Build the sensing end as its own sealed unit and keep every other part out of the beer: a PET bottle preform for homebrew-scale batches, or a 316L stainless tube welded shut at one end when the vessel gets caustic cleaning. Your 21 x 16 mm tilt board enters a tube on its diagonal, so it needs a 26.4 mm bore — a 1-1/4 in [31.75 mm] OD tube with a 1 mm wall, or the body of a preform. Half a metre of beer presses on the shell at only about 5 kPa [0.05 bar], so pressure is not what kills these builds. The wire penetration is; delete it with a wireless design, or pot it properly, and the rest is routine sealing.
What does the enclosure have to survive inside the beer?
Three things: a mild acid — beer sits between pH 4.0 and 4.5 at 18-24 C during fermentation — a sugar-rich liquid that feeds microbes in every scratch, and the cleaning between batches, which reaches 1-2 % caustic soda at 60-80 C. That last one picks your material.
Beer itself is a gentle electrolyte: pH 4.0 to 4.5, 4-8 % ethanol, and 18-24 C for ales or 8-13 C for lagers. The chloride in your brewing liquor is the metal-killer, commonly 50-250 mg/L depending on the water. Left as a damp film in a scratch or a gasket crevice, that is enough to pit 304 stainless at room temperature. It is the same failure that eats 304 in coastal air, where the chloride mechanism is laid out in detail.
The molybdenum in 316L lifts pitting and crevice resistance well above 304. The chloride stress-corrosion threshold stays near 60 C [140 F] for both grades, and caustic clean-in-place (CIP) runs 60-80 C [140-176 F]. That combination decides the wetted material: PET or 316L, with 304 reserved for the dry side of the room. The grades guide carries the full 304 versus 316L comparison.
Food contact narrows the polymers as well. PET is what soda and water bottles are blown from, and it contains no BPA. Silicone O-rings and potting epoxies both come in food-contact grades; the US FDA maintains food-contact listings for both, so buy the grade whose datasheet names the listing.
How deep is the dive, and which rating covers it?
Specify IP68 and write the conditions yourself: half a metre of beer adds only 4.9 kPa [0.05 bar] — 1,000 kg/m3 x 9.81 m/s2 x 0.5 m — but a four-week soak is beyond IP67, whose test is 1 m for 30 minutes.
IEC 60529 draws the line at the second digit: 7 means temporary immersion, tested at 1 m [3.3 ft] for 30 min, and 8 means continuous immersion on conditions agreed between manufacturer and user. On a one-off build you are both parties, so write the conditions into your own spec: 0.5 m of beer, three weeks, 25 C [77 F], sanitation at 60 C [140 F].
IP66 is a hose test — the second digit 6 uses a 12.5 mm nozzle delivering 100 L/min — and it is not an immersion rating. On the type side, NEMA 250 puts temporary submersion at Type 6, nearest IP67, and prolonged submersion at Type 6P, nearest IP68; Type 6P pairs with IP68, never IP67.
On the labels: type and IP numbers on catalog enclosures are manufacturer ratings — ours read rated NEMA 4X / IP66 by the manufacturer; not UL listed. The IP rating chart maps the codes side by side.
Where do submersible enclosures actually leak?
At the wire: the gland, the potted hole or the cap thread is where every failed unit took water. It fails by breathing — washed at 80 C and cooled to 20 C, the trapped air contracts to 0.83 atm and pulls about 17 kPa against the seal.
Run the arithmetic on that breath. Wash the sealed unit at 80 C [176 F] and the trapped air pushes out past a lazy cap thread; drop it into 20 C beer and what remains contracts. Inside pressure falls to 101.3 x 293/353 = 84.1 kPa, about 17 kPa below the beer around it. That vacuum works on the seal for the whole batch, and it wins whenever the exit is a clearance hole.
So design the exit first, and pick one of three routes:
- Go wireless and there is nothing to leak: battery inside, 2.4 GHz through the PET wall at under 1 dB of loss, charging between batches. A plastic fermenter passes the signal, so a gateway a couple of metres away hears the float directly.
- Pot the exit when you need wires: food-contact epoxy around the cable, at least 20 mm of potted length. Specifying an overmolded cable exit covers the molded version of the same joint.
- Treat glands as dry-side hardware. Sealing penetrations in a submersible enclosure shows why a glanded joint belongs on the wall box, not in the beer.
Dry the inside before you close it: 2 g of silica, cap tightened while the body is warm, and the small trapped air volume never reaches its dew point. The condensation guide runs those dew-point numbers if your fermenter lives somewhere cold.
