2000 Honda Insight: 14% better MPG @ 80 km/h (50 mph), with 5 simple aero mods (ABA)

[jime57]

Originally Posted by Piwoslaw

Some cars are fine-tuned to have the external mirror inside a pocket of turbulence, or at least partially inside. Folding it allows the outer edge to hide inside that pocket.
Tuft testing would show which car/mirror combo would benefit from folding.

This is an interesting observation. I have purchased some kniting wool and plan some tuft testing on a variety of items. After reading your comment I will definitely do some testing around the mirror.

My original objective was to hang some tufts on the back window of the Insight and see what they tell me about the departure angle that Honda chose. I suspect it is fine for the stock tail length, but I want to see.

[RobertSmalls]

Sorry I'm late to the party. Metro, nice testing as always.

I just have a tip for those who aren't happy with the mirrors on the Insight. In other cars, you'd adjust the mirrors so you can just see the rear fenders of the car. In the Insight, you want to adjust so you can just see the spot where the fenders would be if the car weren't tapered. This means you won't be able to see the car in the side mirrors, but it gives you the same view of the road as you'd get in a "normal" car.

[jime57]

Darin,

I have attempted to duplicate your results with my own 2006 Insight. I don’t come anywhere close to the 14% that you measured. The list of modifications that you made do not seem to support a 14% improvement in fuel economy. I think that you must have unintentionally tripped over some sort of measurement bias, or error. I certainly do not want to be argumentative, but I think we should get this sorted out, in the interest of the science.

A logical review of what we all here have learned and measured for various modifications would also indicate that 14% is out of reach for this set of modifications. Can we review just briefly? You indicated 5 modifications:
-grill block, full upper and about ½ of lower,
-elimination of right mirror.
-replace the left side mirror with a folding mirror,
-elimination of the passenger side wiper and hatch wiper,
-elimination of the front license plate bracket.

From prior experience here, folks have reported around 3% for grill blocks, plus or minus. Full elimination of mirrors has also been reported at about 3%. Elimination of the tag bracket is probably worthless, since it sits in a stagnation area, and in any case is low enough so as not to disturb the overhood air. The passenger wiper also sits in a stagnation area and is probably worthless. Elimination of the rear wiper might provide a very small improvement, but remember that it is vertical and aligned with the air stream over the hatch.



Regarding mirrors, in this thread:

http://ecomodder.com/forum/showthrea...fied-95-2.html

you made measurements on your metro and reported results in msg #14 for a setup very similar to what you have tested on the Insight. For that experiment, you reported an improvement of 2.3%. The 2.3% is very close to the 2.4% I got by deleting both mirrors on a Toyota Echo. Best as I can recall from reading all the mirror delete tests over the last couple of years, folks have pretty much reported 2-4%. The Insight mirrors are obviously very well designed, so an improvement from deleting them both is quite likely to be 3% or less.

Grill blocks present a bit more of a estimating problem in my opinion because I think I recall some pretty widely varied reported results. In this post:

Testing grille blocking & wheel skirts: +5.7% improvement - MetroMPG.com

You reported an improvement of 2.6% for a well designed and executed test.

In this post:

http://ecomodder.com/forum/showthrea...ivic-1822.html

There was a report of about 4% improvement.

Though I installed grill blocks on my Echo, I do not recall any testing.

As I said, I have tried to duplicate the results. Not believing in short distance testing for the rather erratic and fidgety Insight, lacking a cruise control, and being complicated by Assist/Regen and lean burn, I’ve become convinced that only long range average FE at fixed average speeds hold the answers. I have been using a two way course of 48.1 miles, turn in the middle. The course is along I64 and I95 here in Richmond and is reasonably flat with no really big hills, so I can maintain lean burn about 95% of the time. I did 3 baseline runs at an average RT speed of 50MPH, per my scangauge. The baseline runs were 93, 93.7, and 93.6 MPG. There were two modification measurement runs on 10/17 resulting in FE of 98.3 & 99.4 MPG, and two modification measurement runs on 10/18 resulting in FE of 99.8 & 99.8 MPG (a slightly warmer day).

The average baseline is 93.43
The average of all modification test runs is 99.28
Improvement = 6.26%

I LOVE the improvement, but it doesn’t come close to 14%.

I did make one slight change to your setup. I retained my left side mirror since I can’t safely drive in moderate traffic without it. I’ll spot you 1% for that change. I also note that your course had little traffic, so I’ll spot you 1-2% for the added turbulence on my course.

Bottom line, I see an adjusted FE improvement of 8-9% at best on a flat, lightly traveled road.

I’d be very interest in your comments and where you think one of us may have gone wrong, if not both of us. Perhaps the real number is actually 9-10%, but I’d pretty much doubt anything over 8%

As I said, I highly respect the work you are doing and the integrity that you have given to the testing process, but I think there is something amiss here.

[MetroMPG]

I’d be very interest in your comments and where you think one of us may have gone wrong, if not both of us.

I think it's quite possible we're both right. Read on...

I don't think the data or testing was faulty. I'm satisfied that the variability of the runs wasn't horrible. ( I posted the STDEV in the data near the start of the thread. )

But I do think my results were amplified, accidentally, by the chosen load (road speed):

Originally Posted by MetroMPG

It was such a big change because the mods permitted the car to operate in lean burn 100% of the time at 80 km/h, where without them it felt like it was just on the edge, moving in & out of lean burn at the slightest change in throttle position to maintain that speed.

In other words, the % change might have been less impressive if I'd tested at 70 km/h or at 115 km/h (either fully in, or fully out of lean burn for both A and B runs).

