is worth several points of ABV. Our vapour temperature calculator takes elevation directly, and any chart that does not ask for it is a sea level chart whether it says so or not.

Trust the hydrometer at the strong end. Vapour temperature earns its place through the middle and late portions of a run, where it tracked measured values within roughly 2 to 3 points of ABV here. At the top of the run it tells you the column is running strong and not much else.

Hold your sample temperature steady and a flat correction is defensible. The distillery's constant 3 proof was not laziness. It was a reasonable shortcut, and it survives inversion because the correction slope is nearly flat across the strengths a spirit run covers. Change the sample temperature between readings and the shortcut breaks immediately.

Our thanks to Mitchell T. Abate and Downslope Distilling for the data. Commercial run logs are rarely shared, and a model that has only ever been checked against published tables is a model that has never really been tested.