[Infidel]

I've only had my 06 insight for 2 weeks now, but i've got 1100 miles so far, averaging about 73.0 mpg (thanks to the advice from this forum).

Since there seems to be a conflicing information as far as what's the most FE way to accelerate (NO IMA, vs WOT), I decided to run the following experiment:

Theres a 1.8 mile stretch that is mostly flat near my house. The test run is as follows:

- Reset the segment FCD
- Acellerate to 50mph with varying techniques (depending on data run)
- Hold 50mph using my best lean-burn hold technique
- after 1.8 miles, use max regenerative braking in 5th until it cuts off
- downshift to 2nd, use max regenerative bracking till it cuts off
- quickly make a u-turn, stop, and reapeat back to the starting location. (3.8 miles total)
- record segment FE data, make u-turn, and repeat test.

For some of my runs, I did this with a FULL SoC. (Used regenerative braking untill it refused to charge). Runs labeled NOT Full SoC were done by doing 2 WOT accelerations to 60 to deplete the battery a bit (trying to simulate "real world" vs. "ideal" conditions. Here are the experiments (and results):

Run 1: Using very little acceleration, pretty much the "No IMA" approach. As soon as I hit any assist, I would immediatly back off. I upshifted when the up shift light came on. This was done with a FULL SoC. Result: 75.4mpg over 3.8 miles.

Run 2: Using light acceleration (4-6 bars of Assist). Shifting at around 3k RPM. This was also done with a FULL SoC. Result: 71.6mpg.

Run 3: Using WOT acceleration. Shifting at 4k rpm. Also with FULL SoC. Result: 75.6mpg.

Run 4: Using "No IMA" acceleration. Same shift points as Run1, with 2-bars off a full SoC. Result: 69.0mpg

Run 5: Using WOT acceleration. Sifting at 4k rpm, running 2-bars off a full SoC. Result: 74.4 mpg.

This data seems to imply that WOT acceleration is the best way to go. This makes sense to me, given the P&G discussions I have been reading. I kind of wonder if the WOT approach would be more effective if the shiftpoint was even higher, around 5.5k-6k rpm.

One point to note, however, run 1 was noticably easier to hold lean-burn. I'm assuming this is because background charging wasn't in effect.

I'm now starting to suspect that the #1 factor in maximizing FE is getting into lean-burn as quickly as possible, and holding it for as long as possible, period. All other factors are at least second or third order factors. Also, transitioning from WOT to lean-burn efficiently is *tricky*. It is VERY easy to accelerate to 50mph, and by the time you manage to coax the motor to enter lean-burn, you are down to 40-45mph. If this happens, you are screwed FE-wise.

My technique for transitioning from WOT to lean burn is as follows, feather the throttle slowly easing up until you eventually enter lean-burn. If this takes an extra 5-10 seconds at higher fuel consumption, fine, but the key is to maintain your momentum until you enter lean-burn, because accelerating in lean-burn is very difficult without a downhill slope and/or tailwind to help you.

In any case, I don't know how much of this is "common knowledge", but I found that this data is a bit contradictory to the knowedge base (which encourages WOT acceleration at 2.5-3k rpm shifts) and contradicts the no IMA approach, which only seems viable with a Full SoC, and even then, results in frustrating acceleration.

I've found that WOT acceleration at lower shift points causes you to spend too much time out of lean-burn, sacrificing FE (for tha same reasons that P&G is effective).

I plan on repeating this experiment in the near future with the WOT approach and verying my shift points. I post results here when I do. I hope you guys found this as interesting as I did.

 

[JoeMultihuller]

Hi Infidel, welcome to the Forum and congratulations on your impressive initial lmpg. Thank you very much for sharing your testing with us, and I look forward to your future tests varying your shift rpm.

Inasmuch as the goal is to determine the effects upon FE from varying acceleration techniques, I can't help wondering if acceleration/mpg tests for a fixed distance without stopping wouldn't perhaps result in cleaner data. Remember that you can quickly update your FCD mpg display by hitting the button twice (otherwise, it only updates once a minute) to get your reading at the moment you cross your 1.8 mile distance. As you'll discover, the mileage gains while coasting can be significant and deceleration inconsistency may be contaminating your data. I look forward to future postings from you. JoeS.

