Short answer: put the cameras in a 304 stainless housing with a 5 mm [3/16 in] borosilicate window, nitrile seals and a PUR-jacketed cable through an IP67-rated gland — that build lives in coolant for years. The same scanner in a molded plastic IP65 case survives about 3 days under a coolant waterfall. An IP code tells you which jets the box stops, never which chemistry it survives, so you choose every wetted material against the coolant actually on your floor. Containing coolant at the machine cabinet's own gaps is a separate problem, covered in machine-tool coolant containment at enclosure gaps.
Will an IP65 enclosure survive coolant inside a CNC mill?
No, not on the strength of the rating: IPX5 means the box stopped jets from a 6.3 mm nozzle at 12.5 L/min from about 3 m, and nothing in that test touches oil, amine or biocide chemistry. IEC 60529 tests every digit with clean water, and only water. The second digit 6 raises the stakes to a 12.5 mm nozzle at 100 L/min, which stands in for a wash-down hose, not the drip line of a spindle.
A NEMA 250 Type 4X enclosure adds corrosion exposure — 200 h of salt spray against a 304 stainless specimen — and that still says nothing about a chlorinated cutting oil or a biocide-loaded semi-synthetic. Treat every rating as the manufacturer's rating for the specified design: our enclosures are rated NEMA 4X / IP66 by the manufacturer, not UL listed, and the label never covers the chemistry. What decides survival is the materials list you are about to build, and how much spray you keep off the glass.
Which enclosure materials does CNC coolant attack first?
The soft parts go first: elastomer makers' charts show EPDM swelling past 50 % in straight oil, PVC jackets hardening and cracking, and polycarbonate windows crazing within days, while glass, nitrile, FKM and stainless shrug both fluid families off.
Get the coolant's product data sheet before you order, because the two fluid families do different damage. A semi-synthetic or soluble oil runs pH 8.5 to 9.5 with amines and biocides: bare aluminum etches in it and galvanized parts rot, while stainless ignores it. A straight or chlorinated oil kills EPDM, crazes molded plastic and attacks yellow metals. Most shops also carry tramp oil, so pick for the worst fluid present.
| Wetted part | Semi-synthetic, pH 8.5 to 9.5 | Straight or chlorinated oil |
|---|---|---|
| Polycarbonate or acrylic window | survives mist, hazes from wiping | crazes and cracks in days |
| Borosilicate glass | unaffected | unaffected |
| EPDM gasket | excellent in water-based fluid | swells past 50 %, crumbles |
| Nitrile (NBR) gasket | good to about 80 C | good, under 10 % swell |
| FKM gasket | good | best class, under 5 % swell to 200 C |
| Aluminum housing | etches | fine |
| 304 or 316L stainless | fine | fine |
| PVC cable jacket | acceptable | hardens, cracks |
| PUR cable jacket | fine | fine, the machine-tool standard |
Those swell and etch numbers come from published elastomer and metalworking-fluid compatibility charts, and they decide this selection more than any ingress rating does.
Glass or plastic: which window stays clear in coolant?
Borosilicate glass: about 92 % visible light transmission, untouched by either coolant family, and it wipes clean, where polycarbonate picks up stress microcracks that no wipe can remove.
Order 5 mm [3/16 in] borosilicate for any window up to about 8 in [200 mm] on a side — standard sight-glass practice — and go thicker above that so the pane stops flexing against its gasket. Retain the glass in a stainless frame on a nitrile gasket, or bed it in RTV silicone when it will never come out; the screw and support spacing that keeps a long gasket line sealed is laid out in gasketed window seal spacing and support.
Light matters to a scanner: uncoated glass loses about 4 % per surface, so the pane costs you around 8 %, and an anti-reflective coating takes that under 1 %. Spend on the coating when your code contrast is already marginal. The door seal around it follows sheet-metal gasket design for IP65 practice — one continuous compressed bead, sealed at the corners.
How do I keep the coolant waterfall off the camera window?