Which body wins: PET preform, 316L tube, or a potted dip?
For beer only, a PET bottle preform with a sealed cap wins. The moment the vessel sees caustic CIP or a brewery washdown, go to a 316L tube welded shut at one end; pot the board in food-contact epoxy when the sensor is small and disposable.
| Build | Body | Wall seal | 2.4 GHz radio | Best for | Limit |
|---|---|---|---|---|---|
| PET preform | Beverage-grade PET, ~25 mm body, 100-120 mm long | Screw cap plus silicone O-ring | Passes, under 1 dB loss | Beer-only homebrew vessels | Hot wash under 60 C |
| 316L welded tube | 31.75 mm OD x 1 mm wall, one end capped and welded | Welded; cable exit potted | Blocked by the metal | Caustic CIP at 60-80 C | Needs a cable or an antenna port |
| Potted dip | Food-contact epoxy over the bare board | The potting is the seal | Passes, small loss | Small, disposable sensor boards | One-way build, no rework |
The preform is bottle stock: a 28 mm neck, a 25 mm body, and the same food-contact pedigree as the bottle it becomes. It also lets your radio live inside the beer, which deletes the one part that leaks.
The 316L tube is how brewery probe builds are made: a tube closed with a full-penetration weld at one end, the other end carrying the cap or the potted cable. It shrugs off caustic and peracetic acid for years, and it blocks 2.4 GHz completely, so budget for the cable.
For the dip, choose an epoxy listed for repeated food contact; the cheap polyester casting resins shrink as they cure and carry no such listing. A flat gasketed food container seals a lid, not a feed-through, and it tumbles in a rolling ferment — save it for the dry side.
What size tube does a 21 x 16 mm board need?
A bore of at least 26.4 mm, because the 21 x 16 mm board enters on its diagonal: the square root of 441 + 256 is 26.4 mm. A 1-1/4 in [31.75 mm] tube with a 1 mm wall has a 29.75 mm bore and fits.
The PET preform is easier. With roughly 3 mm of wall around a 25 mm body, its bore is 18-19 mm, and your board rides down the axis on its 16 mm side. A 65 x 18 mm lithium cell fits the same bore in a 100-120 mm body, with the board above it and the temperature sensor on the same board.
If you want the slim 1 in [25.4 mm] OD tube, re-lay the board. With 0.5 mm of wall the bore is 24.4 mm, which takes a diagonal up to 24.4 mm — a 20 x 13 mm layout, for instance. Add about 10 mm of body at each end for potting and ballast, so a complete unit runs 120-150 mm long.
How do you make the float sit at the right angle?
By ballast: the sealed tube must weigh a few grams less than the beer it displaces, and you trim with weights at the bottom until it floats about 25-35 degrees from vertical.
Take the preform as a cylinder 25 mm OD and 120 mm long, and run the numbers:
- Displaced volume: 3.14 x 12.5 x 12.5 x 120 = 58,900 mm3, call it 58.9 cm3.
- At a starting gravity of 1.040 kg/L it displaces 58.9 x 1.040 = 61.3 g of beer; at a final gravity of 1.010, 59.5 g.
- Dry parts: 45 g lithium cell, 10 g of board and wiring, 2 g of desiccant — 57 g. Screw 2 g of stainless ballast to the closed end: 59 g total.
At 59 g the float still swims at final gravity with 0.5 g to spare, and it rides deeper as the beer thins from 1.040 toward 1.010 — that shift in attitude is exactly the signal your accelerometer reads. Trim the ballast until the unit hangs 25-35 degrees from vertical at mid-fermentation. Then calibrate it in sugar solutions of known density.
The rule the example hands you: dry mass a few grams under the displacement at final gravity, ballast low, sensitive axis along the tube.
Where does the rest of the electronics live?
Outside the beer, in its own 316L wall box by the vessel: it needs no food rating, it stays dry, and you can open it to swap batteries and charge the radio.
A compact radio gateway lands in an 8 x 6 x 4 in 316L wall-mount enclosure. The 10 x 8 x 4 in size adds room for a terminal row, and a controller with its power supply fits the 12 x 10 x 6 in box. All three carry a hinged gasketed door and a removable mounting panel; browse 316L wall-mount enclosures if the wash zone decides your size.
