How does the US Environmental Protection Agency decide how far an electric vehicle can go on a single charge? The simple explanation is that an EV is driven until the battery runs flat, providing the number that goes on the window sticker. In practice, it’s a lot more complicated than that, with varying test cycles, real-world simulations, and more variables than a book of Mad Libs, all in an effort to give you a number that you can count on to be consistent and comparable with other vehicles on the road.
The start of EPA mileage testing
The EPA started testing vehicle fuel economy in 1971, and that initial testing still plays a major role in how modern cars are measured.
The year before, President Richard Nixon signed the National Environmental Policy Act of 1969 (followed by the Clean Air Act of 1970) and established the EPA with a mandate that included lowering motor vehicle emissions. Part of the EPA’s plan to reduce emissions was to let buyers know just how much fuel a car would use so they could cross-shop cars effectively.
Testing started with a route called the Federal Test Procedure. The EPA adopted an 11-mile (18-km) route that was originally done on real roads in Los Angeles. The route had an average speed of 21 miles per hour (34 km/h) and a top speed of 56 mph (90 km/h). Tailpipe emissions were measured, fuel economy was calculated, and the “city” fuel economy rating was born.
By the time the 10-mile (16-km) Highway Fuel Economy Test was added in 1974, the tests were performed in a lab on a dynamometer. Running tests on the dyno made them more consistent and easier to repeat, though it wasn’t perfect.
Small changes and tweaks were made over the years, with the biggest change announced in 2005. That year, the EPA announced changes to the test to meet new highway speeds, account for heating and air conditioning use, and make the test more relevant to real-world driving. Drivers weren’t able to hit the published numbers, and the EPA wanted to fix that. The system was introduced for the 2008 model year and is largely the one we use today.
Modern range testing
Today, automakers have two different test options for EVs. The automaker can decide that it wants to perform a “single cycle” test. On that test, the car drives the EPA city cycle over and over again until the charge runs out, then does the same on the highway cycle, starting with a full charge. The process is repeated for reliability. The alternative is that the automaker can perform a multi-cycle test that has completed four city cycles, two highway cycles, and two constant speed cycles.
The test cycles
The city cycle
The EPA’s Urban Dynamometer Driving Schedule is the official “city cycle” test loop. It is a complicated graph of time, vehicle speed, and allowable acceleration. The total test time is 1,369 seconds, the distance simulated is 7.45 miles (12 km), and the average speed is 19.59 mph (32.11 km/h). As with all of the tests, the exact speed required at each second of the test is laid out in a spreadsheet.
The highest speed reached on the test is 56.7 mph (91.25 km/h), and there are several periods where the vehicle sits stationary. Stationary seconds of the test made more sense when it was designed to measure a gas vehicle’s idle emissions and consumption, but it does still have some relevance today when it comes to climate control use and energy required to accelerate the vehicle.
The highway cycle
For higher speeds, vehicles complete the Highway Fuel Economy Driving Schedule (HFEDS). This test has a top speed of 59.9 mph (96.4 km/h) and an average of 48.3 mph (77.73 km/h), and it takes 765 seconds to complete.
Only the UDDS and HFEDS tests are required to certify an EV. But a top speed of 59.9 mph is a much lower highway speed than most drivers will experience.
Driving more quickly or using climate control can greatly impact range. More tests were introduced to help give a more realistic range, and they’re part of the 5-cycle test covered below.
The three extra cycles
The 5-cycle test adds three more tests to the UDDS and HFEDS.
US06 is a high-speed, high-acceleration test. It simulates an 8.01-mile (12.89-km) route and takes 596 seconds to complete. The top speed is 80.3 mph (129.23 km/h), and the average is 48.4 mph (77.89 km/h). Vehicles are required to accelerate to the test’s varying speeds more quickly than in the HFEDS, using more energy. It’s meant to better simulate real driving.
