Coasting experiment: engine on VS engine off on a fixed route = 12.9% gain

[MetroMPG]

(Originally posted March/06)

Background:



I used to do a fair amount of engine-off coasting in my 89 accord. but after I got the Firefly and the ScanGauge, I let that habit wane because switching the key off to kill the motor unfortunatly also put the SG to sleep.

And unfortunately for me, my car also uses the ISO obd protocol, which takes longer to initialize (roughly 15 seconds) when switching the key back on than the newer CAN protocol. The result of that was seriously skewed fuel consumption on the SG from engine-off coasting.

So I stopped doing as much engine-off coasting.

Then krousdb alerted me to the possibility that I could install a kill switch to shut off the engine without using the key, and the SG might stay active. I added the switch - and it worked.

So it was time to find out what I had been missing by not coasting the Blackfly with the engine off. And holy smokes, I'd been missing A LOT.

The test:

2 laps (of the same route) in city driving. One I let the engine idle when coasting, the other I switched the engine off (if the coast opportunity is longer than about a block in length).

The route:

- 9.7 km of city driving
- in light/moderate traffic
- stops or turns every 3-5 blocks
- the route avoided traffic lights where possible (to preserve as much predictability between the two loops; 3 traffic lights were encountered)
- several small hills (under 40 feet)
- 3 longer or steeper hills (one, approx 70 feet elev.)
- 4 stop signs
- lots of turns at intersections

The test 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

Car was warmed up prior to test with 1 hour EBH; 20 minutes city driving (errand run)

The results:

- 60.3 mpg (US) - lap 1 "engine OFF while coasting"

- 53.4 mpg (US) - lap 2 "engine idling while coasting"

Notes & observations:

- this test is subject to experimenter influence/bias

- the traffic & light patterns slightly favoured the "idling" loop. of the 3 lights i went through, one was green for this lap, but they were all red on the "engine off" lap. some minor traffic flow issues also helped the "idling" loop.

- time to complete: 17:57 sec (engine off lap); 17:44 (idling lap)

- on the "off" loop, i killed the engine 13 times; 3 key starts and 10 clutch 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

- under these circumstances, engine-off = 12.9% better than engine on

- obviously since this is a "new" technique, I haven't yet calibrated the SG accuracy using engine-off coasting at fill-up time.

[MetroMPG]

Since running this test, of course I've been killing the engine more for coasting than I used to.

My segment mileage, according to the SG, has shot way up. example:

68.1 mpg (US) over 30.5 km of sub/exurban driving (23 mph avg speed / 42 mph max; rain, wet roads; 6 C / 46 F ambient; 1.25 hrs EBH.)

This trip was above the 70 MPG level on the way back to the house, and then I got hit by 3 red lights in a row and a train crossing at the bottom of my best hill, and these things dragged it down to 68.1

It's obvious that through using engine-off coasting, a driver would see the "hybrid FE flip flop": where non-highway driving FE is going to exceed highway-only numbers. And largely for the same reasons (hybrids shutting down the ICE at every opportunity in low speed conditions).

[Rower4VT]

How did you hook up your kill switch? I was thinking about doing the same thing. I know I would do a lot more EOC if I didn't have to use the key to do it. My plan was to mount two switches by the e-brake (easy to rest my hand right there). One switch would be in-line with the "run" circuit, to kill the engine and then turn it back on immediately so my blinkers still work and I can bump-start. The second switch would be a "press and hold" style to start the engine back up if I didn't bump-start. Any comments or suggestions? I like to use the starter to start the engine from EOC as bump-starting, no matter the gear/speed, isn't the smoothest.

[MetroMPG]

I wired a momentary interrupt switch in line between the distributor and the pickup coil. Not sure if that applies to your car or not - it's new enough it may have individual coils per spark plug.

Don't use an on/off switch for engine kill. You will occasionally forget to turn it back "on". (Ask me how I know this.)

I suspect if you're having trouble doing a smooth bump start, that you might not be doing it the best way. I really should start a new thread on this, because it comes up regularly.

[NoCO2]

Once I get my scan gauge I will definitely be interested in doing the kill switch thing to the Camry assuming I don't get a different car by then. A how-to in the DIY section would be great if you want to write one up. I know a bunch of people would appreciate it.

[boxchain]

I've been looking at rigging up a kill switch, and figured on going to the fuel injector, but I'm worried about throwing a CEL. I wanna rig up an old school footswitch like what used to be a high/low beam switch, 30-40 years ago.

Until I figure this out, I'll do my usual, turn the key. I don't turn it all the way to the next click, but I can do it just a tad (10 deg?) so the engine goes off. NEVER keep the key in anything but the regular ON position while its moving.

Bump start in high gears, 5th or 4th if going slow. Don't do anything else, just put it in gear and let out the clutch. I rarely let the clutch out all the way, push it back in when you start to feel a loss in momentum then rev the engine to match your speed or switch gears. Then let out the clutch all the way.

[MetroMPG]

Posted: How to smoothly bump start / clutch start an engine