Bring cables in through the bottom face with a drip loop, and use glands rated for hose duty — the box holds its 4X rating only down to the rating of its weakest fitting. Sealing a gland around a smaller cable covers the sizing trick when your sensor cable runs thin.
For US installs, NEC 110.28 says an IP code never replaces an enclosure type rating, so buy this box by type — Type 4X covers hose washdown. If the project grows past monitoring, replacing a brewery's OEM controller with an Arduino-based system walks the wiring side of that upgrade. Every box is made to order and delivered in 10 business days, counted from your order — custom sizes from your approval of the drawing.
What goes on the order?
Two items: the submersible unit you build from a preform or have welded from 316L, and the outside box, which you order by grade, size and rating — 316L, 8 x 6 x 4 in or 12 x 10 x 6 in, rated NEMA 4X / IP66 by the manufacturer.
| Line item | Specify this |
|---|---|
| Submersible body | PET preform, 28 mm neck, 110 mm body — or 316L tube, 31.75 mm OD x 1 mm wall x 130 mm, one end capped and welded |
| Wall exit | Wireless, with nothing through the wall, or a potted cable exit in food-contact epoxy, 20 mm of potting |
| Float trim | 2 g stainless ballast, adjustable; 2 g desiccant sachet inside |
| Outside box | 316L wall-mount, 8 x 6 x 4 in, 10 x 8 x 4 in or 12 x 10 x 6 in, hinged gasketed door, removable panel |
| Box entries | Bottom-face cable glands rated IP66 / 4X, with a drip loop |
| Delivery | 10 business days from your order; custom sizes from your drawing approval |
Custom starts at one piece: you set the material, the width, height and depth, and every cutout on the custom enclosures page, and we build it from one piece. The stainless size finder filters every wall-mount and HMI box by grade and outside size. We build and cut the outside boxes; the submersible body stays yours — a preform from the bottle trade, or a welded tube from your fabricator.
The decision, by case: beer-only homebrew vessel, go preform and wireless. Caustic-cleaned brewery vessel, go welded 316L with a potted cable. Small disposable sensor, dip it in listed epoxy. The controller always lives outside the beer, in 316L whenever the wash zone can reach it.
Enclosures in this article
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 8 × 6 × 4 in | 316L stainless | 8 x 6 x 4 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 10 × 8 × 4 in | 316L stainless | 10 x 8 x 4 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 12 × 10 × 6 in | 316L stainless | 12 x 10 x 6 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Is PET from a bottle preform food safe for weeks in beer?
Yes. PET is the polymer that soda and water bottles are blown from, it contains no BPA, and the preform is simply that bottle before the blow mould, so the food-contact pedigree is identical. Give it bottle treatment: sanitize warm, keep any hot wash under about 60 C, and fit a fresh O-ring once the old one stays flattened.
Can I pot the whole board in epoxy and skip the housing?
Yes, and a smooth potted block is the easiest shape in the bucket to sanitize. Use an epoxy listed for repeated food contact rather than a general-purpose casting resin, pour in two thin layers so the cure stays cool, and give it 24 h at room temperature before it meets the beer. The trade is final: a potted sensor cannot be opened again, so treat it as a consumable.
How long will a battery last in a sealed floating sensor?
Count the duty: a WiFi microcontroller transmitting 3 s every 10 min averages about 30 mAh a day, so a 3000 mAh cell runs roughly three months. The six-axis chip is the quiet drain, because leaving both of its sensors running draws about 3.6 mA and cuts the same cell to under a month, so run the accelerometer alone in low-power mode.
Will the stainless tube block the wireless signal?
Yes. A closed stainless tube is a shield, and 2.4 GHz stops at the wall, while PET passes it with under 1 dB of loss. If you need the metal, plan a potted cable or an antenna that rises above the liquid, and keep the receiving gateway within 2-3 m either way, because beer absorbs 2.4 GHz strongly.
How do I keep moisture out of a sealed float for a whole season?
Seal it dry. Assemble in a warm room, drop in a 2 g silica sachet, which holds about a third of its own weight in water, and tighten the cap while the inside air is warm, so the sealed volume starts out below its dew point. A 30 cm3 tube holds so little air that one sachet lasts a season.
What cleaning chemicals can a 316L tube take?
The standard brewery set: 1-2 % caustic at up to 80 C, peracetic acid at its usual 0.5 % use strength, and hot water rinses. 316L is the grade breweries and dairies run for exactly this duty. Keep chlorine bleach weak and cold if you use it, under about 200 mg/L at room temperature, because concentrated chlorides attack even 316L.
