Can a door leak reach the bottles before it reaches the thermostat?

Yes. Cold air pools at floor level, so a gap near the sill drains the coldest air away from the lowest shelves first, and bottles racked low near the door can sit in a warmer pocket than a wall-mounted sensor several feet away registers, especially where no return draws air across that corner.

Does a full cellar hide a leak longer than a lightly filled one?

For a while, yes. The bottles are the room's thermal mass, and glass and liquid hold far more stored cold than the same volume of air, so a heavily loaded cellar absorbs each afternoon's exchange and shows a shallower, later sag than the same room half empty. The runtime climbs first in either case.

Does door type change how quickly a leak shows up?

Yes. A frameless door has no perimeter frame at all. It is 13.5mm tempered glass carried on pivot hinges and sealed by magnetic gaskets at the jambs and head, so seal pressure shifts across the whole opening when the panel or a hinge moves. A framed or wooden door concentrates its sealing at fewer points, which makes any single failed point count for more.

Can a door that latches normally still leak at the sill?

Yes. The latch governs how the door meets the jamb, not how its bottom edge meets the floor, and those are separate seal points that fail independently. On a door fitted with an automatic door bottom, the gasket drops from inside the bottom rail as the door closes, so a mechanism no longer dropping fully leaves a latched, flush-looking door with an open sill beneath it.

Does opening the door during the day distort the record?

It adds a spike, but a recognizable one. An entry event shows as a sharp step in temperature that decays over minutes as the room's air remixes, while a leak shows as a slow, repeating curve that returns at the same hour each day. Noting entry times beside the readings keeps the two apart.

How long a record does it take before the pattern is readable?

Roughly a week of readings taken at the same two or three hours every day, one of them in the late afternoon and one shortly after sunrise. A single day cannot separate a sag caused by one unusually hot afternoon from a sag that repeats, and the repetition is the entire signal. Leave the setpoint and the fan speed untouched for the duration and note both alongside the readings, since changing either mid-record erases the shape the diagnosis reads.

Wine Cellar Door Leaking Temperature: Read the Runtime Shape

Choosing Between Them for a Specific Run

Once a schedule has been issued to anyone, changing a value quietly is worse than not changing it. A revised schedule carries a new revision number, a new date, and a marker against every row that moved, so a reader holding both versions can tell in seconds which lines are different. Reissuing a corrected sheet with no revision marker leaves two documents in circulation that look interchangeable and are not.

The reissue also has to reach the people working from the old one. A cellar's fixture order, the racking shop drawings, and the electrical rough-in can all be running from the same schedule at the same time, and a Kelvin value corrected on the designer's copy changes nothing until the version carrying that correction lands with whoever placed the order.

Fixture substitutions happen at bid, at order, and occasionally after delivery, and a fully populated schedule is what sets the terms of one in advance. A substitution proposed against a complete row has to hold every value on that row, including the ones a supplier would find inconvenient to match. A fixture matching wattage, beam angle, and trim size while dropping below a stated CRI floor is a different specification, and a populated row makes that visible without an argument about intent.

The burden of showing it sits with whoever proposes the swap. A substitution request arriving as a catalog number and a price, with no line-by-line comparison against the scheduled values, has not been made yet; it has been asked about. Requiring that comparison in writing is what keeps the substitution a decision the specifying professional is still making.

Threshold seals are the one envelope point that behaves like the door because it is part of the door, and Cachet's guidance on it is blunt: threshold seals are where amateur installations most often fail, and a gap of one-eighth of an inch there undermines the seal for the entire door. The test that sorts any of these is the one the runtime already answers: whether the added load arrives on the room's daily schedule or independently of it.

The Sheet Has to Say What It Is

TIP: A schedule reissued after fixtures are ordered should say so on the sheet. A revision that arrives post-order is a change order in practice, and marking it as one keeps the cost conversation attached to the line that caused it.

What you observe	Points to
Present since installation, color consistent, dim points evenly spaced	Spacing and mounting height
Boundary lines up with a material seam in the racking, present from day one	Rack depth or finish change
Color shifts across a boundary, not just brightness	Mixed LED bins from a partial replacement
Whole run dims, flickers, or fades together; appeared after a period of looking fine	Driver channel losing output

The first thing most owners notice is not the temperature reading. It is the compressor. The setpoint is untouched and the room still reads close to target for most of the day, but the unit is running in longer stretches, or restarting sooner after it shuts down, than it did a month ago. What changed is how hard the equipment works to hold the room where it already was.

That extra runtime is the equipment working against a load it did not have before. A sealed cellar under unchanged ambient conditions and unchanged use presents a repeatable heat load, so the duty cycle should look much the same week to week aside from seasonal drift. A sustained increase with no change in setpoint, room use, or outside temperature means something is now exchanging air with the room that was not before. A door is the most common place for that exchange to start, since it is the only moving, gasketed opening in an otherwise sealed envelope of R-19 walls and R-30-plus ceiling insulation.

A door leak, an undersized unit, and a compromised wall or ceiling seal all produce longer runtime, and only one of them is a door problem. Total runtime establishes that something has changed. It does not establish what.

Temperature alone will not separate a door leak from other causes, because several unrelated problems push a cellar warm. Humidity is not the separator either, though it is often presented as one.

Cachet's target range is 55°F, plus or minus two degrees, at 60 to 70 percent relative humidity. What governs an air exchange is not the percentage on each display but how much water each air mass actually carries. House air at roughly 72°F and 50 percent relative humidity holds more moisture per pound of dry air than cellar air at 55°F and 65 percent does, so an exchange at the door can push cellar humidity up. On a hot, dry afternoon that same house air can hold less than the cellar air, and the identical exchange pushes humidity down instead. The direction is not stable from day to day, which is why it cannot be read as a diagnostic sign.

