For cooling systems engineers and design engineers at CRAC/CRAH and liquid cooling OEMs.
The seal gets specified late. That’s the pattern. Not because engineers forget it, but because it looks like a commodity decision. The cooling unit is 90% designed, the airflow path is locked, the compressor is selected, and someone notes that the door perimeter needs a gasket. A standard EPDM D-profile goes on the BOM. The project moves.
Three things happen next, roughly in order: the unit ships, the install environment turns out to be more aggressive than the spec assumed, and the seal starts to fail, not catastrophically, not all at once, but enough. Thermal performance degrades. Condensation shows up where it shouldn’t. Noise levels creep. And when the field team pulls the door off six months in, they find a gasket that has compressed down to roughly half its original cross-section and isn’t recovering.
That’s not a quality problem with the gasket. That’s a specification problem that was made in week fourteen of a twenty-week design cycle, when nobody had time to interrogate it.
What ‘Standard’ Actually Means in a Cooling Application
Standard rubber profiles are designed around general sealing requirements: door-to-frame contact, basic environmental exclusion, a compression range that works across a broad set of use cases. For general HVAC equipment going into a climate-controlled building, that’s often fine. The unit sees relatively consistent temperatures, the door opens and closes a predictable number of times per day, and the environment doesn’t change much.
Data center cooling equipment operates differently. CRAC and CRAH units run continuously. The compressor cycles. The internal environment oscillates between cold supply air and warmer return air. The door, if it’s being opened for maintenance, is being opened under load, against a pressurized air path. And the unit itself may be sitting in a hot aisle/cold aisle configuration where the temperature differential across that door gasket is significant.
Under those conditions, a standard gasket that wasn’t specified to the actual use case starts failing in predictable ways:
- Compression set accumulates faster than the profile was rated for, because the seal is being held at compression continuously rather than intermittently.
- Silicone foam compounds, if chosen for temperature range, can be too soft for the closure force the door hardware delivers, leading to over-compression at the hinge and gap formation at the latch.
- EPDM closed-cell sponge profiles, often the default for HVAC doors, don’t recover well after repeated thermal cycling if the operating temperature exceeds the compound’s optimal range.
- Liquid cooling circuits introduce a different problem entirely: the seal at a coolant line penetration isn’t just an air seal. It’s a fluid barrier. The compound selection has to account for the coolant chemistry, and many standard nitrile profiles aren’t compatible with glycol-based coolants at elevated temperatures.
None of these are catastrophic failures. They’re drift failures, performance degradation that happens slowly enough to be missed until the downstream consequence is obvious.
Where the Specification Decision Actually Lives
The engineering question isn’t just “what compound?” It’s “what are the actual boundary conditions this seal will see, and what does that mean for compound, cross-section geometry, durometer, and closure force?”
That question almost never gets asked late in a design cycle, because late in a design cycle the door geometry is fixed and the hardware is selected. The gasket has to work with what already exists. Custom geometry becomes difficult. So engineers reach for the closest standard profile that fits the space.
The right time to answer that question is earlier, when the door geometry and closure force are still adjustable, and when the thermal and fluid environment of the unit is understood well enough to select a compound that actually matches the application. At that point, the gasket specification isn’t a commodity decision. It’s a design decision with real consequences for the unit’s long-term thermal performance and maintenance cost.
The practical implication for OEMs: the sealing supplier needs to be involved before the BOM is finalized, not after. That’s not a pitch, it’s a description of where the decision has leverage.
Compound Selection: The Variables That Matter
A simplified map of the decision, by application type:
CRAC/CRAH door perimeter seals
The primary requirements here are consistent compression set resistance, closure force compatibility, and thermal stability across the unit’s operating range. Silicone sponge performs well in high-temperature applications and has excellent compression set properties, but requires careful durometer selection relative to the door’s closure force. EPDM closed-cell sponge is the common alternative, lower cost, good UV and ozone resistance, acceptable compression set if the compound is properly specified. The failure mode when this is wrong is gradual: air bypass that shows up as thermal inefficiency before it shows up as a visible gap.
Liquid cooling penetration seals
Glycol-based coolants are aggressive toward standard nitrile compounds over time. EPDM is chemically resistant to water-glycol systems and is the correct starting point for most liquid cooling circuit penetrations. The geometry matters here too, a pipe penetration seal needs to maintain compression against the pipe OD over the thermal cycling range of the coolant circuit. If the seal relaxes under repeated heat-cool cycles, you get a slow leak that’s hard to diagnose until the coolant level drops. Silicone performs at higher temperatures but may not have the chemical resistance profile for all coolant formulations. Viton offers the best chemical resistance across a range of fluids but comes with a cost premium and hardness considerations.