Move the window out of the stream: re-aim the box or add a shield so the nozzle trajectory misses the glass, then run an air knife at about 80 psi [5.5 bar], figuring 3.5 scfm of air per inch of knife, to sweep mist off the view line.
- Placement first. Relocate or tilt the enclosure so the trajectory misses the pane, or hang a 304 stainless shield beside it. Running cost zero, and most waterfall problems die here.
- Air knife second. A 6 in [150 mm] knife draws about 21 scfm at 80 psi while the spindle runs, and it holds a dry corridor across the glass for the exposure time of each read.
- Spinning-disc window third. At 3000 rpm a 100 mm disc sees about 500 g at its rim — 0.05 m x (314 rad/s)² is roughly 4900 m/s² — so the coolant film cannot hold on. It needs standoff room in front of the lens and a drive, and it adds one rotating seal of its own.
- If you accept the blast, plan the window and seals as wear items and keep spares on the shelf.
Mist-only exposure is the easy case: spray the housing with water and wipe it at shift end. Water is harmless to glass, nitrile and PUR, and it removes the film that slowly bakes onto any surface.
How does the camera cable get out without leaking?
Through a rated gland in a punched, deburred hole: an IP68 gland whose clamping range matches your cable's jacket OD, or an IP67 M12 connector when you want to swap the scanner in the field, keeps the whole assembly at its rating.
The assembly is rated by its weakest part: an IP68 gland and an IP67 connector keep you above an IP66 target, while a loose grommet in a field-drilled hole leaves the box effectively unrated. Match the gland to the jacket as cable gland clamping range shows, and torque it down as covered in installing a cable gland without losing the rating.
Punch the entries in a bolt-on plate rather than the body when you expect changes — the case for a gland plate on a stainless enclosure is exactly this one. On the scanner side, machine vision camera cable selection covers jackets, bends and connectors for this environment. A glass-tape-and-RTV wrap over the cable also seals — harsh-environment camera builds ran plate steel, RTV and corner-screwed glass for years — but you cut it out to service it, so save it for cables you will never touch. Lock the machine out and prove absence of voltage before you open the box to re-land anything.
What do I order for two QR scanners over a coolant waterfall?
A 16 x 12 x 8 in 304 stainless wall-mount body with a 150 x 100 mm [6 x 4 in] borosilicate window: 0.5 m² of surface sheds the 20 W at an 11 C rise, 46 C inside at 35 C ambient, 4 C under the 50 C scanner limit.
Run the numbers. Two scanners at 10 W each put 20 W into the box. The 16 x 12 x 8 in [406 x 305 x 203 mm] body carries about 2 x (0.406 x 0.305 + 0.406 x 0.203 + 0.305 x 0.203) = 0.50 m² of surface, and stainless sheds heat at 3.7 W/m²K, so the rise is 20 ÷ (3.7 x 0.50) = 11 C. Inside temperature: 35 + 11 = 46 C [115 F]. At a 40 C [104 F] machine ambient you reach 51 C, one degree over, so step up to a 20 x 16 x 8 in body: 0.78 m², rise 20 ÷ (3.7 x 0.78) = 7 C, inside 47 C. The enclosure heat calculation guide walks the same method with fans and bigger loads.
Materials: with a semi-synthetic at pH 9 you take 304 stainless — aluminum would etch — and nitrile seals. With straight oil, 304 and nitrile still hold; move to 316L only for chlorides or salt air. Order the window anti-reflective and 5 mm thick.
| Part | Order this | Why it survives |
|---|---|---|
| Housing | 304 stainless, 16 x 12 x 8 in, one-piece body | alkaline semi-synthetic etches aluminum, not stainless |
| Window | 5 mm borosilicate, 150 x 100 mm, AR-coated | 92 % transmission, immune to oil and biocides |
| Door seal | Nitrile gasket; FKM near hot surfaces | under 10 % swell in straight oil |
| Cable entries | 2 x M20 IP68 glands, 1 blanked spare hole | the assembly holds the gland's rating |
| Cable | PUR jacket sized to the gland clamping range | PUR is the machine-tool standard against oil |
| Mounting | standoffs holding the glass clear of the nozzle line | takes the pane out of the waterfall |
You set the material, the width, height, depth and every one of those cutouts on a body like this, and it is delivered in 10 business days counted from your approval of the drawing. Lay the window, the glands and the mounting holes out in the cutout designer, then send the layout with your custom enclosure request.
