Acceleration mini-experiment
 

(Cross-posted from https://www.fitfreak.net/forums/eco-f...ke-granny.html)

A. Slow acceleration. Shifts at 2500rpm.
B. Rapid acceleration. WOT in higher gears. Shifts around 2500 ? 3000 rpm in lower gears, and at 2500rpm in higher gears. Some wheelspin in 1st and/or 2nd gear due to the gravel road surface.
Note that cruising and deceleration were the same in both cases.

The car is a 2007 Fit LX with standard transmission.

https://img62.imageshack.us/img62/2821/courseni1.th.jpg
The course is a 7.8km (approx. 5 mile) rural block chosen for minimal conflicting traffic. (Nevertheless, one run had to be aborted due to Mennonites.) I stop at 5 points -- each corner and one additional location -- and add one additional deceleration and re-acceleration.

1. IJKL 1.7km.
   IJ - Accelerate from standstill to 80km/h and hold.
   JK - Decelerate (in gear, no throttle).
   KL - Brake to a stop (clutch / neutral).
2. LMNP 1.4km.
   LM - Accelerate from standstill to 80km/h and hold.
   MN - Decelerate (in gear, no throttle).
   NP - Brake to a stop (clutch / neutral).
3. PQRS 0.9km.
   PQ - Accelerate from standstill to 70km/h and hold. (With gentle acceleration, there is not room to reach 80km/h.)
   QR - Decelerate (in gear, no throttle).
   RS - Brake to a stop (clutch / neutral).
4. STUV 1.6km.
   ST - Accelerate from standstill to 80km/h and hold.
   TU - Decelerate (in gear, no throttle).
   UV - Brake to a stop (clutch / neutral).
5. VWX 1.5km.
   VW - Accelerate from standstill to 80km/h and hold.
   WX - Decelerate (in gear, no throttle). Final speed is around 50km/h.
6. XYZI 0.7km.
   XY - Accelerate from 50km/h to 80km/h and hold.
   YZ - Decelerate (in gear, no throttle).
   ZI - Brake to a stop (clutch / neutral).

I made three runs using each method, alternating methods to try to avoid any changes due to other factors. The runs weren't perfect; in the B cases I sometimes overshot the target speed, and was unable to shift from 1st and 2nd quickly enough to shift at the target 2500rpm. On trials 3A and 3B other traffic slowed me briefly (part of segments LM and ST respectively).

In each case I had the car off briefly while recording information prior to the initial run or from the previous run. I turned the engine on, reset the Scangauge trip counter, and drove.


Run    Initial   Max   Avg   Max   Max    Fuel
      Air Water Water Speed Speed Engine  Eff .
      ?C   ?C    ?C   km/h  km/h   RPM   L/100km
1A    10   82    84    51    82    2844   6.6
1B    14   86    84    61    83    3559   6.3
2A    14   84    84    53    81    2710   6.6
2B    13   85    84    64    86    3308   6.5
3A    13   84    83    51    79    2738   6.5
3B    10   82    83    62    85    3513   6.2


Rapid acceleration to cruising speed is more fuel-efficient than slow acceleration. Presumably the greater fraction of the drive spent in high gear at cruising speed more than makes up for the hard acceleration.

Excellent! I was wanting to do a similar experiment sometime in my car, but it looks like you beat me to it!

I'm having a hard time following your test here. Was your fuel total for the whole five mile loop? If so that a lot of varibles for 5 accelerations in such a short distance and the total burns are close. Why not just do runs of 1 mile accelerating to 80kph and holding with the different acceleration techniques and recording the results.

how much harder did u excellerate tho like 500 rpms or like 1k rpms i didnt get that

You should still do it if you get a chance. Different car, different route, and more runs would all help.

Well, a 'circular' route gets me back to the starting point without having to do 3-point turns on a public road. I don't have a dyno or a closed track, so I take what I can get. Also, the 6 segments are all slightly different, so they should help average out any aspect of a single segment that might unfairly favour one method.

In the slow method (A) I accelerated as gently as I could manage while still visibly accelerating. I got to the maximum speed not long before the deceleration point. (In fact on the long sides of the rectangle I chose the deceleration points as the next clear landmark after I reached 80km/h.)

In the rapid method (B), in the upper gears (3rd through 5th) I held at wide-open throttle until I reached the target RPM or speed. In the lower gears, I didn't get that far since I had trouble shifting quickly enough to stay near the 2500rpm target shift point (the high maximum rpm figures on the B runs come from throttle overrun while shifting).

I'd be interested in seeing a comparo of the various accel. experiments people have run. There's one at CMPG in a Civic CVT, I think SVOBoy did one last year as well.

Also, what's more valid: measuring fuel consumption over a fixed distance (including both the acceleration & cruise parts), or accelerating up to speed (at the different rates being compared) and then driving a fixed distance before recording the MPG?

I'd also like to try doing this. The more cars/results, the merrier.

I just changed by acceleration style just a few days ago in the Geo. I used to use medium-light throttle and shift from 1st to 2nd @ 7km/h, 3rd @ 30km/h, 4th @ 40km/h, and finally 5th at 50km/h.

