Tested: The Most Efficient Way To Accelerate With An Automatic (Toyota Matrix)

[JRMichler]

I tested my truck the same way and got similar results with a manual transmission. Now I just step on it, shift at 2500 RPM, and go. That gives about the same acceleration as most other people around here.

[MetroMPG]

Originally Posted by mechman600

Believe me, 60 km/h in 740m is very slow. Painfully slow. Like old man with a beard driving a Volvo slow.

Or from another perspective, that's the ForkenSwift accelerating flat out. OK, maybe the next point on the graph. Or maybe not...

Thanks for posting this!

[E.Roy]

Thanks, I think this will be applicable to my 2000 Corolla 1.8 4spd auto

[Arragonis]

I disagree, these figures are interesting but not very scientific - more runs would not make this better. The problem is the SG2.

The SG2 reads data from the car via the OBD interface - this does not include MPG or average MPG.

The SG2 calculates these by sampling values from the OBD interface - things like road speed, injector timing and dwell - which combined with engine size and the owner's adjustments sort of works out to MPG.

These sampling calls are based on time - when you do the slow runs the SG2 makes more calls and probably makes a more accurate estimate of MPG. On the faster runs fewer calls are made and therefore than result is less accurate with reality.

You can adjust this via the interface speed setting but I think that just means the SG2 should ask for data more often.

I've seen this myself in my previous Aygo. When I drove round town (usually between 0-25 MPH) I could get the SG2 to match my tank refills more or less exactly. When I drove at higher speeds (e.g. my 200+ mile trips to England this summer at 65-75 MPH) the SG2 would be up to 15% out. What was interesting was that it would be both under and over by 0-15%, not always under or over.

I think that at 0-25 the SG2 sampled the engine enough to make an accurate guess at what my MPG was. At 70 the same sample rate could miss a lot of "detail" so the calculation would get far less accurate - e.g. it could miss me coasting faster and then using DFCO and count me accelerating uphill slower more often so the average would fall.

[mechman600]

Could be. Maybe we need someone with an MPGuino (more accurate, I presume) to perform the same test?

[christofoo]

Originally Posted by Arragonis

I disagree, these figures are interesting but not very scientific - more runs would not make this better. The problem is the SG2.

The SG2 reads data from the car via the OBD interface - this does not include MPG or average MPG.

The SG2 calculates these by sampling values from the OBD interface - things like road speed, injector timing and dwell - which combined with engine size and the owner's adjustments sort of works out to MPG.

These sampling calls are based on time - when you do the slow runs the SG2 makes more calls and probably makes a more accurate estimate of MPG. On the faster runs fewer calls are made and therefore than result is less accurate with reality.

You can adjust this via the interface speed setting but I think that just means the SG2 should ask for data more often.

I've seen this myself in my previous Aygo. When I drove round town (usually between 0-25 MPH) I could get the SG2 to match my tank refills more or less exactly. When I drove at higher speeds (e.g. my 200+ mile trips to England this summer at 65-75 MPH) the SG2 would be up to 15% out. What was interesting was that it would be both under and over by 0-15%, not always under or over.

I think that at 0-25 the SG2 sampled the engine enough to make an accurate guess at what my MPG was. At 70 the same sample rate could miss a lot of "detail" so the calculation would get far less accurate - e.g. it could miss me coasting faster and then using DFCO and count me accelerating uphill slower more often so the average would fall.

Am I wrong in believing the OBD-based-MPG-gauges are calculating fuel rate primarily from the MAF sensor? OBD-II PIDs - Wikipedia, the free encyclopedia

This would be particularly a concern with acceleration testing since it will underestimate fuel-consumption if you get into WOT, unless the driver ensures that is stays closed-loop, but even then I'm suspicious MAF isn't linear across the full BHP range.

My SG2 'calibration' has drifted upwards from +11% to +30% over the course of 30,000 miles or so that I've had it. Apparently I have a sensor that is aging and yet the MPG hasn't changed noticeably.

[elhigh]

Originally Posted by mechman600

Like old man with a beard driving a Volvo slow.

Hey, now.

Oh wait, the Volvo's gone.

[MetroMPG]

Originally Posted by mechman600

Could be. Maybe we need someone with an MPGuino (more accurate, I presume) to perform the same test?

I suppose MPGuino would be best: it counts injector pulses, where the SG is extrapolating fuel consumption from air consumption.

[riva2model64]

Great write up man. I actually hear a lot of stories from people online (and even real life) going from soft to brisk acceleration and seeing an MPG increase.
Even though the test didn't reflect that hearsay very much, but it also didn't reflect the common belief that slower acceleration is more fuel efficient. The acceleration efficiency also probably has a lot to do with the vehicle being tested.

Even so, it's interesting to note that there wasn't an FE difference greater than 7% regardless of the acceleration method.

Which leads to another question . . which acceleration method is best for the life of the vehicle (least wear and tear)? It affects MPV (miles per vehicle) heheh. .

[gone-ot]

Originally Posted by E.Roy

Thanks, I think this will be applicable to my 2000 Corolla 1.8 4spd auto

...likewise applicable to 2003-2008 Pontiac Vibes with 1.8L engine & 4-speed auto.