.jpg )
How replacing the ceiling tiles turned into a bigger project than expected
What happens when an electrical standards laboratory’s environmental control system gets old, cranky and hard to maintain? When the lab in question is the Fluke Everett Primary Electrical Laboratory (EPEL), the answer was to replace the air conditioning system, as part of a bigger endeavor to renovate the lab. The project was ultimately successful, but it almost derailed early on because of what some people might consider a small detail: replacing the lab’s ceiling tiles.
Since 1984, the EPEL ceiling was covered in compressed-fiber tiles, each perforated with hundreds of holes. Air flowed through the holes, providing a fairly uniform air supply into each of the four lab rooms. The old tiles were showing their age and needed to be replaced. Simple, right? Just go out and buy replacements. That’s what we thought, until we discovered that nobody makes these tiles anymore. By then we’d already disposed of the old tiles and the new air conditioning system was already being installed. What do you do when you can’t have what you were counting on?
The first solution: individual air supply vents
The team brainstormed and came up with “Plan B,” which involved installing individual air supply vents in various locations throughout the four rooms of our lab. Once the vents were installed, my colleague Dennis was asked to take measurements and determine how well the new system was working.
The air conditioning controller has one temperature sensor per room. Dennis took a Fluke Calibration 1620A “DewK” Digital Thermometer-Hygrometer to use as a reference; he set it up in one of the rooms, let it acclimate, and recorded the temperature in multiple locations within the room. Then he moved it to another room, let it acclimate, and repeated the process. He did that for all four rooms.
In the meantime, the other folks who work in the lab were also making notes. They would be setting up a test and they’d find it was too hot or too cold in that part of the lab, so they’d have to move. Soon it was clear: we could do a lot of things in the lab, but we couldn’t do everything in every corner of each room. We decided we could do better.
The second solution: do-it-yourself
As we brainstormed to find a better solution, someone suggested that we make our own perforated tiles. But how many holes should each tile have, and how big should the holes be? We wanted to simulate (more or less) what we had before. The engineers in Facilities and the maintenance guys in the shop thought the area of coverage (the percentage of holes per tile) was two percent. But nobody was sure and nobody could prove it.
I kept looking around and we found two of those old ceiling tiles, just lying against a wall in one of the Facilities spaces. It was like finding the Rosetta Stone!
Now, what was the area of coverage? I didn’t want to count every hole in one of those old two-by-four-foot tiles to figure it out. Instead, I marked off an area of about one square foot, and counted the holes within that area. Then I scaled that up to determine how many holes were actually in the whole tile.
The next thing was to figure out the size of the holes. I got a box of drill bits and started fitting them into the holes until I got a drill bit that closely represented the hole. Now I could calculate the area of coverage…and it turned out to be two percent after all.
Now the question becomes, how many holes and how big should they be, to make a ceiling tile with a two percent area of coverage? I created a spreadsheet and picked different sizes that corresponded to common drill bits. I made a table that showed, for each size of hole, how many holes would be required to get a two percent area of coverage.
Then I passed my spreadsheet around in the EPEL and in Facilities. Ok guys, what do you like? After some discussion, everyone agreed on an array with 300 holes per tile.
What is the best way to make the holes in each tile? We looked at having them laser cut, and we provided a sample to a local vendor with instructions to cut some holes. The laser-cut holes worked, except for one thing: they caused the tile to smell like burnt toast. We decided to look for another solution.
Drilling the tiles would be expensive, but we decided it was preferable to having a standards lab that smelled like burnt toast. So we found a vendor, got a price quote, and had the work done. The holes are very sharp and clean, with 300 holes per tile in the 100+ tiles needed to cover the ceilings in all four rooms.
Making it all work in the lab
By the time the tiles were done, we had placed a Johnson Controls sensor and a primary DewK sensor in the ceiling of each of the lab’s rooms. At that point, we had air coming out of the plenum above the tiles, down through the holes. Although that tells us the temperature of the air going into the room, it doesn’t necessarily represent the temperature further down throughout the rest of the room, where you’ve got equipment operating. More work was needed.
I went back to Facilities and asked them to make some simple sheet metal plates, about two feet in diameter. I put a plate, with holes in the middle for the sensors to fit through, on top of the tiles in each room, so it blocks the air from coming through the holes near the sensors. That way, we don’t have an air wash directly over the sensors.
After everything was in place, I repeated the survey that my colleague Dennis had done earlier. I was able to show that we now had something that worked throughout each room of the lab. The final step was to adjust the Johnson Controls sensor so that it measured the same temperature as our reference, the DewK.
And from there, things have actually been working pretty well.
If you would like to learn more about this project, read the paper I presented at NCSLI 2013, Temperature & Humidity Control in a Primary Standards Laboratory. It’s posted on the Fluke Calibration website.