Fast or Slow Acceleration? Test Results
 

As a new aspiring hypermiler, I have been doing a lot of reading online. Every now and then I come across a post where someone refers to a recommendation by BMW suggesting that brisk acceleration up to speed will give better overall fuel efficiency that slow acceleration.

I just wanted to share the results of some informal testing that I did with my new Civic EX. My conclusion: While the numbers may not be extremely precise, I think the trend is fairly clear (and predictable). Slower acceleration over a longer distance will save you more fuel than quick acceleration over a shorter distance. At least it does for my Civic with CVT.

You can see my methods and results at Acceleration Experiments | Civic MPG. Sorry, can't make this a link since I am a newbie.

--Chris

So w/ the CVT at the slow, med, and high (1.8K, 3.3K, and 5+K RPM), it was held at said RPM the whole time? Because I think that might be what is killing your MPG far more than throttle position itself!

The CVT is a different beast than traditional transmissions. It is capable of loading the engine before increasing engine rpm which is a big efficiency benefit. Automatics increase rpm (and load) as throttle is increased. Manuals increase load as throttle is increased.

Still, good testing, thanks for doing it!

It would be interesting to see what load the engine is at at those rpm ranges.

So this means that with a CVT if you throttle it higher, the engine will first increase load at a relatively steady slowly increasing RPM and then depending on power demand indicated by throttle position, increase RPM by shifting gears if necessary? If that's true we really do need a guaged "engine load" reading rather than the butt dyno, no? Maybe a throttle position reading too?

I know that is how it works on the Prius with its e-cvt. As soon as you're off idle rpms, your engine is optimally loaded.

I'm not exactly sure how OEMs are programming normal CVTs. But, theoretically, that is how they can work.

I do have a bluetooth ODB2 reader, so I should be able to get throttle position from that I am guessing. I will have to check it out. I also read about vacuum gauges for measuring load. Does anyone know if this would be available through ODB2?

I enjoyed doing this little experiment and am planning to do more as time permits. I will include some load and / or throttle position experiments.

Yes, I did try my best to keep the RPMs close to those levels. As mentioned, CVTs have a different behavior that is new to me. I can keep the RPMs steady and slowly the speed will increase.

Thanks for putting that link in there Doax!

Nuverde, you don't have an Ultragauge? This would be a great and cheap way for you to get great data, such as load and a ton of other things, including reading trouble codes.

Wouldn't the same data be available through a bluetooth ODB2 reader and a smartphone? I just ask since I already have both of these. I haven't used them on this car though. Just on previous vehicles for error codes.

Maybe. I don't know what is on the items you refer to. But do they offer calculated engine load? Calculated instant MPG, trip MPG, and tank MPG? Engine load is the great critical calculated reading for access to the best hypermiling strategy "DWL" or "driving with load."

In a CVT when you give it gas the engine will basically pick a certain RPM and it will stay there until you change the throttle position. In my Altima I tested for acceleration consumption, but I stopped measuring at the same point on the road. Faster acceleration got me to speed quicker which allowed just about the same average MPG to the same spot in the road. Of course more rapid acceleration got me to speed faster, while slower acceleration took longer. Bottom line was the MPG was just about the same regardless of the throttle position until you really accelerated quickly which cost MPG, measured to the same point.

Our 2006 Murano would accelerate to 70 MPH without the RPM changing at 1700, with a constant throttle position.

regards
Mech

On my old Civic, Scangauge told me that 75 engine LOD was roughly equal to 12.5 MAP vacuum.

On the Fit (and probably the new Civic), it's not so simple as there are electronics in between you and the actual throttle.

Everyone is talking about CVTs and whatever but I think you're all missing the elephant in the room. Cruising 0.6 miles at 65mph takes WAY more energy and that's not negligible. Someone on Spyderchat did a test with different shift points and a 2ZZ-GE engine and concluded that the slowest acceleration uses the least fuel as well, but that is hardly an indicator of what engine load is most efficient.

That is a very good point, it would be interesting to look at fuel used just to reach target speed as well (i.e. merging/p&g), that would have a bias towards faster acceleration, and you can just EOC out the rest of the distance if time doesn't matter anyway (sans CVT).

Indeed, the point of my little experiment was not to pinpoint the most mpg efficient load/mpg/gear ratio/other. The point was for me to "discover" how I should be accelerating up to (and maintaining) my cruising speed using the least amount of fuel. In the end, isn't that the goal: to use the least amount of fuel to traverse a certain distance?