SC03 is the air conditioning test. It is a 596-second test simulating a 3.6-mile (5.79-km) route. The top speed is 54.8 mph (88.19 km/h), and the average is 21.6 mph (34.76 km/h). This test is much like a shortened version of the UDDS, with the addition of air conditioning.
The last drive cycle is the Cold FTP. It’s an exact copy of the UDDS listed above, but the ambient temperature must be −7.0 ± 1.7° C at the start of the test, with an average of −7.0 ± 2.8°C during the test. Instantaneous temperature temperatures may be above −4.0° C or below −9.0° C, but not for more than three minutes at a time during the test. At no time may the ambient temperatures be below −12.0° C or above −1.0° C.
The vehicle must be preconditioned, which means it is soaked for 12−26 hours at cold temperatures and then driven through the UDDS once. Preconditioning temperatures can be above −4.0° C or below −9.0° C, but not for more than three minutes at a time during the preconditioning period. At no time may the ambient temperatures be below −12.0° C or above −1.0° C. The average ambient temperature during preconditioning must be −7.0 ± 2.8° C.
Rolling road, not rolling on the road
Testing is done on a dynamometer, not on real roads. The dynamometer needs to be able to simulate the weight, inertia, and other road loads that affect the vehicle, including the vehicle’s total aerodynamic drag. The test starts with a “cold start,” another carryover from internal combustion models.
Automakers can determine road load, but it is published in EPA documents. The 2023 Tesla Model Y RWD, for example, uses 11 hp for its 50 mph road load, and the 2023 Jaguar I-Pace with 20-inch wheels uses 15.3.
No brand-new or worn-out vehicles
Before a vehicle can be tested, it must have covered at least 1,000 miles (1,609 km) repeating what is called the Standard Road Cycle (SRC). This cycle is a seven-lap drive on (or on a simulation of) a 3.7-mile (5.95 km) course. Acceleration and deceleration rates are set out in a chart saying how the test should be driven.
The battery must also be aged (the SAE’s term for it) using the same rules, but it can also be aged on a bench instead of in the car. Bench-test aging must simulate at least 1,000 miles on either the SRC or one of several other approved cycles. At the start of the range test, battery amp-hour capacity needs to be checked to make sure it’s within manufacturer specs.
The vehicle also can’t have more than 6,200 miles (9,979 km) accumulated, and the vehicle “shall be stabilized.” What counts as stabilized is up to the manufacturer, but it’s understood to mean everything is broken in.
Vehicles are tested at loaded vehicle weight. LVW is the curb weight plus 300 lbs (136 kg), and the curb weight is defined as the total weight of the vehicle, including “optional equipment that is expected to be installed on more than 33 percent of the vehicle line,” but not a driver or other payloads. An all-wheel-drive dyno is recommended, but an extra weight factor is added if a two-wheel-drive dyno is used.
Temperature check
Ambient temperatures are strictly controlled. The standard tests must be done at an ambient temperature of 20–30° C. Events surrounding the test, like recharging and sitting, must be done in the same temperature range.
Dynamometer specs are set out by government regulations. The test system must hit defined air flow rates—approximately 45m3/s of flow at 60 mph (95.56 km/h) for test SC03 as an example—and there are defined grids for both rectangular and circular fan outlet grids. Fixed-speed fans are allowed, but the maximum capacity is limited based on the test, and axial flow must be maintained.
The dyno must record vehicle speeds at a minimum of 10 Hz and ambient temperature and humidity at 1 Hz. Other recording specifications apply to gas vehicles. There are also calibration frequencies and other regulations meant to keep automakers in line and accurate.
The math, road load forces, and methods used to apply real-world forces to a vehicle attached to a machine are also set in regulation. Vehicles must coast down at realistic rates, even if they’re not on a real road. Regulators have set out exactly how nearly all calculations must be made.
Other forces that the dynamometer exerts on the vehicle are set out in SAE standards J2263 and J2264. Without getting into the intricate details, the point of the standards is to make the dyno’s rollers an accurate simulation of how the vehicle stops, goes, and coasts on an actual road.