Now I use medium-heavy throttle and shift from 1st to 2nd @ 10km/h, 3rd @ 35km/h, skip 4th and go directly into 5th @ 50km/h. The acceleration is a lot quicker and the MPG loss during the new acceleration isn't too detrimental.

Of course, it only works on dry pavement. Winter driving in snow/ice wouldn't work out too well with that technique.

https://crxmpg.com/accelmpg.html

Here's mine, *shrug*

I think there is something to the medium-heavy acceleration being better. If you take a look at my gaslog you'll notice a pretty steep increase a few tanks back. It corresponds to my change in acceleration technique. I used to accelerate in each gear to 1500 rpm's at a throttle setting of ~15 according to the SG. At the point where my FE took a leap up was when I switched to accelerating to 2000-2500 rpm's at a load of ~70-75% according to the SG. The TPS reads significantly higher with this technique. I should also note that I usually kill the engine and coast as far as possible once I reach my top/target speed.

And let's not forget the motherthread on this topic: Throttle Position During Acceleration and its effect on FE

Most of these tests are with MT? I think auto would fare a little different........ might have to do some testing with the sable when I have time.

ye aive noticed this too. at least with my truck. if i shift at 2000-2500rpm i get around 25mpg. i tried shifting less than 2K which didnt last long cuz truck bogged down easily. but i saw a big loss in my mpg (22mpg) im gonna try 3K and see what happens.

That’s weird, when I tested the slow acceleration proved better every time. I can't remember, I posted them here somewhere, but if the rapid acceleration it was like 15 mpg up to 65mph then cruising at 65 it was55 mpg. On the slowest acceleration it was like 30 mpg up to 65. The faster acceleration just could never catch up to the slow one in overall mpg’s for the run.

keep in mind mine is a truck...with a underpowered 4banger. higher rpm's might be easier on the engine or the PCM might sense that the truck is under a heavy load and try to dump more fuel in to compensate...

Brock, could that be a gas/diesel difference?

With a diesel's much flatter torque curve (and presumably far "broader" BSFC map), and lack of throttling losses (one of the things the relatively inefficient gassers overcome with this driving style that get them into a friendlier BSFC zone), I'd expect to see little or no gain driving an oil burner like this.

Can you comment on the benefit (or lack thereof) of doing P&G in a diesel? I suspect if there is any, it's much less pronounced than in a gasoline car.

I read somewhere that if you go WOT in low RPM, it is more efficient. Not too low. 1500 rpm i think. It would be less drag on the intake stroke. This would be dependant on the ECM programming ("your results may vary"), because if the fuel mixture is enrichened, then the benefit is less.

Remember in high school physics class, there is a measurment for the rate of acceleration? Miles per hour squared, i think.

Heck with the physic caluclations I used to do in school, and a scanguage, you could do graphs of MPG vs Trottle position vs RPM. Then M/H^2 vs Fuel.

Also you could do a quarter mile fuel usage testing... instead of measuring speed, you measure fuel consumption. At the "traps" be going 45 MPH every time. Test from (WOT to 45mph then cruise) to (take the whole quarter mile to accelerate to 45 MPH.) Graph it.

Just wanted to throw some idea out there. My manual transsion car is out of commision this week, I'm fixing a power steering leak. And I don't have a scanguage yet.

Mmm-hmm. I think that this is the key. Accelerating as quickly as you can w/o making the ECU put you in open-loop or in a richened state.
Ya, that would be perrrrfect.

Yes, this test, via the Throttle Position During Acceleration and its effect on FE thread, is what made me want to try this myself. In my case, though, I wanted to try it without involving any hypermiling techniques.

I suspect you are right; the car does go open loop under full throttle, but acceleration (by 'feel', anyway) does not change very much between heavy and full throttle. Given that, and that I expected from others' tests that rapid acceleration to be more efficient, I decided to use WOT to (presumably) put rapid acceleration in a worse light, i.e. to see whether it's more efficient even if the driver is not careful with the throttle.

I think the next test I'd like to do is between WOT and rapid closed-loop acceleration, provided I can manage the latter reasonably reliably, e.g. with a mechanical stop under the accelerator.

Metro yes this could very well be the difference between diesel and gas. And yes I tried two weeks of trips home running P&G every other time and got within 1% of my non-P&G trips. So I can only assume P&G doesn't add much for diesel's. Of course when warm I idle at about .25L or 8oz per hour.

Well automatics take a whole different approach,, I did some testing in the sable.
I used a 4 mile stretch with a 55 mph speed.
I drove 2 miles stopped turned around and came back 2 miles.

I don't have a bunch of individual numbers for ya, but it goes like this.
Full throttle.. runs were between 28 and 29 mpg
granny style were 31 to 32 mpg
The best was moderate throttle up to 35 mph where it shifts to OD then very slow accel to 55. This netted 34 to 35 mpg range.

This wasn't a surprise to me, thats how I normally drive.
The problem with a auto is you can't short shift it to stay at lower rpms.
For example, a full throttle run in the sable is ... 1st gear up to 45 mph 6200 rpms.... then to second gear up to 77 mph 6200 rpms. This engine won't shift at WOT until it hits the rev limiter at around 6200 rpm, trying to shift it manually does not help.