My finding was that for my car, accelerating slowly up to 65mph and cruising for a short amount of time until I reached the one mile mark used LESS FUEL than accelerating quickly and then spending more time cruising at 65mph (until the one mile mark).

Now I concede that more experimentation could be done to determine the sweet-spot in the low acceleration range... maybe someday.

I think you are assuming slowly merging onto an empty hiway with a long on-ramp and then maintaining a speed. Around town the best technique is to use your engine efficiently (if the transmission allows it) and avoiding the brakes (don't over-accelerate, coast to stops). You also have options if the hiway is mostly empty, such as drive slower and pulse and glide, drive with load, etc.

Also there appears to be some issues with the accuracy of the imid
What is your Civic SI's best MPG? | 9th Generation Honda Civic Forum
It may not be recording fuel consumption directly, unfortunately (indeed since you are inferring fuel used from displayed mpg...)

Experienced hypermilers understand the compromises involved. I've p&gd a beater 5 speed all over the state to the tune of 55mpg. As well as route selection becomes a factor also, i.e. efficiently pulse and glide on surface streets, or a 65mph burn on the hiway, might only make a 4 minute difference in commute and cut your mileage in half.


It's complicated... My prius (arguably a cvt) loses a lot of efficiency in the transmission if I go over 2400 rpm at lower speeds, but my 5 speed civic doesnt.

Nice testing.
I have a question about your procedure. When you say 1 mile, is that 1 mile from the 0MPH starting point or 1 mile from where it hit 65MPH?

Thanks!

That is one mile from the 0mph start. So the distance covered was 1 mile for all test runs, regardless of acceleration rate.

It would be interesting to accelerate (1) with different % engine loads to the same average speed, plus (2) coasting back down to 0 mph and comparing the results.

You're going to get better mpg with slower acceleration through the pulse, but a faster acceleration will give a better ratio of pulse to coasting. For example, A accelerates to 60 mph in .25 miles, and coasts for .75 miles, whereas B accelerates to 60 mph in .75 miles, and coasts for .75 miles.

A is accelerating 1/4 of the time, and B accelerates for 1/2 of the time. Thus B gets better mpg for the pulse, but A gets better mpg for each complete cycle, due to a greater percentage of coasting. This doesn't apply to keeping the same speed the whole time.

What you've actually discovered is that driving at a lower average speed burns less fuel. That was never in debate.

I demand a retest!!! :eek: Can you add 2 entries? 1 for the rpm below vtec engagement and just above it? Vtec engagement is 2500 rpms+- a few conditions, right?

Run Data:

Acceleration Run 1 Run 2 Run 3 Run 4 Avg. MPG Avg. Gallons Used
Slow (~1800rpm) 28.6 27 28.4 26.6 27.65 0.036
Med. (~3300rpm) 24.9 23.2 25.2 24.43 0.041
Fast (5000+ rpm) 19.9 21.3 21.1 20.77 0.048

Actually, the more I learn, the more I am convinced that trying to isolate a single variable for testing is beyond my patience. The engine in this civic (R18a) has a vtec that is kinda backwards from what we are used to thinking. In short, between 1000 and 3500 rpm, low cams may be engaged, IF the conditions are right for fuel economy. See http://asia.vtec.net/Engines/RiVTEC/ for lots of details. Cool stuff.

Between that, the CVT and ECON mode, the car is doing a ton of mpg optimization in real time. This car is made for MPG. If you want a 'performance' Civic, you get the Si, with traditional vtec.

The insights and I believe HCH the switch over point is 2500 rpms. Thats assuming the car is warmed up and its not floored. Otherwise it will engage it at lower rpms. I think for the SI its much higher like 4800-5700 rpms.

It seemed you got a testing method, so I didnt think it would be any trouble to test that as its something I wondered if using the lower end of vtec or the upper end of the stock grind gave better mpg. You know cam profile 1 vs 2.

I would argue that 1800rpm isn't 'slow' acceleration, that's actually probably around BSFC. The argument for accelerating 'fast' isn't to rev to 5000+ RPM, its' actually to use high loads to accelerate.

We know more RPM will use more fuel.

The real issue is shifting at the same RPM but varying the throttle opening. I've always found heavy foot low rpm (~2000) to work the best (carb, EFI, petrol, diesel, NA, or turbo it's always worked). Most people trying to save fuel are afraid of using the accelerator when they shouldn't. Although light throttle doesn't 'waste' fuel, if you can move faster for no economy penalty you might as well. I think this point would bring hypermilling to the interest of a lot more people too.