This heavy regulation makes it all the more unusual that the actual driving is done by people. A chart sets out the exact speed the vehicle must hit on a second-by-second basis, but achieving it accurately is done by a human foot and eyeball.
Test procedures
For most tests, all of the vehicle’s accessories need to be turned off. That includes radios, lights, and the rest.
Tests that need the air conditioning system to be on require a set temperature of 22° C for automatic systems, and that temperature needs to be measured at least once every five seconds to make sure it stays in spec.
Manual A/C vehicles need the climate control unit to be set to full cold, max, and recirculation, with the highest fan speed settings. Manual systems must also be adjusted to different settings (like 50 percent fan, fresh air, and whatever temperature setting achieves 13° C at the vents) between seconds 186 and 204 of the relevant test.
If there is an auto setting at key-on, the automaker can use that. And if the vehicle has three or four-zone climate control, they can “use good engineering judgment” to set the rear controls. What constitutes good engineering judgment? That’s up to the automaker, though if the EPA disagrees, they can make a company re-do the test.
Measuring power
Instead of using the EV’s computers, battery power must be measured independently. Automakers need to add DC voltage metering, an amp-hour meter, and a watt-hour meter, and these must be installed so they capture all current into or out of the main pack. All AC energy going from the wall into the charger when the car is plugged in must also be measured.
Data recording
Eighteen parameters need to be recorded, though some are not required for every test. There are obvious measurements, including battery capacity and temperature, plus the amount of AC recharge energy, discharge energy, and the time that each part of the test started and ended.
After all that, we’ve finally arrived at the actual test. For vehicles using thermal conditioning, the battery must be fully discharged by driving it at a constant 65 miles per hour on the dyno. This is called the Constant Speed Cycle (CSC). Once discharged, the EV needs to be charged (using AC charging) while the vehicle sits for at least 12 and no more than 36 hours. Charging needs to start within three hours of finishing the discharge. The amount of energy required to recharge the battery is recorded for at least 12 hours or until the car is done charging.
For vehicles not using thermal conditioning, the vehicle must sit for 12 hours or until charging is done.
No matter which test the automaker chooses, the vehicle has to be rolled onto the dyno, not driven, and it can’t be rolled for more than a mile. The dyno test must start no more than an hour after the car is unplugged from the charging equipment.
The vehicle is then driven. For the city test, the vehicle repeats the UDDS test until the battery runs flat. The tests are done in pairs, where the vehicle must soak for 10–30 minutes between each pair. The soak is done with the power “off,” the hood is closed, the cell fans are off, and the driver can’t have their foot on the brake.
For the highway test, the vehicle does pairs of the HFEDS test. Between each loop, there is a 15-second pause with the key on, and after each pair, there is a soak of 0–30 minutes.
The standard says that the charge is done when the vehicle can no longer maintain the speeds specified in the test. Or when the manufacturer says it should be stopped because of safety reasons like high battery temperature or low voltage. Manufacturers can also set their own safety checks to stop the test. When it’s done, the driver must brake to a stop within 15 seconds. That marks the end of the test.
For both tests, DC discharge in amp-hours must be measured. It’s measured for the full time, including soaks. This info is plugged into an equation to come up with the window sticker’s consumption data and the MPGe number. MPGe is calculated using the assumption that a gallon of gas has 33,705 watt-hours of energy. Measurements taken here are also used for the estimated charge time on the EPA label.
Range math
On the single-cycle test, range is the total test distance driven from the start of the test until the vehicle comes to a stop at the end. The city distance is the total driven on the UDDS, and highway range is the HFEDS total.
To give automakers an official range rating, the two numbers are combined. The city value is weighted at 55 percent, the highway number at 45 percent. Decades ago, the EPA decided that’s a typical ratio for an average driver.
