Data Center CRAC Unit Upgrade in Phoenix, AZ

Scenario

A Tier III colocation facility in Phoenix, AZ upgraded 24 rooftop CRAC units serving server rooms. Despite low ambient humidity (summer wet-bulb often ≤62°F), high internal latent loads from dense IT equipment (25 kW/rack) drove significant condensate generation. Existing ½-inch ABS drain lines frequently clogged due to mineral scaling from evaporative pre-coolers upstream—requiring emergency shutdowns. Constraints included roof-mounted units with limited vertical drop (<6 ft) and strict uptime SLA (99.995%), mandating zero drain overflow events.

Given Data

• Cooling Capacity: 180,000 Btu/h (per CRAC unit)
• Dry Bulb Temperature: 105°F (design peak)
• Wet Bulb Temperature: 62°F (low-humidity desert condition)

Calculation

Using the Condensate Drain Line Sizer:

1. Estimate latent load:
   • IT equipment generates near-pure sensible load, but humidification systems and personnel contribute latent load. Per ASHRAE TC 90.4, latent load ≈ 12% of total cooling capacity → 180,000 × 0.12 = 21,600 Btu/h
2. Mass flow: 21,600 / 1050 ≈ 20.6 lbₘ/h
3. Volumetric flow: 20.6 / 8.34 ≈ 2.47 gal/h = 0.041 gpm
4. Tool input yields:
   • cooling_capacity: 180000
   • dry_bulb_temperature: 105
   • wet_bulb_temperature: 62
   • → condensate_flow_rate: 0.04 gpm (rounded)
   • → pipe_diameter: 0.50 in (½-inch)

Result and Decision

Although the tool suggested 0.50-inch pipe, field data showed ½-inch lines failed within 8 months due to calcium carbonate scaling from pre-cooler bleed water mixing with condensate. Engineers selected ¾-inch stainless steel tubing (not PVC) with internal electropolished finish, installed with 1/2″ per foot slope and automatic flush valves triggered every 72 hours. Pipe size was increased not for flow capacity—but to reduce velocity (<1.5 ft/s) and minimize scaling adhesion.

Lesson

Low flow rate ≠ low maintenance risk: In arid climates with evaporative assist, mineral-laden condensate demands material selection and cleaning protocols—not just diameter sizing. Always validate tool assumptions against actual water chemistry and failure history.