Ask most people speccing a liquid cooling loop what matters in a filter housing, and you'll get answers about micron rating, flow rate, maybe pressure drop. Almost nobody asks what the housing itself is made of until a few years in, when corrosion byproducts start showing up somewhere they shouldn't.
That's usually the point where 304 stainless or brass gets swapped out for 316L. Here's the actual reasoning behind that switch.
 |
SS316L filter housing engineered for long-term corrosion resistance in CDU liquid cooling systems. |
The corrosion math nobody runs:Stainless steel's resistance to pitting in chloride and glycol environments is quantified by PREN Pitting Resistance Equivalent Number based on chromium, molybdenum, and nitrogen content. Run the numbers and the difference between grades gets concrete fast:
304 stainless comes in around PREN 18-19, since the alloy has no molybdenum. 316 jumps to roughly 24-26 once 2-3% molybdenum is added. 316L lands in that same 24-26 range the molybdenum content is identical to standard 316. The difference is carbon: capped at 0.03% in 316L versus 0.08% in 316.
That's the detail people get wrong. The "L" doesn't improve pitting resistance over 316 — the molybdenum does, and it's the same in both. What lower carbon buys you is resistance to sensitization: when 316 gets welded, carbon migrates to the heat-affected zone and can precipitate chromium carbides at the grain boundaries, depleting chromium right along the weld seam. That creates a corrosion-prone stripe exactly where you can't afford one. A filter housing is nothing but welded seams and fittings, so this isn't theoretical. 316L keeps carbon low enough that it doesn't happen.
304's missing molybdenum is the bigger problem for CDU coolant specifically, since most glycol-water blends carry chloride-based corrosion inhibitors and chloride is what drives pitting in the first place. That's why 304 housings that hold up fine in plain water start pitting within a couple years in treated glycol.
What it costs you if you get it wrong
Pitting in a housing doesn't usually fail the housing first it fails quietly, by shedding. Once a pit forms, the surface starts releasing metal ions instead of blocking them. Those ions travel downstream and settle wherever flow slows or geometry narrows, which in a CDU loop means cold plate microchannels often 100 to 300 microns wide, narrower than a human hair. You don't get a dramatic housing failure. You get a slow, undiagnosed rise in cold plate thermal resistance that looks like a cooling problem until someone finally opens a microchannel and finds deposits.
Brass fittings have their own version of this called dezincification, where zinc leaches out of the alloy under certain water chemistries, leaving a porous, weakened structure and putting copper ions into the coolant.
When 316L isn't actually necessary
To be fair to 304 and brass they're not wrong choices everywhere. A commissioning flush filter used once at startup doesn't need 316L's long-term resistance; it's in service for hours, not years. Facility-water-side strainers running plain treated water see much less aggressive chemistry than the technology cooling loop does. Specifying 316L across every position regardless of duty cycle isn't better engineering, it's just higher cost with no matching benefit.
Where 316L earns its place is the positions that stay wetted continuously, for years, in glycol-based coolant the main cooling loop's secondary filter housing and the side-stream polishing circuit. Those are the two spots where sensitization resistance and long-term pitting resistance actually get exercised over a decade-plus service life [VERIFY typical service life figures].
How this fits the bigger picture
Housing material is one variable in a filtration strategy it doesn't substitute for correct micron rating, proper staging, or monitoring. For a deeper look at how primary, secondary, and side-stream CDU filtration stages work together, Brother Filtration's engineering breakdown covers where each position sits in the loop and what it's rated to catch.
Conclusion:
If you're speccing a housing for a position that stays wetted year-round in glycol coolant, 316L is worth the incremental cost not because of vague "corrosion resistance" language, but because of a specific, well-understood failure mode that 316L is metallurgically built to avoid. For short-duty or low-aggression positions, it's reasonable to spec down and save the cost. The mistake is picking one material for an entire system without asking which failure mode each position is actually exposed to.
Comments
Post a Comment