The refrigerator in your kitchen is likely the biggest and most expensive appliance you own. Equally big: the job of testing the things out to figure out which ones are worth the money.
Refrigerators were some of the first large appliances that the CNET Appliances team tackled here at our Louisville, Kentucky testing facility, and it took almost two years of hard work to perfect our protocols. We modeled our tests after the the ones conducted by well-established industry testing groups like the Association of Home Appliance Manufacturers (AHAM) and consulted with countless industry specialists to make sure we were getting it right.
The resulting tests are a lot of work, but they yield the sort of hard data that we need to identify the standouts. Here's a complete rundown of the entire process, from start to finish.
First things first: if you want to test refrigerators, you're going to need a controlled environment to test them in. After all, ambient temperature affects fridge performance -- if it's hot in your home, your fridge will need to work harder to keep things cold. Differences of even a few degrees can skew performance testing results, as can differences in humidity.
That all adds up to three words: climate control chamber. You can't test fridges without one, so we designed and built our own.
The room is big enough to accommodate four fridges at once. We situate each one into one of those wooden storage slots in the picture above -- they're designed to help mimic the real-world placement of refrigerators in modern kitchens, which typically limits the refrigerator's exposed surface area to the front and the front only.
The climate control chamber is hooked up to a dedicated HVAC unit and humidity control system that work together to keep things locked in right where we want them: 76 degrees F, with the relative humidity at 30 percent. We monitor those numbers through the system itself, and also through a set of sensors placed in the center of the room. Each time we run a test, we record the minute by minute temperature and humidity in the room alongside the temperatures in the fridge -- that way, we can be certain that everything stays steady for the duration.
Taking a refrigerator's temperature
If you want an accurate look at a refrigerator's cooling performance relative to other refrigerators like it, then you'll need to account for a few important variables. The first and most important is the amount of stuff you've got crammed inside -- the weight of which relative to the size of the fridge has a huge impact on how that refrigerator will perform.
To keep things even, we load each refrigerator with mass relative to its capacity. Our grocery analogues include bottles of water, two-liters of soda, and Tupperware containers filled with a propylene glycol and water solution. The same goes with the freezer compartments. Each one gets loaded with a fixed percentage of mass relative to capacity -- our carefully weighed grocery analogues include cans of frozen juice, frozen packages of spinach, and blocks of wood.
We make sure to load each fridge and freezer as evenly as possible, being sure not to overload any one shelf or drawer with too much weight. This is true for volume, too -- if all of the space is blocked off on a particular shelf, the air won't circulate evenly. For instance, in that picture above, the weight appears to be distributed evenly, but the volume is a little too crammed on the top shelf. We'd likely move a few of those bottles to the other shelves to help spread things out a bit.
Once each fridge is properly loaded up with the exact amount of mass (down to a tenth of a pound), we can start adding in the thermocouples that take our minute-by-minute temperature readings. Really just a thin wire that detects temperature at the tip, we stick each one into a jar of our propylene glycol solution, then seal it in with some industrial putty. The reason? Sealing the thermocouples in a jar helps protect against intermittent air currents throwing our readings off, and suspending them in fluid helps to average out the surrounding ambient temperatures of whatever shelf we stick them on. We use the propylene glycol mix because the stuff won't freeze if the fridge runs cold (it's a primary agent in antifreeze, as a matter of fact).
We put a jar on each shelf and in each drawer, making sure that the temperature-sensing tip of the wire isn't touching the sides of the jar. Then, we run the wires out of the fridge, keeping them as flat against the inside of the door as possible, and tape them in place. If we've done it right, the tiny gaps in the door's seal will only have a marginal affect on the results, if any.
Next up is the freezer. We don't use jars of propylene glycol here -- instead, we stick each thermocouple into special blocks of wood that we've drilled thin holes into. Like with the jars of glycol, the wood protects the thermocouples from the cycling air currents, and instead, gives us a more accurate look at the average temperature in each section of the freezer.
The thermocouples from the fridge and the freezer plug directly into a wall panel behind each fridge station; the data then feeds into our fridge testing computer. We run a custom-built piece of software that tracks the temperatures on a second-by-second basis -- every minute, it'll spit out readings for each thermocouple into a waiting spreadsheet.
Before we can start taking readings though, we need to "burn in" each fridge. It takes time for a refrigerator to reach its temperature setting and then stabilize -- usually at least twelve hours. We make sure to let each one sit for a full 24 hours. After that, it should be stabilized down at the default temperature (typically 37 degrees) -- that's the first test we run.
