The Necessity of Online Conductivity Monitoring in Liquid-Cooled CDU Systems – A Technical Analysis and Engineering Practice

1. Background: New Operational Challenges in the Liquid Cooling Era

In this context, the coolant—often described as the “lifeblood” of liquid cooling systems—directly determines system stability and equipment safety. Conductivity, as the most sensitive indicator of ionic purity in the coolant, is evolving from a supplementary metric into a critical “lifeline” for system safety.

2. Root Causes of Rising Conductivity: Ionic Accumulation

In a closed-loop liquid cooling system, rising conductivity is not accidental. The underlying cause is continuous accumulation of ionic species. Primary sources include:

1. Metal Ion Leaching. Systems contain multiple metals—cold plates (copper or aluminum), piping (stainless steel), and fittings (brass, etc.). Galvanic corrosion between dissimilar metals releases active metal ions into the coolant. Non-metallic components (seals, hoses) may also leach ions.

2. Inhibitor Depletion and External Contamination. Additives such as corrosion inhibitors and biocides degrade over time. Rising conductivity in deionized water, propylene glycol (PG25), or ethylene glycol (EG25) solutions typically signals dissolved salt accumulation, inhibitor failure, or external contaminant ingress.

3. Operational Conditions. Prolonged high-temperature operation accelerates chemical aging, and temperature itself affects conductivity readings—making accurate temperature compensation essential for reliable measurement.

Once conductivity exceeds safety thresholds, the coolant loses its dielectric properties, leading to leakage currents, signal interference, and in severe cases, server short circuits or entire rack failures.

3. Risk Quantification: From Trace Ions to Major Incidents

The hazards of abnormal conductivity rise can be quantified across three dimensions:

Electrical Safety Risk. Higher conductivity means increased ionic concentration and enhanced fluid conductivity. In direct liquid cooling (DLC) and immersion systems, coolant comes into direct or indirect contact with electronics. Exceeding limits produces leakage and shunt currents, reducing efficiency and potentially causing short circuits. Globally accepted limits: ≤5 μS/cm for deionized water, and ≤1000 μS/cm for ethylene/propylene glycol-based coolants (at 25°C).

4. International Standards and Industry Consensus

Online conductivity monitoring has become a standard requirement in liquid cooling systems. Relevant international standards and industry guidelines clearly mandate such monitoring:

• ETSI TS 103 586 (European Telecommunications Standards Institute) provides a technical framework for liquid cooling solutions in ICT equipment, emphasizing continuous monitoring of key coolant parameters.

• OCP (Open Compute Project) has published foundational specifications for immersion coolants, defining limits for conductivity, pH, particulates, etc., widely referenced across the industry.

• ASTM International has established a dedicated subcommittee (D27) to develop unified standards for dielectric coolants in immersion cooling, reflecting the growing industry focus on fluid quality.

• Major international equipment manufacturers and operational guidelines commonly require that coolant conductivity (at 25°C) be kept below 1000 μS/cm, and for deionized water systems, below 5 μS/cm.

However, laboratory batch sampling has temporal blind spots—coolant chemistry can change significantly within hours. Consequently, continuous online monitoring has become industry best practice.

5. Engineering Implementation: Technical Requirements for Online Conductivity Sensors

To meet these monitoring needs, sensors used in liquid-cooled CDU systems must possess the following capabilities:

• Wide Range Coverage: From ultrapure water (near 0 μS/cm) to degraded high-conductivity states (several thousand μS/cm), supporting ranges such as 0–20, 0–200, 0–2000, and 0–5000 μS/cm to accommodate deionized water and glycol solutions.

• Strong Interference Immunity: Variable-frequency pumps, switching power supplies, etc., create complex electromagnetic environments within CDUs; sensors must incorporate electromagnetic isolation to ensure accuracy.

• Real-Time Response: Second-level response times to capture trend changes promptly.

• Material Compatibility: Wetted materials must be corrosion-resistant and non-contaminating—316L stainless steel body with EPDM seals is a mature industry choice.

The RK500-13LC liquid cooling conductivity sensor exemplifies these requirements: ±1%FS accuracy (0–5000 μS/cm@25°C), 1 μS/cm resolution, ≤1s response time, IP68 protection, and dual 4-20mA plus RS485 outputs for flexible integration into legacy or digital control platforms. It has been deployed in AI computing centers and new energy battery precision manufacturing for long-term online monitoring and trend-based early warning.

6. Conclusion

Liquid cooling is reshaping the infrastructure landscape of data centers and high-performance computing. In this transformation, the coolant is no longer a simple heat transfer fluid but a system-level component carrying multiple critical functions—electrical insulation, corrosion protection, and thermal management.

Conductivity, as the most sensitive and direct quantitative indicator of coolant health, has become an indispensable safety barrier in CDU systems. The shift from reactive repair to proactive risk warning, from periodic sampling to real-time awareness—this is not merely an upgrade in monitoring methodology, but a fundamental paradigm shift in liquid cooling operations and maintenance philosophy.