I didn't plan it that way, but it was interesting to discover. I tested at that speed because it's the highest posted limit on the secondary roads I usually travel.

This is also key: remember, I tested with the hybrid system (IMA) intentionally shut off (worn out battery), so my car is going to give significantly different results near the "edge" of lean burn than yours will. Yours may hold lean burn longer, and as you point out, throttle movements will have different results for you if they call IMA into play.

For consistency, I don't think I'd attempt testing with IMA enabled, even if mine worked.

Hope that sheds some light.

[jime57]

Originally Posted by MetroMPG

Testing info/data added to post#2.

---

What prompted me to do this test was a pair of recent highways trips (the same trip, repeated on 2 weekends) to visit friends at a cottage.

It's 293 km (182 mi.) round trip, on secondary roads with a max posted limit of 80 km/h (50 mph). It includes passing through a number of small towns and one small city.

The first time I did the trip before the mods were all in place. My goal was simply to drive with load to keep the car in lean burn at ~100 mpg on the instant display, or higher when at the posted speed limit. I got something like 89 mpg US for the first round trip.

The second trip, in nearly identical weather conditions with all the mods in place, I used the same techniques (~100 mpg target), but I managed to do the trip in 20 minutes less! How? Because the car spent more time going faster on the highway while still maintaining at least 100 mpg indicated. And there was quite a bit more traffic (stop & go) in the city portion of the second trip.

And I ended up with 94 mpg US for the second round trip.

I'm still not convinced that either of our tests quite gells. It is true that I was out of lean burn about 5% of the time, but the fuel economy for those brief periods is still about 2/3 of the lean burn values, so that wouldn't seem to account for the rather large difference. BTW, I tested at the same speed that you tested at and my baseline speed was nearly identical to yours, and very tightly grouped - so I'm doing something right.

From the above quote, you gained 5 MPG or 5.6% on a long range test - almost identical to what I got. Did you have the IMA system turned on for those trips?

It is said by many Insight "experts" that using the assist/regen will lower fuel economy numbers, since the process is inefficient to some degree - but I haven't seen any numbers on that. I certainly think I see that inefficency but it is a small one if the assist is only occasional.

I think I kinda said it, but I'll say it again, I'm very concerned about testing an Insight over short distances because everything in the system is so twichy. All driving errors are magnified, and after all, we are trying to measure small absolute changes. If you had some sort of systematic error in the way you approached your "test mile" there could be errors. I think my main concern is that the numbers for the results just don't seem to add up, given what we all already know about the effect of those mods-about 3% for the grill blocks and about 3% for the mirrors.

In any case, I do not have the luxury of a quiet, level test course. Richmond is a rather large and busy place and I drove around a bunch to find a course that I thought worked even in the averages.

[jime57]

It is interesting that 3-wheeler has not revealed any fuel economy data taken directly for FE measurements. I met Jim last summer and talked to him at some length. Very pleasant, helpful and thoughtful gentleman. I did not bring up the measurement subject, but we had discussed it earlier on this forum. He had been a fan of roll-down testing instead, but he has found some technical GPS problems with that mode of testing.

I'm waiting anxiously for him to post some tuft testing on his beautiful boat tail. I'm not sure he considers the project finished though.

[MetroMPG]

Originally Posted by jimepting

From the above quote, you gained 5 MPG or 5.6% on a long range test - almost identical to what I got. Did you have the IMA system turned on for those trips?

No, IMA has been off since I got the car, except for a few brief test drives.

Also, I wouldn't call my longer trips a "test". It's uncontrolled anecdotal evidence - and I think it's just a coincidence it's close to your results. I didn't even attempt to drive at the same speed on the highway portions of each trip. In fact, I intentionally drove a little faster on the second trip, because I was able to hold lean burn at higher speeds with the mods in place.

In any case, I do not have the luxury of a quiet, level test course.

Where I test, it's straight, level & nearly devoid of traffic. Yet another reason we shouldn't attempt to directly compare our results.

I think perhaps the big lesson to take from this is that because these cars are so "twitchy" when near the lean burn transition, and because small inputs and other conditions will therefore have big effects on MPG, it's futile to compare different cars and expect to see the same results.

Even tested side by side in identical conditions, I'd expect two Insights to return different results when testing near the lean burn transition, for reasons such as:

- different number of miles on them
- different states of tune
- different levels of tire wear (mine are very close to the wear bars)
- using different types of fuel (no ethanol in mine at the moment)
- different drivers (foot control / lean burn coaxing)

I'm not very surprised that we're seeing different results for the same mods.

Originally Posted by jimepting

He had been a fan of roll-down testing instead,

That's definitely the best way to measure whether "A" is any different than "B". I bet if we conducted coast-down testing of these mods, the results would be more consistent, car to car.

[jime57]

OK, I'm still not satisfied and I don't consider the matter "laid to rest" but you have worn me down for the moment

One thing we agree on, these little beauties are certainly twichy and difficult to measure. On the other hand, I consider a 6% improvement to be a MAJOR improvement and am happy you pointed the way to it! I am also extremely happy as a 2 1/2 year owner to have one of the rare toys.

On a very positive note, you have rekindled my modding interest, so I'll probably go on to other mods. Thanks

[MetroMPG]

I also consider myself really lucky to have found this car (especially at a good price).

"so I'll probably go on to other mods. "

Boat tails for everyone!

[jime57]

Originally Posted by MetroMPG

I also consider myself really lucky to have found this car (especially at a good price).



Boat tails for everyone!

Yeah! But I sure would like to see some results from 3 Wheeler's work before I plunge into that large amount of work