A sealed cabinet that swings 25 K builds roughly 86 mbar against its own gasket. Whether one vent carries that or five depends on the equalisation time you are willing to accept, and that is the number almost nobody publishes. The arithmetic is below, written out, so you can check it against your own cabinet instead of taking a figure on trust.
A cabinet is a fixed volume of air. Heat it and the pressure rises, because the air has nowhere to go. For a rigid box the arithmetic is the ideal gas law held at constant volume: absolute pressure tracks absolute temperature, so the rise equals 101.325 kPa multiplied by the temperature swing and divided by the cold absolute temperature.
Twenty degrees at dawn, forty-five by mid afternoon. That is 101.325 × 25 ÷ 293.15, which comes to 8.6 kPa, or 86 mbar. Put the same cabinet in the sun in July and the internal swing reaches 40 K, which is 138 mbar.
| Internal swing | Ideal pressure rise | In mbar | What it loads |
|---|---|---|---|
| 15 °C → 35 °C | 7.0 kPa | 70 | Door gasket, gland body |
| 20 °C → 45 °C | 8.6 kPa | 86 | Gasket plus latch preload |
| 25 °C → 55 °C | 10.2 kPa | 102 | Panel bow, weld at the corner |
| 20 °C → 60 °C | 13.8 kPa | 138 | Everything above, at once |
Rigid, perfectly sealed box. A real cabinet gives some of it back through panel flex and minor leakage, so a field measurement comes in lower, often by a third.
That reduction is not a reason to skip the vent. The flex is the wear. Every millimetre the panel moves is a gasket being compressed and released again, and a gland being pushed sideways on its thread. Do that twice a day for five years and the failure is not a dramatic one: it is a gasket that no longer returns to shape, and a cabinet that passes its factory IP test and leaks at year three.
Two transients are worth checking on top of the daily cycle. Transport, if the cabinet is sealed at low altitude and then hauled over a pass. And a cold rain hitting a panel that has been in the sun, which contracts the air inside far faster than sunset does. Both are several times the daily figure for a few minutes, which matters when you choose the time below.
The flow you need is the air that has to move, divided by the time you are prepared to give it. Written as a formula, with litres, kilopascals and hours:
Q [L/h] = V [L] × ΔP [kPa] ÷ ( t [h] × 101.325 )
Everything in that formula is a property of your cabinet except one thing, and it is the only thing that moves the answer by an order of magnitude: t. Nobody quotes it. A supplier who says "one vent per two hundred litres" has picked a t for you and not told you which.
| Free air volume | Equalise in 1 h | 4 h | 8 h | 24 h |
|---|---|---|---|---|
| 5,000 L floor cabinet, ~5 m³ | 426 L/h · 5 vents | 107 L/h · 2 vents | 53 L/h · 1 vent | 18 L/h · 1 vent |
| 20,000 L container, ~20 m³ | 1,705 L/h · 19 vents | 426 L/h · 5 vents | 213 L/h · 3 vents | 71 L/h · 1 vent |
Calculated at 8.6 kPa (a 25 K swing) and 90 L/h per M12 vent plug. Vent counts rounded up. Change either assumption and the count changes; that is the point of the table rather than a single recommended figure.
Which time is honest? The daily swing is not a step change. A cabinet on a concrete pad in direct sun takes three to five hours to climb from dawn to peak, and about the same to fall back overnight. Matching a four hour ramp is the defensible target, and on the table above that is two M12 plugs in a five cubic metre cabinet, not nineteen.
One hour is a design margin for a fast transient, not a weather figure. Twenty four hours is sizing for a slow leak. Both are legitimate choices for specific projects. Neither is a default, and a datasheet that gives you a count without saying which one it assumed has not answered your question.
