I'm a facilities manager handling cooling and infrastructure orders for a mid-sized colocation provider. I've been doing this for about six years. In my first year (2019), I made a classic mistake that cost roughly $3,200 in damaged gear and a 72-hour service outage. Everything I'd read about industrial dehumidification said 'just match the CFM and get a unit that works.' In practice, I found that the conventional wisdom skips the most critical step: defining the process load, not just the space volume.
The failure happened in August 2019. I ordered a standard off-the-shelf desiccant dehumidifier for a 400 sq ft server room. It looked fine on paper—matched the room CFM, had a built-in humidistat, and cost about $1,800. The result came back: the unit couldn't keep up with the latent heat load from 12 server racks running at 70%. The humidity spiked to 75% RH in under three hours, condensation formed on the cold aisles, and I had to emergency-evacuate two racks. That's when I learned that a Munters dehumidifier isn't a commodity; it's a precision instrument that requires a specific specification process.
After that disaster, I created a five-step checklist for selecting dehumidifier Munters units for our data centers. Here's the system I use now. It's designed for anyone managing a server room, a small data center, or a critical storage environment.
This is the step I missed. You don't buy a dehumidifier for the room; you buy it for the heat and moisture generated by the equipment inside it. Most vendors will ask for square footage. That's not enough.
Here's what I check now:
I add these together and then oversize the dehumidifier's capacity by 20% (surprise, surprise—undersizing was my original sin). The Munters evaporative cooling hybrid units are good here because they can handle both sensible and latent loads, but you still need the correct base capacity.
This one caught me off guard in 2022. We spec'd a standard desiccant wheel unit for a new server room. The room was fine at 75°F. But the supply air entering the dehumidifier's evaporator coil was only 65°F (due to a chilled water loop upstream). The coil surface temperature was below the dew point, and the unit started frosting internally. Performance dropped 40% in two hours.
If you're using a Munters dehumidifier with a pre-cooling coil, ensure the entering air temperature is above 55°F at the evaporator, or you'll need a hot gas bypass valve. Most standard units don't include this unless you specify it. I now add a line item to my spec sheet: 'Confirm min entering air temp at evaporator coil.'
In September 2022, I installed a new unit on a raised floor. The condensate pump had a rated lift of 10 feet. The drain line needed to go up 15 feet (through a ceiling plenum to the main building drain). The pump was undersized. The condensate pan overflowed, water seeped into the subfloor, and we discovered it during a routine inspection. It was $800 in cleanup and a 1-day delay in commissioning.
I now check three things:
Most dehumidifier Munters units specify the condensate pump specs on the technical data sheet. Don't skip the math here. I recommend adding $150–$200 for an external lift pump if the internal pump isn't sufficient.
You'll see an 'Energy Efficiency Ratio' on the spec sheet. Take it with a grain of salt. That number is usually calculated at 80°F and 60% RH. If your actual operating conditions are different (e.g., 70°F and 45% RH), the efficiency changes.
I've started asking for a 'part-load efficiency curve' from my vendors. If they can't provide it, I run the unit at partial load for 24 hours and measure the actual kWh consumption. On one job, the sticker said 1.5 kW. The actual consumption at our operating point was 2.1 kW. That changed my total cost of ownership (i.e., not just the unit price but the operational expense over three years).
Interestingly, I've found that Munters evaporative cooling units can be more efficient in indirect mode than direct expansion units, but only if the ambient temperature is below 85°F. In hotter climates, the desiccant wheel needs to regenerate more frequently, which increases energy use. The conventional wisdom says 'evaporative is always more efficient.' My experience with 20+ installations suggests otherwise.
I'm not 100% sure why this isn't standard, but it's not. Most units come with one humidity sensor. If it fails (and they do, especially in high-vibration environments like near an AHU), the unit runs in a default cycle and the humidity drifts.
I now specify a second digital humidity sensor (like a basic Sensirion SHT30 sensor wired to the BMS) in the return air stream. I also calibrate both sensors annually. The cost is about $80 for the sensor plus labor. The cost of a sensor failure is far higher.
I once had a sensor drift from 55% to 65% RH over six months without detection. We caught it during a scheduled PM check. The sensor was $45 to replace. The potential damage to servers from three months of elevated humidity? That's a $15,000 risk.
Even after all this, my team still messes up occasionally. Here are two that make me kick myself when I see them on a review.
I've seen a maintenance supervisor do this. A kerosene heater adds about 6.8 lbs of moisture per gallon of fuel burned. You're literally counteracting your dehumidification effort. Use an electric heater if you need to pre-warm the space.
It does. If you're in a cold climate with a heat pump, the heat pump's defrost cycle often dumps cool, moist air into the space. That means your dehumidifier needs a higher peak capacity to handle those defrost events. I learned this the hard way in January 2024.