Before the highway number is plugged in, though, it is reduced. The EPA takes the HFEDS number and multiplies it by 0.7. That’s meant to account for aggressive driving and HVAC use. The EPA says that most automakers use this factor for their figures, but it’s one place where automakers can tweak their figures.
Automakers can use the 0.7, derived 5-cycle data covered below, an approved 5-cycle equivalent, or “using adjustment factors which are based on in-use data.” If an automaker thinks its models perform better on the road, it can use a better number and get a higher range. Most do not choose to do this, according to the EPA.
The choice to use thermal conditioning or not thermal conditioning is another variable an automaker can select. The former uses heating or cooling of propulsion systems and/or the cabin, the latter does not.
Using the thermal conditioning method, the vehicle must be plugged in while conditioning and AC energy going into the system is to be measured. This method measures AC input before the test and DC output during to calculate power use.
Using the no thermal conditioning method, DC output is measured during the test, and AC input to recharge is measured after the test.
If a vehicle is tested not utilizing the thermal conditioning method, it must be placed on charge within three hours of finishing the test.
Multi-cycle test
Running the Single Cycle Test required two full days of testing, tying up dyno time and adding expenses for the EPA and automakers. To help solve that problem, the Multi-Cycle Range and Energy Consumption Test (MCT) was created.
The Multi-Cycle Range and Energy Consumption Test runs eight cycles together: the UDDS, a 15-second pause, the HFEDS, a 10-minute pause, the UDDS, a 0-30 minute pause, a CSC, a 0–30 minute pause, and repeat. The battery is expected to completely discharge during the second CSC cycle. The first CSC cycle length should be adjusted by engineers to make sure that the final CSC is 20 percent or less of the total test distance.
For this test, another adjustment is made. The first city cycle happens with a full battery, which limits regenerative braking. As a result, the first UDDS is scaled to be valued at one-third of the others.
Total range is calculated instead of directly measured here, based on the distance traveled per unit of energy used and the battery’s usable energy.
Short multi-cycle range test
While the MCT cut the time to test a 150-mile (241-km) EV from 18.5 hours down to 4.5, automakers and the EPA still wanted a faster test. The Short Multi-Cycle Range and Energy Consumption Test (SMCT) is that test. Vehicles run two full UDDS cycles, pause for 0-30 minutes, run an HFEDS, pause 15 seconds, run US06, pause 15 seconds, run HEFDS again, pause 0-30 minutes, then run UDDS twice more uninterrupted.
In this test, the battery isn’t fully drained on the dyno. At the end of the cycles, it is pushed to a soak area and is discharged into a load bank. That discharge is measured, and the test ends when the battery can no longer output enough energy to maintain 61 mph (98.17 km/h).
Range is calculated as usable battery energy divided by energy discharge per unit of distance measured for each of the highway and city cycles. It’s then put into the same 55:45 weighted equation as the single-cycle test.
The Short Multi-Cycle Range and Energy Consumption Test Plus Steady State (SMCT+) is largely the same as the SMCT. The main difference is that at the end the vehicle drives a CSC distance until the battery is discharged. It’s meant for an EV that can’t discharge to a load bank.
5-cycle test
EPA rules let automakers use the 5-cycle Test for range in addition to the other tests. This method includes the UDDS and HFEDS as well as US06, SC03, and a -7° C UDDS. The tests are run separately and used not to calculate range but to come up with an adjustment factor. That adjustment is applied to whichever two cycle test was used.
Because the five cycle tests are meant only to measure electric consumption and are run individually, they don’t start with a full battery. The battery only needs to be above 50 percent charge. The exception is the cold test, where it needs to be fully charged. Again, there are small differences between pre-conditioned and not-pre-conditioned vehicles that mostly change when the energy used to charge is measured.
An adjustment factor is calculated by dividing the 5-cycle DC power discharge total (when not using thermal conditioning) by the 2-cycle discharge. With thermal conditioning, the AC discharge is used. Multiply the uncorrected 2-cycle range by that number, and you have your 5-cycle range figure.
Simple, huh?