The count is not a property of the cabinet alone. Divide your required flow by what the part in your hand actually passes, and that figure moves by a factor of three between two plugs with identical threads. Both suppliers are quoting honestly. They are measuring different membranes, and sometimes at different pressures.
| Thread | Published airflow at 70 mbar | Converted |
|---|---|---|
| M5 × 0.8 | ≥ 600 ml/min | 36 L/h |
| M8 × 1.25 | ≥ 600 ml/min | 36 L/h |
| M12 × 1.5 | ≥ 1,500 ml/min | 90 L/h |
| M16 × 1.5 | ≥ 4,000 ml/min | 240 L/h |
| M20 × 1.5 | ≥ 4,500 ml/min | 270 L/h |
| M24 × 1.5 | ≥ 7,000 ml/min | 420 L/h |
| M40 × 1.5 | ≥ 8,500 ml/min | 510 L/h |
One supplier's nylon vent plug datasheet, retrieved October 2026, quoted as minimums (Hoonsun, HX-PF series). Not a cross-industry average. See the seven supplier comparison for how far apart published M12 figures actually sit.
A concrete case. Gore publishes the PolyVent Standard at 450 ml/min at 70 mbar, which is 27 L/h. The nylon M12 in the table above is 90 L/h at the same 70 mbar. Same thread, same pressure, 3.3 times the air. Size your cabinet for the first and buy the second and you have installed three times the vents you needed. Do it the other way round and the cabinet is under-vented by seventy per cent, which will not show up until the gasket starts weeping in its second summer.
So: ask for the airflow at your working differential, not at whatever pressure flatters the number. Then divide. If a supplier cannot tell you the pressure their figure was taken at, the figure is not usable.
Daily breathing is the one above: tens of mbar, driven by the sun, repeating every day for the life of the site. A membrane vent handles it, and the membrane is what keeps the water out while it does.
The second event is a cell venting or a thermal runaway. That is tens of kilopascals arriving in seconds, with hydrogen and other gases behind it. It is the job of a deflagration vent or a bursting disc, sized to the enclosure volume under NFPA 68, NFPA 855 or EN 14491. A breather is not that device and cannot be made into it. Its membrane area is chosen in litres per hour, not cubic metres per second, and it has no burst pressure and no certified vent area to offer.
The distinction is worth stating plainly because combining the two jobs in one part is a common sales claim, and the two sizing methods do not overlap. If a single part is offered for both, ask for its burst pressure and its effective vent area. A membrane assembly has neither number, and the answer will tell you what you are actually being sold. We wrote the two sides out separately in breathing versus runaway if you need the argument in more detail.
The quiet failure sits between them. A vent that has clogged still passes its IP68 retest, because that test measures water ingress and not airflow. Dust, salt and pollen close the pores, the label stays valid, and the cabinet keeps breathing through its gasket instead. That is why the count is only half the design: the mounting position, sheltered and facing down, decides how long the number you calculated stays true. The clogging case covers the test that would catch it.
Five inputs and one drawing: the free air volume inside the gasket in litres, the internal temperature swing you are designing for, the equalisation time you will accept, the IP rating the cabinet has to hold, and the thread or wall thickness you can tap. Send those and you get a count and a part number back, or an honest note that the count is one and the money is better spent on a door gasket.
It is a safe rule, and it is published for battery packs rather than cabinets (Fellten's PRV datasheet gives one valve per 200 litres), where the same part also has to cope with cell venting. Run it through the arithmetic above and it works out to about 11 to 38 minutes to equalise a 25 K swing, depending on which M12 you fit. Weather never asks for that. On a cabinet that only sees sunlight, the rule usually lands at five to ten times the parts you need. On a pack, keep the conservatism.
Yes, and it is often the better build. Flow adds, two ports give you redundancy, and you can place each where it does not sit under standing water. Check the sum of the airflow figures rather than comparing thread sizes, because membrane area per unit of thread differs between suppliers.
No. The rating belongs to the membrane and the seal of that specific assembly, not to its diameter. A larger plug passes more air at the same differential and does not pass a harder water test. Size for airflow, then confirm the rating against the test report for the part you are actually buying.
Then the count is fixed and the time is the variable. Put your existing count into the table to see the equalisation time you actually have, and decide whether the gasket can live with it. That is also the honest answer when the calculation says one vent: knowing it is one is worth more than buying five.
If you would rather start from the other end, the pressure calculator takes your volume and temperature swing and returns the differential and the flow, and the vent plug range lists what the parts themselves are rated to.
Three numbers and we will return the count, the thread and the part, with the arithmetic shown so you can check it against your own cabinet.