What exchange does reliably is make the humidity trace less settled. Readings that wander in both directions while runtime climbs point toward air trading across an opening. Humidity that moves steadily in one direction while runtime holds is more often a coil or condensate problem: the evaporator coil runs deliberately below the room's dew point, moisture condenses out of the air crossing it, and the condensate leaves through a drain, so a fault in that path has its own settled direction.

The signal that does separate causes is when the runtime happens. Air exchange across a gap scales with the difference between the two air masses on either side of it, and a gap trades air both ways at once, divided by density across the height of the opening. Ambient dry-bulb temperature in coastal Orange County swings widely between early morning and late afternoon, while the cellar side of the door barely moves if the equipment is healthy. That widening difference drives more air through a marginal seal in the late afternoon than at any other hour.

On a log, the result is a room that holds close to setpoint overnight and through the morning, sags in the late afternoon, then recovers once the house cools and the two air masses come back toward each other. A unit that is undersized or losing capacity does not usually produce that shape. Its limitation is total capacity rather than the size of a difference across one opening, so it runs long across the whole day, including the overnight hours when ambient load is lowest. A late-afternoon sag that reliably clears by morning points at an opening rather than at the equipment.

Before the door gets blamed, it is worth confirming the unit is not undersized or degraded, since both produce longer runtime.

Cachet installs from a defined set of systems sized to the room: WhisperKOOL self-contained and through-wall units and Breezaire self-contained units with Sentry III+ controls, both from 500 to 2,000 BTUH, CellarPro and US Cellar Systems ducted equipment from 1,800 to 3,500 BTUH, and Wine Guardian's ducted Pro Series split system on a 24V thermostat circuit compatible with Nest or Honeywell controls. A unit sitting near the bottom of its rated capacity range for the room it serves will run long even with a flawless envelope, and that runtime carries no daily shape at all.

A door leak is also easy to confuse with a seal failure elsewhere in the room: the wall-to-ceiling joint, a penetration around ducting or wiring, or a vapor barrier gap behind the racking. Those locations trade air too, but the daily pattern is usually flatter, because most are not exposed to the swing a door on an exterior or garage-adjacent wall sees. A gap in the ceiling assembly exchanges against attic conditions, which run on their own schedule rather than tracking the room's hour for hour.

If the compressor is running long, the sag arrives in the late afternoon and clears overnight, the humidity trace has become unsettled rather than moving one way, the unit checks out for rated capacity and age, and no obvious gap sits elsewhere in the envelope, what remains is the door: its gasket, its threshold, or the latch pressure holding it closed against the frame.

A short, dated record makes the next conversation quick: the hours the compressor runs longest, the room's temperature and humidity at those hours set against a period when it held setpoint normally, and which door type is installed. Cachet builds four door types, wooden, metal with a thermal break, framed glass, and frameless glass, plus French pairs and sliding or barn-style configurations, and each seals differently at the threshold and jamb, which changes what an inspection checks first. Condensation on the glass is a separate question with its own answer, and it does not settle whether the door is leaking.

Cooling equipment is installed and commissioned at stage 9 of Cachet's build, doors and finishing go in at stage 10, and inspection and walkthrough is stage 11. A door issue on a finished cellar is diagnosed against a system commissioned before the door was hung, and a clear symptom log turns a service visit into a confirmation rather than a search.

Can a door leak reach the bottles before it reaches the thermostat?
Yes. Cold air pools at floor level, so a gap near the sill drains the coldest air away from the lowest shelves first, and bottles racked low near the door can sit in a warmer pocket than a wall-mounted sensor several feet away registers, especially where no return draws air across that corner.
Does a full cellar hide a leak longer than a lightly filled one?
For a while, yes. The bottles are the room's thermal mass, and glass and liquid hold far more stored cold than the same volume of air, so a heavily loaded cellar absorbs each afternoon's exchange and shows a shallower, later sag than the same room half empty. The runtime climbs first in either case.
Does door type change how quickly a leak shows up?
Yes. A frameless door has no perimeter frame at all. It is 13.5mm tempered glass carried on pivot hinges and sealed by magnetic gaskets at the jambs and head, so seal pressure shifts across the whole opening when the panel or a hinge moves. A framed or wooden door concentrates its sealing at fewer points, which makes any single failed point count for more.
Can a door that latches normally still leak at the sill?
Yes. The latch governs how the door meets the jamb, not how its bottom edge meets the floor, and those are separate seal points that fail independently. On a door fitted with an automatic door bottom, the gasket drops from inside the bottom rail as the door closes, so a mechanism no longer dropping fully leaves a latched, flush-looking door with an open sill beneath it.
Does opening the door during the day distort the record?
It adds a spike, but a recognizable one. An entry event shows as a sharp step in temperature that decays over minutes as the room's air remixes, while a leak shows as a slow, repeating curve that returns at the same hour each day. Noting entry times beside the readings keeps the two apart.
How long a record does it take before the pattern is readable?
Roughly a week of readings taken at the same two or three hours every day, one of them in the late afternoon and one shortly after sunrise. A single day cannot separate a sag caused by one unusually hot afternoon from a sag that repeats, and the repetition is the entire signal. Leave the setpoint and the fan speed untouched for the duration and note both alongside the readings, since changing either mid-record erases the shape the diagnosis reads.