During the test
After all of that, we're finally ready to start testing. Once we tell our software to start, it will automatically begin logging those minute-by-minute temperatures. All we need to do is wait 72 hours, and we'll have the tens of thousands of temperature readings needed to know how well the fridge performs.
But people don't typically just let their fridge sit for three days straight. Most of us open our fridge at least a few times a day, which affects how the fridge performs. It's important for us to simulate that kind of average, everyday usage.
That's why we run regularly scheduled door openings during each test, one for every twelve hours of the test. That makes a total of five door opening cycles (when it's time for the sixth, the test is over). Here's how a door opening cycle works: we open the fridge for 30 seconds, then close it and open the freezer for thirty seconds, then close it and wait 30 seconds, then repeat all of it three more times.
Once the test is done, we crank the fridge down to its lowest setting (usually 33 degrees F, but sometimes 34). After waiting another 24 hours to burn in the fridge at that stabilized temperature, we start all over again, and run another 72-hour test.
Typical test results
Once both tests are done, we'll typically have 100,000 temperature readings to sift through, if not more. The easiest way to wrap our heads around all of those numbers is to graph them out.
We color-code each graph based on a section-by-section basis. Blue lines represent the main refrigerator shelves, green and purple lines represent the door shelves, and red lines represent the drawers. By the way, if any of those drawers offer variable temperature settings (like that "CoolSelect Pantry" in the graph above), we set them to the opposite setting with respect to the rest of the fridge (i.e. we crank it all the way down during the default test, then crank it all the way up during the coldest setting test.)
The other useful way of visualizing (and simplifying) that data is to average each section out, then plot the temperatures on a color-coded heat map of the fridge. We use a standardized color scheme for each fridge -- all you really need to know is that blue is good and orange is warm.
This performance data gives an objective look at each refrigerator's cooling abilities, making it well worth the week or so that it takes to accumulate it. But, it's only part of the story -- which brings us right to:
With performance tests done, we unload everything out of each fridge, then wheel it over to our storage testing area. We've got a full load of standardized test groceries meant to simulate the fridge contents of a busy family right after a trip to the store, everything from beer and soda to pudding cups and cottage cheese.
We look for a couple of things as we load each fridge up. Can our entire assortment of condiments fit neatly into the door shelves? Do we need to rearrange any shelves to accommodate tall items? Are we forced to pack things in tighter than we'd like and block off the air flow? Is the butter bin big enough to fit both a box of sticks and a tub of margarine? Is there a good spot to keep an open can of dog food isolated away from the rest of the fridge's contents?
We load each fridge up a couple of times to get a good sense of how easy it is to get things in and out, and to keep things organized. A lot of it is subjective, but it's a very helpful test, and one that simulates real-world usage quite well.
After we've got a feel for how the fridge handles our standardized grocery load, we bring out the big stuff: six stress test items. There's a pitcher, a cake tray, a party platter, a casserole dish, a roasting pan, and an extra large pizza box from one of my favorite local pizza joints. With the fridge fully loaded, we try and make space for each one -- first one at a time, and then all six at once.
It's another useful test, helping us to get a good sense of how well each fridge makes use of its own space. A fridge with a well-designed interior can actually outperform a bigger fridge with a less thoughtfully constructed arrangement of shelves -- we've seen it happen more than once.
Another important fridge factor that we take into consideration is how much energy it uses. Testing this out for ourselves is a more complicated endeavor than you might think, so instead, we rely on the independent testing done by Energy Star, and the "Energyguide" it releases with each fridge that it rates.
Those numbers tell you how much power the fridge will draw per year, but it's important to put them into context. A bigger fridge will probably use more power than a smaller one, because it has a bigger job to do, but that doesn't necessarily mean that it's less efficient.
I like to think about efficiency as the cost of cooling each cubic foot. The easy way to look at this comparatively is to divide the yearly energy cost listed on that Energyguide by the refrigerator's capacity. This tells you how much you're spending each year to keep each cubic foot cold. A more efficient fridge will give you a lower number. Fridges that perform well in our tests and have a low cost per cubic foot relative to similar models will almost always score very well with us.
From here, it's simply a matter of spending some time with each fridge and testing out its construction and features. How spillproof are the spillproof shelves? Does the ice maker jam easily? Is it a pain to move the shelves around? Does the stainless steel smudge too much? Are storage-minded features like slide-in shelves and drop-down bins actually useful? We look at all of it. Then, we share everything we've found.
As you can see, it's a long process testing each refrigerator out, but we think it's time well spent. Refrigerators aren't cheap, they sit front and center in your kitchen, and you likely use yours just about every day. If you're buying a new one, it's worth it to buy the one that's right for you -- and that makes it worth it to us to keep testing them out.