The coil is a dehumidifier you did not ask for
Any cooling coil cold enough to take heat out of a room is usually also below the dew point of the air passing over it, so water condenses on it and runs away down the drain. The share of the total capacity that goes into doing that rather than into lowering the temperature is the sensible heat ratio, or SHR. An SHR of 0.75 means a quarter of the machine is a dehumidifier.
For a house in July that is exactly what you want. For a wine room it is the problem. The arithmetic is direct: latent capacity is total capacity times one minus SHR, and dividing that by about 1,060 BTU per pound gives pounds of water an hour. Multiply by the fraction of the time the unit is actually running, and you have the daily condensate.
| Coil SHR | Latent share of a 2,500 BTU/hr unit | Condensate at a 36% run fraction |
|---|---|---|
| 0.65 — general purpose AC | 875 BTU/hr | 7.1 lb a day, 0.86 gal |
| 0.75 | 625 BTU/hr | 5.1 lb a day, 0.61 gal |
| 0.85 — typical wine unit | 375 BTU/hr | 3.1 lb a day, 0.37 gal |
| 0.95 | 125 BTU/hr | 1.0 lb a day, 0.12 gal |
| 1.00 — no condensation | 0 | none |
That table is the entire argument for a purpose-built cellar unit over a window air conditioner in a hole in the wall. The window unit will hold the temperature perfectly well. It will also take three quarters of a gallon of water a day out of a room that has nowhere to get it back from.
The air holds almost nothing, so the coil wins
A 640 cubic foot cellar contains about 48 pounds of air. At 55 degrees and 60% humidity that air is carrying 38.4 grains per pound, which is 0.26 pounds of water — 119 grams, about four fluid ounces, roughly a sixth of what is in one bottle of wine. A coil at an SHR of 0.85 on a 2,500 BTU/hr unit condenses that much in two hours of run time, and one at 0.65 does it in under an hour.
The only reason a cellar does not go to zero humidity in the first hour is that everything else in the room is a moisture reservoir: the wood racking, the cardboard, the corks, an unsealed masonry wall, the slab. Those buffers are what actually hold the humidity, and they are why the drying takes weeks rather than minutes. It is also why the recovery after you fix it is slow.
Leakage can run either way
Whether air leaking into the cellar helps or hurts depends on the grains difference, not on the relative humidities. Air at 72 degrees and 45% carries 52.4 grains per pound. Air at 55 degrees and 60% carries 38.4. So in that case the house air is wetter and leakage adds about 0.9 lb a day, partly offsetting the coil. Come winter, when the house drops to 25% humidity, the house air is at 29 grains and the same leakage now takes 0.6 lb a day out of the cellar instead. The seasonal reversal is why a cellar that is fine in August is a desert in February, and the calculator will show it if you run both conditions.
What to do with the number
The daily make-up figure is a size, not a schedule. If it comes out around a quarter gallon a day, an open pan of water with some surface area may cover it, and a lot of small cellars do exactly that. If it is a gallon a day, that is a real humidifier with a real water supply and a real drain, and the question of what happens when it fails while you are away becomes worth thinking about. If the number is negative, the room is gaining moisture rather than losing it, and the thing to work out is where that moisture is going instead — the answer is usually a cold surface, which is a condensation problem.
Related
For the sizing question this page deliberately does not answer, the walk-in cooler load calculator builds the BTU per hour from an envelope, door traffic and product. For humidification driven by ventilation across a whole house rather than by a coil, the humidifier sizing calculator is the closer fit; for the opposite problem in a damp basement, the dehumidifier sizing calculator and the dehumidifier run time calculator. The same grains arithmetic applied to vent air in a crawl space is the crawl space vent air calculator. How long the room holds temperature is the wine cellar holdover calculator.
Questions people ask
Why is my wine cellar so dry?
Almost always the cooling coil. If the unit is a general purpose air conditioner or a coil running well below the room dew point, a large share of its capacity is condensing water and sending it down a drain. At an SHR of 0.7 a 2,500 BTU/hr unit at a one third run fraction removes around six pounds of water a day, and a sealed room has no way to replace it. Purpose-built cellar units run the evaporator warmer specifically to avoid this, at the cost of needing more airflow for the same cooling.
What is sensible heat ratio and where do I find mine?
It is the fraction of a cooling coil total capacity that lowers temperature rather than condensing water, so 0.85 means 85% temperature and 15% water. It is on the equipment performance data, quoted at specific entering air and ambient conditions, and it moves with those conditions — the same coil has a different SHR in a 55 degree room than in a 75 degree one. If you cannot find it, run the calculator at 0.75 and 0.9 and treat the two answers as the range.
Can I just put a humidifier in the cellar?
You can, and many people do, but size it from the number on this page rather than from a guess, and think about the failure modes first. A humidifier adding a gallon a day to a sealed cold room is adding it to every cold surface too, so the dew point figures matter: anything below the room dew point will wet, including the evaporator, glass in the door, and any spot where insulation is thin. A room where the humidifier and the cooling unit fight each other continuously usually has a coil SHR problem underneath, and no amount of added water fixes that cheaply.
Does a passive or non-cooled cellar have this problem?
No, and that is the point of them. With no coil there is no condensate term at all, and the only moisture flows are leakage and whatever the walls and slab give up. Set the capacity to zero and the calculator reduces to the leakage arithmetic. Passive cellars usually have the opposite problem, which is too much humidity, plus a temperature they cannot control.
Why do the numbers barely move when I change the room humidity target?
Because the coil condensate does not depend on the room humidity in this model at all — it comes from the capacity, the SHR and the run fraction. The target only changes the grains difference driving the leakage term, which is usually the smaller of the two. That is a real effect and not a modelling shortcut: a coil below the dew point condenses at a rate set by the coil, not by how close the room is to a setpoint.