Why can't a data centre spares case be airtight?

A: A dead seal is drawn out of shape by differential pressure, and the lid then dumps one violent exchange of dusty air.

Why is 10^4 ohm foam not a violation?

A: It only collects; charge then crosses the 1 MΩ stud, lifting the path into the 10^6–10^9 window.

Decay is over 2 s. Where first?

A: The foil-to-foam bond, then stud resistance. Foam failure is less likely than an aged joint.

Why change desiccant on 30 days rather than colour?

A: Indicators read local moisture and lag in low airflow; a 20% weight gain averages the pack.

Does weekly wiping wear the conductive layer?

A: A neutral cleaner does not; dry scrubbing does. Wipe damp, avoid alcohol and abrasives.

Data Centre Spares Case

Quick answer: A data centre spares case holds two numbers: 10^6 to 10^9 ohm on the surface and ISO Class 8 inside, bridged by a conductive liner and 1 MΩ stud.

Why Two Control Numbers
The Resistance Window
The Dust Side
Liner and Stud, Not a Sealed Box
Wiping and Desiccant
FAQ

1. Why Two Control Numbers

A data centre spares case carries NICs, optical modules, fan trays and drive caddies. In the rack they sit on a grounded chassis; lift one out and the technician plus the case body become the only route for charge.

So the case answers two questions: how fast charge leaves, and how much dust follows the lid in. They pull against each other. Dust control wants a tighter gasket and fewer openings; static control needs a path from liner to stud, plus a little air exchange against negative pressure.

A fully sealed box scores well on dust and fails on static. Field use: 12 openings a shift, tens of seconds each, in aisles where fibre and floor-wax dust hangs. The 0.5 µm band is what lands on gold fingers.

2. The Resistance Window

Surface resistance on a data centre spares case belongs between 10^6 and 10^9 ohm. Below 10^6 ohm, charge leaves too fast: a device lifted out with pins on a conductive surface discharges in microseconds, current hits amps, and the gate oxide is punched through. That failure rarely shows at the bench; it returns as intermittent packet loss in service.

Data centre spares case open on grey, foam in the lid, slotted black grid holding dark modules and a coiled white cable

Above 10^9 ohm, charge will not leave. Foam, dividers and walls are insulators, charge creeps away, and a technician walking up becomes the path through the module. Below 30% RH the same motion generates several times the voltage.

Inside a data centre spares case the 10^4 ohm foam is the collection layer, not a contradiction. It touches the module, takes charge off the surface and passes it through bonded copper foil to the stud, which carries a 1 MΩ resistor in series and lifts the path to the 10^6 ohm order. Collect low, release limited.

Metric	Target	Test method
Surface resistance	10^6–10^9 ohm	Parallel electrode, 500 V
Conductive foam	10^4 ohm	Surface meter
Stud resistance	1 MΩ	Megohmmeter across stud
Charge decay	< 2 s	Static meter, 1000 V

Charge decay under 2 seconds is the acceptance value: apply 1000 V to the liner, remove it, time the fall to 10%. Slower usually means the foil-to-foam adhesive aged, not the foam.

3. The Dust Side

Interior cleanliness targets ISO Class 8: no more than 3,520,000 particles at 0.5 µm and above per cubic metre. That is looser than a rack intake, tighter than an office. A data centre spares case need not be a cleanroom, only avoid dragging coarse particles in while the lid is open.

Dust arrives two ways. Displacement at opening, where 30 litres of interior air mixes with aisle air and two or three tenths is exchanged. And the gap between wall and liner, where dust settles and is shaken loose at the next opening. Open less for the first; press a gasket onto the second. Humidity sits at 30–60% RH. Under 30% charge generation climbs; over 60% metal parts condense, and an optical end face is the last place for that. The band serves both numbers and sets the desiccant charge.

Metric	Target	Test method
Cleanliness	ISO Class 8	Counter, lid open and shut
Particle size	0.5 µm	Lowest counter channel
Humidity	30–60% RH	Logger, 7 days
Desiccant	50 g per 30 days	Weigh, replace at 20% gain

4. Liner and Stud, Not a Sealed Box

The shell is PP or ABS, an insulator around 10^14 ohm. Making the shell conductive is neither practical nor necessary; put conductivity in the liner, let the shell carry structure and sealing.

Data centre spares case moulds on a workshop floor, red overhead crane above, steel mould blocks with guide pins, one worker in blue

Foam is milled cell by cell with 3 mm ribs between cells, so the sheet is one electrical body, not a dozen islands. Copper foil bonded to the back runs to the stud; the body is stainless, holds a 1 MΩ resistor, and leaves a 6 mm face outside for a clip.

Skipping the sealed box is the trade this data centre spares case makes. A double-lip gasket compresses 30% at the rim and stops the bulk of dust; a breather valve cracking near 10 kPa takes the differential from transport and temperature swing. A dead-sealed case has its gasket drawn out of shape and dumps one violent exchange when opened.

Check stud continuity on the wiping cycle: probes across the body should read 1 MΩ ±5%. Tens of megohms means a resistor or joint has degraded while surface resistance still looks perfect.

5. Wiping and Desiccant

Wipe the data centre spares case liner weekly with a lint-free cloth and a neutral cleaner, then air it 10 minutes. Weekly rather than per shift, because foam does not load up visibly; what gets dirty is the cell floor and the rib corners.

A 50 g charge lasts about 30 days in a 30 litre case opened 12 times a shift. Silica gel equilibrates near 20% of its own weight at 40% RH, so 50 g takes on roughly 10 g before saturation, after which humidity tracks the aisle. Judge by weight, not colour.

Metric	Target	Test method
Liner wiping	Weekly	Signed maintenance card
Desiccant change	30 days	Weigh, replace at 20% gain
Stud continuity	1 MΩ ±5%	Megohmmeter
Gasket compression	30%	Feeler gauge

The routine runs under 15 minutes a week, far cheaper than reworking a module wrongly judged good. A data centre spares case earns its keep on what comes out, not what goes in.

Conclusion

The difficulty is not any single number but holding two. Resistance inside 10^6–10^9 ohm lets charge leave without leaving violently; ISO Class 8 and 30–60% RH keep the interior clean after 12 openings. The fix is not to build the box shut: a liner through a 1 MΩ stud handles release, a double-lip gasket plus breather valve handles differential. Put both maintenance acts on the card and the case holds.