How long do the window and seals last before you replace them?
Check them quarterly under a waterfall and about once a year in mist: pull any nitrile gasket that has swollen 10 % or gone soft, and re-bed or replace the window once wiping stops clearing the haze.
Keep one spare window, one gasket set and a pair of gland seals on the shelf, and treat read rate as your early-warning instrument. When first-pass QR reads slide a few points over a shift, the pane is filming and wants a wipe before the residue bakes on. If you are replacing parts monthly, stop buying spares and buy the air knife — or the spinning disc — instead.
Decide by what hits the box. Waterfall: glass, nitrile or FKM, an air knife and quarterly seal checks. Heavy mist: glass and nitrile, wiped every shift. Semi-synthetic carrying tramp oil: pick for the oil, and keep EPDM off the enclosure entirely.
Enclosures for this job
Catalog sizes that suit the job this article answers, picked from its subject; open one for its full specifications.
| Enclosure | Material | Outside H x W x D / display | Rating |
|---|---|---|---|
| 16 × 12 × 8 in | 304 stainless | 16 x 12 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 20 × 16 × 8 in | 304 stainless | 20 x 16 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 24 × 20 × 8 in | 304 stainless | 24 x 20 x 8 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
| 30 × 24 × 10 in | 304 stainless | 30 x 24 x 10 in | Rated NEMA 4X / IP66 by the manufacturer; not UL listed |
Questions
Does a higher IP rating, like IP67, add any coolant protection?
No. IPX7 means 1 m of clean water for 30 min and IPX8 is continuous immersion on conditions agreed with the maker, and both still say nothing about oil, amines or biocides. The window, seal and jacket materials decide survival, so a glass-and-nitrile IP65 build outlasts a plastic IP68 one beside a spindle.
How long does a polycarbonate window actually last in coolant?
Under a direct waterfall, field results on molded plastic housings run about 3 days to visible crazing, with cracks starting at hinges and other stressed features. In oil-free mist you may see several weeks before haze closes in. Straight oil and tramp oil are the killers: they wick into surface stress and open microcracks no wipe can remove.
When does the housing need 316L instead of 304 stainless?
Go 316L when chlorides are in play: coastal salt air, brine washdown, or a coolant carrying chlorinated extreme-pressure additives. Its 2 to 3 % molybdenum holds off the pitting that 304 succumbs to. Straight oil and an ordinary semi-synthetic at pH 9 leave 304 untouched.
Can I bed the window in RTV silicone instead of a gasket?
Yes. A high-temperature RTV rated to about 200 C bonds glass to a steel or stainless frame and has carried camera windows through brutal service for years. The cost is service: you cut and scrape the bead to swap the pane, where a nitrile gasket swaps in minutes. Use RTV for permanent windows, gaskets for wear items.
How much compressed air does an air knife over the window use?
About 3.5 scfm per inch of knife at 80 psi [5.5 bar], so a 6 in knife draws around 21 scfm while it runs. It only needs to run with the spindle, so duty is a fraction of the shift. A spinning disc uses far less air but adds a rotating seal of its own to maintain.
What tells me a seal is failing before the camera gets wet?
Three signs: the gasket has grown more than 10 % in thickness or width, it stays compressed when you open the door, or you find oil tracking on the inside face of the gland plate. Any one means the elastomer is swelling, so swap it that shift. A nitrile seal costs you minutes of downtime; a drowned scanner costs you the machine.
