40 mpg compact sedan comparison

MORE AT MOTORTREND.COM

• » $28K High-Performance Two-Door Comparison

• » 2013 Nissan Altima 2.5 SL First Test

• » Gunning for 200 MPH in the 2013 Ford Shelby GT500

Every once in a while, an advertising claim comes along that strikes a nerve. It makes you stop and glance up at the TV screen, or flip back to that previous newspaper page to contemplate its meaning. For me, it's usually a hair-loss treatment. However, for most people, it's probably something that impacts their pocket book, and a pretty good example of that in the car business recently has been the claim of "40 mpg!"

It's a powerful number -- 40 mpg. Car engineers do sophisticated analysis to figure out how to reach it, and the numerical game often comes down to just getting to the threshold where the number rounds up to 40.

It just so happens that we're practiced at rounding up, too -- in this instance, rounding up six significant examples of this rarified 40-mpg breed to determine if any of them are also cars you'd be happy to own.

The select six we've summoned to our investigation are the Chevrolet Cruze Eco (28/42), Ford Focus SFE (28/40), Honda Civic HF (29/41), Hyundai Elantra (29/40), Mazda3 (28/40), and Volkswagen Jetta TDI (30/42). A diesel, but no hybrids? Unlike the public's uneasiness with battery-aided driving (hybrid sales have been stuck around the 2.5-percent range), diesels are hydrocarbon kissing-cousins to their gasoline brethren. We know them. We're comfortable with them. And we were also curious to see how the Jetta TDI stacks-up against its high-tech gasoline alternatives.

In search of 40-mpg answers, we drove from our El Segundo nerve center to the high desert north of Los Angeles and two destinations we normally visit for our Car of the Year programs. At the first, we grunted our way through a zillion mind-numbing laps of the Hyundai-Kia Proving Ground's 6.4-mile oval in a sequence of stepped, constant speeds to establish our fleet's steady-state fuel-efficiency.

The next day, each car was repeatedly piloted around our familiar Tehachapi real-world driving loop, a richly informative 27.3-mile mix of small-town stop-sign streets, curvy rural roads, elevation changes, and freeway miles. (Another plus is that it's 80 miles from Los Angeles' super-saturated roadways, which are ever on the brink of cardiac arrest.)

Let me explain the meaning of the results you'll see on the following pages. Each car's average Tehachapi mileage, and fuel cost per mile doing so, are obvious. But we've added a figure to this that represents how depressed this is compared with these lap's theoretical mpg modeled from our constant-speed mileages -- that is, what they'd ideally do were they unimpacted by accelerating, fidgeting with the throttle, hill-climbing, and stopping. If you don't drive, uh, very gingerly, pay attention to this one.

Testing the Car's Mileage -- and Our Patience
Measuring things like 0-60-mph times and panic stops is old hat for us. But discerning fuel mileage, that's another issue. At a cruise-controlled 60 mph, one of our cars, the Chevrolet Cruze Eco, returned 49.4 mpg. That's 0.0202 gallon per mile, or, if you were to picture that gasoline as a series of teaspoons, one burned every 4 seconds.

A Nissan GT-R can accelerate from 0 to 238 feet in 4 seconds; 238 feet I can measure. But teaspoons of gasoline? Being combusted deep within a densely packed engine? Measure this?

We monitored the cars mass airflow sensors (excepting the Elantra's, which employs a different strategy), and, as long as the car is in this steady mode, knowing the air passing through the engine points directly to its fuel consumption. This is what was displayed in a filtered, averaged way on our cars' instrument panels. As a check of this information, we simultaneously logged the cars' slow decline in fuel tank level, and, after calibrating the fuel tanks (incrementally measuring their level versus gallons added), their constant-speed fuel consumption data could be backed up.

Divergence
Our driver's mileage did vary. You hear all the time that you shouldn't expect your driving to duplicate the EPA's super-scientifically derived mileage numbers. Well, before we started our Tehachapi lapping, our drivers were firmly directed to observe the posted speed limits, which inadvertently put our own mileage reproducibility to the test. And the results were actually more divergent than I expected. Below, we list numbers -- and name names.

Oddities? Our differing driving habits had nearly 10 times the influence on the Elantra's mileage than on that of the Focus. Perhaps the Hyundai's more susceptible to enthusiastic outbursts? Mr. Martinez' idea of sticking to the speed limit was taken as more of a suggestion.

Under Pressure
Every time gas prices spike, the AAA and the EPA nag us to check our tire pressures. After our 40-mpg test-o-rama was complete, I did a little experiment with our long-term Civic Si (the performance-opposite to our test HF). At a cruise-controlled 60 mph, with the tires set to the door sticker numbers, the Civic returned 38.42 mpg. After stopping and immediately lowering them by 5 psi, the mileage (deduced from the mass airflow sensor) declined to 38.18 mpg, or 0.6 percent. Car engineers would kill their dear old moms for 0.6 percent. And all from a little squirt of compressed air.

AutoEnginuity
To raise our data-gathering game we've reached out to Jay Horak and his OBD2 scan tool company, AutoEnginuity, which has become a leader in its field. Don't know what OBD2 is? Feel around under your dash, and you'll discover a hexagonal-shaped receptacle that technicians use to diagnose troubles and check your car's emission system. We're employing it to log speed, rpm, mass airflow, and fuel-tank level -- and, in the case of the Jetta, its diesel consumption in gallons per hour. Cool.

An Anomaly
To measure the cars' speeds accurately, we've turned to our longtime GPS pals, Vbox USA (aka Racelogic). Their data loggers have become the industry standard for GPS-data recording, and they've brought along additional units to supplement those we use for our regular performance testing. A curiosity we've noted is that all the test cars (save the Jetta) are actually traveling slightly faster than their speedometers indicate. Our best explanation is their new (unworn) tires.

'Theoretical mileage?'
We applied each car's various constant-speed mpg results -- acquired under near-perfect circumstances at the proving ground -- to every portion of our Tehachapi Loop having the same corresponding speed to calculate each car's absolutely ideal loop mileage. The difference tells us how real world roads -- and driving -- impacts those constant-speed curves.