EMI gaskets for switchgear and power distribution cabinets
EMI gasketing in a data center context is a different material family entirely, fabric-over-foam or conductive elastomer, with the sealing function secondary to the shielding function. The common failure here isn’t compound-related; it’s geometry. An EMI gasket that isn’t maintaining consistent contact across the closure surface creates shielding gaps. The specification question is about contact resistance, closure force, and the geometry of the door interface, not just material selection. Standard fabric-over-foam profiles from catalog suppliers work for many applications. Custom cross-sections become necessary when the door geometry doesn’t match standard profiles, or when the closure force is low enough that a standard profile’s initial compression resistance leaves gaps.
Vibration isolation for compressors and fan assemblies
Isolation mounts are load-bearing parts. The durometer and geometry of the mount determines the isolation frequency and the deflection under load. An undersized mount doesn’t isolate, it transmits. An oversized mount deflects excessively and can affect the alignment of connected components. Custom molded isolators, specified to the actual weight and frequency of the unit, outperform catalog parts for demanding applications. For standard CRAC/CRAH units with well-characterized vibration profiles, catalog isolators often work. For custom or high-load cooling systems, the geometry and compound need to be engineered to the specific application.
The Domestic Sourcing Variable
Lead time on custom rubber profiles has become a real design constraint. Offshore suppliers can offer cost advantages on standard profiles at volume. But custom cross-sections, low-to-mid volumes, and programs where the specification isn’t final until relatively late in the design cycle are all conditions that favor domestic manufacturing.
The relevant variables are lead time, minimum order quantities, and revision flexibility. A cooling unit OEM on a 12-week cabinet launch window doesn’t have room for an 8-week lead time on a revised gasket profile. If the first article reveals a fit issue, which it often does, the time to revised tooling and re-sample matters. Domestic suppliers can typically turn revised extrusion tooling in two to three weeks. Offshore tooling revisions, with shipping, can double or triple that.
That’s not an argument against offshore sourcing in all cases. It’s a description of where the tradeoff actually bites. Programs with stable geometry, high volume, and long lead time tolerance can absorb the offshore model. Programs with design iterations, compressed timelines, or low-volume custom profiles often can’t.
What to Look for in a Sealing Supplier for Cooling Applications
A few practical criteria, in the order they tend to matter:
- Compound range. Can they run EPDM, silicone, nitrile, neoprene, and Viton in-house? Or are they a single-compound shop that will substitute when a spec calls for something they don’t run?
- Extrusion and molding in-house. Custom profiles often require both, an extruded base profile with a molded end cap or corner piece. Suppliers who subcontract one or the other introduce a coordination point that adds time and variability.
- Die-cut capability. Flat gaskets, blanking panels, and custom die-cut parts are a separate manufacturing process. If the supplier can’t run die-cut work, you’ll be managing two vendors for what should be a single-source sealing program.
- First article process. A supplier that can turn around first article samples and provide dimensional reports within a defined timeframe is a different risk profile than one that can’t commit to sample timing.
- Minimum order flexibility. Cooling unit OEMs often need low-volume custom profiles during development and higher volumes at production. A supplier with rigid MOQs creates friction at both ends.
The Specification Decision Is a Design Decision
The seal isn’t a commodity part. It’s the interface between the thermal environment inside the unit and everything outside it. When it’s specified correctly, to the actual compound requirements, the actual closure force, and the actual fluid chemistry, it’s invisible. It does its job for the life of the unit.
When it’s specified to a standard catalog profile because the decision got made late and there wasn’t time to interrogate it, it drifts. Performance degrades. Field teams pull doors and find gaskets that have been running at 30% of their original cross-section. And the root cause traces back to a week-fourteen BOM entry that nobody had time to question.
The fix is to treat the sealing specification as a design input rather than a procurement afterthought. That means involving a sealing supplier earlier in the design cycle, when the geometry and hardware are still adjustable, and when compound selection can be matched to the actual boundary conditions of the application.
That’s where the leverage is. Not in the gasket itself, but in when and how the decision gets made.
Universal Polymer & Rubber manufactures custom rubber seals, gaskets, and vibration isolators for cooling system OEMs. Capabilities include EPDM, silicone, nitrile, neoprene, and Viton compounds, with in-house extrusion, molding, and die-cutting. Manufactured in the U.S.