Posted Wednesday, May 31/06 in Driving efficiently

As I mentioned in my previous record tank entry, I used to do a fair amount of engine-off coasting in my Honda Accord. The main reason I abandoned the technique in the Firefly was because it messed up the ScanGauge, which would go to sleep when I killed the ignition (and skew its fuel economy information while it missed counting distance travelled).

As soon as I learned that a kill switch that bypassed the key kept the ScanGauge active with the engine off, I immediately set out to run a test to find out what I'd been missing.

Turns out, I'd been missing a lot.

The test ...

I mapped a loop of city/suburban driving, which I drove twice.

On the first lap, I coasted the car as much as practical (e.g. when approaching a stop, a turn, descending a hill, etc). I stopped the engine whenever I estimated I would be coasting for 10 seconds or longer. I stopped at the start/finish and recorded my average fuel consumption.

For the second lap, over exactly the same route, the engine remained on for the duration: I coasted in neutral in the same places as the previous lap. Again, when I returned to the start/finish, I recorded my fuel consumption.

The test route ...

• 9.7 km / 6 miles of city driving
  
• light/moderate mid-day traffic
  
• stops or turns every 3-5 blocks
  
• the route avoided traffic lights as much as possible (to preserve as much predictability between the two loops)
  
• 3 traffic lights
  
• 4 stop signs
  
• lots of turns at intersections
  
• several small hills (under 40 feet elevation change)
  
• 3 longer or steeper hills (one, approx 70 feet elevation.)

Conditions ...

• light drizzle; damp/wet roads
• Conditions at SUPN6 as of (2:00 pm EST) 1900 GMT on 03/10/2006:
  Wind Direction (WDIR): SW ( 220 deg true )
  Wind Speed (WSPD): 22 kts
  Wind Gust (GST): 25 kts
  Atmospheric Pressure (PRES): 29.44 in
  Air Temperature (ATMP): 45.7 F / 7.6 C
• The car was warmed up prior to the test with 1 hour of block heater use and 20 minutes of city driving. In case the drive train wasn't fully warmed up at the start (which would give the "advantage" to the second lap), I did engine-off coasting first to minimize the potential difference in the results.

Results ...

• Lap 1: engine OFF while coasting - 60.3 mpg (US) ... 3.9 L/100 km ... 72.4 mpg (Imperial); elapsed time 17:57
• Lap 2: engine idling while coasting - 53.4 mpg (US) ... 4.4 L/100 km ... 64.1 mpg (Imperial); elapsed time 17:44
• On the engine-off loop, I killed the engine 13 times; 10 clutch starts and 3 key starts
• The engine was off for 3.9 km of the 9.7 total (thanks, Google Earth!), or 40% of the distance travelled.
• Max / avg speed (idling loop): 39 mph / 17 mph
• That worked out to 12.9% better fuel economy using the engine-off technique.

Observations ...

• This is a pretty unscientific test. It's impossible to control for traffic flow, lights, and probably more importantly, experimenter bias and the placebo effect. So take these results with a grain of salt and consider it an approximation, not cold hard fact.
• That said, I believe the traffic & light patterns slightly favoured the "idling" loop. Of the 3 lights I went through, one was green on that lap, but they were all red on the "engine off" lap. Some minor traffic flow issues also helped the "idling" loop.
• Oh the irony. All that time I'd been driving leaving the engine ON while coasting... just to keep the ScanGauge alive to give me feedback so I could get better fuel economy.
• The ScanGauge does keep its normal accuracy when coasting with the engine off. I've run a few tanks through the car since re-embracing this technique, and it has stayed within 1-2% of my "manual" calculations on the back of the gas receipt. (I can adjust the ScanGauge to remove any variation, but it's not big enough to worry about.)
• Like hybrid & electric car drivers, I'm aware of the fact that the car doesn't make much noise when coasting with the engine off at low speeds. I've taken to lightly tapping the horn when coming up behind pedestrians or kids playing on or near the road, to let them know I'm there.
• Since re-adopting this technique, it's safe to say it has had the single largest impact on improving my city fuel economy. I now regularly see individual round trips of 80+ mpg (US) in suburban/urban driving, according to the ScanGauge. Results like that depend on carefully planning a route that is "coasting-friendly" (traffic and terrain), and on some of those routes the engine is OFF for over 40% of the total distance travelled.