Refrigeration
The Checklist That Keeps Our Underground Freezer Running: Daily, Weekly and Monthly
A lot of the trouble in our logs didn’t start as a big failure. It started as something small: a plugged strainer, a pan that didn’t drain, a level that crept up.
I run an underground freezer in the Valmeyer bluffs. The refrigeration side is a two-stage ammonia plant feeding a brine loop, with 34 refrigeration units out in the rooms. Our checklist lives on one spreadsheet, split into daily, weekly, every-other-week and monthly columns, with a box to tick for each weekday. In 2024 we tightened up the engine-room part into its own daily sheet.
Here’s what’s on it, and for most items, the problem from our own logs that it’s there to catch. This isn’t a list of saves. Some of these problems got past us before the check existed, which is exactly why the check exists now.
Daily
Room temperatures. Every room, every day. Our setpoints vary by room, roughly -15 to 8 °F, so “normal” is different from room to room. You only know a room is drifting if you know what it usually reads.
The compressor-room log. For each compressor, high side and low side: suction pressure and temperature, discharge pressure and temperature, oil level, oil pressure and oil temperature, written down by hand.
People ask why we still write this down when the controls system records it. One line from the safety paperwork answers it: computer monitoring does not replace daily logs. We learned that for real when we moved the plant from the old controllers to a new platform over the winter of 2024 to 2025. For a while the graphs and alarms weren’t reliable. The paper log was the record we trusted.
Oil is the one I watch hardest. After a chiller problem in 2023 got brine where it didn’t belong, we were still purging oil out of the system two years later. Oil level and pressure are where trouble on the refrigerant side tends to show up first.
Intercooler and chiller levels. Both chillers, every day. In spring 2025 we had compressor-room shutdowns that nobody could explain. They traced back to a high-level probe and an alarm delay, and the high-level alarm wasn’t even showing up. Looking at the levels yourself is how you notice a vessel running fuller than it should before the plant decides for you.
Purger check. A quick look that the purger is running and not in alarm. The weekly count below is the part that tells the story.
Water treatment and chemical pump filters. Check that the chemical feed is working and the filters are clear. This one cost us in August 2026. A plugged strainer on the supply pump for the tower’s water-treatment controller killed pH and blowdown control. Conductivity spiked past 7,000 and pH went over 9, and scale built up on the tower in about four days. Cleaning it meant descaler and pressure-washing the tubes. The fixes that came out of it: move the chemicals from drums to bulk tanks so they never run out, and set the controller alarms to send email.
The units. A walk past the evaporators, with the air-unit readings kept on their own sheet.
Wireless temperature sensors. We put wireless temp sensors in the rooms in 2024. The daily check is just that they’re all reporting. A sensor with a dead battery looks a lot like a room with no problems.
Weekly
Purger reset and total. Once a week we write down the purge counter and reset it. Comparing one week’s total to the last is a cheap way to spot a change in how much the system is purging.
Check the drain pans on the units, from the lift. This is the one nobody wants to do, because it means going up in a lift. It’s also the one that catches frozen drains. Our daily logs from January 2025 have a unit found frozen again, a drain we suspected was frozen on another, and heat pads in the pans. In July 2026 a drain line refroze and got new heat tape. Water that can’t leave the pan turns into ice on the coil.
Clean the compressor room. A clean floor shows a new oil drip or leak the day it starts.
Roadways and entrance. Sweep, pick up trash and clear leaves at the entrance.
Every other week
Sweep the docks.
Monthly
Grease the motors.
Power-wash the compressor room.
Clean the signs.
When it’s cold, and the extras
Run water in exposed lines when it freezes. In a December 2025 cold snap a cooling tower froze. Anything that sits outside the cave sees real winter, even if the freezer rooms never notice.
Check fuel in the standby generators.
Keep the temperature sensor batteries good. Same reason as the daily check: a quiet sensor isn’t a good sensor.
How to build your own
- Tie each item to a failure. If you can’t say what a check is there to catch, it’ll be the first one people skip.
- Write numbers, not checkmarks, where it matters. Pressures, levels, oil and purge counts are only useful if you can compare today to last week.
- Keep the paper log even with a controls system. Software changes, and alarms get missed. A log you wrote yourself is the record you can stand behind.
- Put the unpleasant checks on a fixed day. The drain pans get done because they’re on the weekly sheet, not because someone feels like going up in the lift.
- Add items after incidents. Most of our list is lessons, written down so we don’t pay for them twice.
The temperature part of this lives on our log sheet. If you want a starting point, get the free cold storage temperature log template through the signup box below, and build the rest of your checklist around it.
Related: The Day From Hell, on what a bad week in a heat wave looks like, and Snowstorm Behind the Racks, on frost from an air leak.
Get the free temperature log template
Get the free cold storage temperature log template, plus new posts by email. It's a printable sheet and CSV for daily freezer and cooler checks. See the template. No spam, unsubscribe anytime.