 

[sno779]

I agree with Joe. I think you need at least three tests out and three tests back without the u-turn to clean up the data. My old 2000 even with replacement bateries doesn't ever charge up to the top. Personally I use the 4-5 bar assist technique through 3rd, skipping 4th, and accelerating slightly over the speed limit, (2-3 mph) put it in 5th and coasting down to the speed limit to set the lean burn. I skip 4th up to 45 mph speed limit then use them all over that. Enjoy your new car.....Louis

 

[Guest]

Hi Infidel:

___My experience was a bit different then yours? I would bring her up w/out assist in 1, 2, and grab lean burn from 3 on up while still accelerating without assist as per my usual. I do not remember seeing less then 70 mpg on the FCD from dead cold over a 1 mile distance except when it was below 30 degrees F. The best I remember seeing using the above acceleration technique from dead cold to 50 mph was 102 something after 1.5 miles with little to no elevation assist and all on E10. That one was in the heat of summer of course … Besides having a pack that for all intents and purposes was brand new when I sold her, the packs longevity was surely not an issue using the above technique. I hope that helps?

___Good Luck

___Wayne R. Gerdes

 

[Rick Reece]

In your WOT tests you are not gettting a true mpg value unless you account for the energy that must be put back into the battery. Thus with a short test track you are leaving the battery at a lower level than when you started. If you continue this long term the car would need to charge the battery back up thus making it harder to maintain lean burn since the short test track does not allow enough time for the background charging to replace the energy expended.
If you wish to do more experiments I would suggest a 25 mile test loop with perhaps 5-10 0-50 segments. Then between tests do a reset of the battery so that the SOC of one run does not effect the SOC of a following run. As you have already fouynd the key to above average fuel economy is maintaining lean burn as much of the time as possible. Unfortunately background charging will make this harder so with a non-MIMA equipped Insight it usually best to minimize usage of the IMA for best fuel economy. (MIMA allows you to control background charging and thus not impact the lean burn. Note you will still have to replace energy taken from the battery but you can choose how and when such as comming to a stop using MIMA vice the brakes although do check that no one is following too close).
Have fun, RIck

 

[Infidel]

Well, the weekend before last I tried some more experiments. Unfortunately, the results were fairly inconclusive. My test run this time was as follows:

All tests started with 2-5 bars off a full SOC.
- Reset FCD
- Accelerate to 50 mph, varying shift points from 2500 to 5500 rpm
- Once at 50, maintain lean-burn until 1.0 miles.
- At 1 mile, double tap FCD and record mpg.

I did several runs at each shift point. (2500, 4000, & 5500), in both directions, and about 2-3 runs per direction, per shift point. The results were all over the map. I literally looked like a random distribution from 67mpg all the way up to 77mpg.

I agree with Rick. The test loop is just too short to get any meaningful data out of it. I think the only method that is even close to statistically relevant is to reset my FCD for my commute to and from work every morning and log that info (along with my technique). My commute is about 90 miles round trip, consists of about 60 miles of freeway driving, with about 20 stops total on the street segment.

Earlier last week, I tried this route using the "No IMA" approach. I usually averaged between 75-78 mpg for the 90 mile segment. This week, I'm trying the Full assist, low rpm approach. I've been shifting between 1500 and 2000 rpm, acellerating at just enough throttle to give me 75-100% of full assist. (basically eyeing the FCD, and trying to shoot from the hip regarding where the point of diminishing returns is in accel/vs more assist/vs less mpg). I never accelerate at less than 25mpg. Acellerating at 10-15mpg (WOT in second) is clearly not twice as fast as 25mph acceleration w/assist.

On monday I got 80.2 mpg for the 90 mile loop.
Today, I got 81.5mpg for the same loop. (including some extra stop & gos in traffic this morning). I also added the radiator block today.

Based on what I've seen to date in the last 3 weeks, it really feels like the car has different efficiency acceleration "bands." I.e, in second gear, accelerating at 10 mpg is less than twice as quick as accelrating at 25mpg, but accerating at 40mpg is more than twice as slow as 25mpg.

However, the efficiency bands change based on what gear you are in, and what speed you are moving at.

I think the key to cracking this problem is by datalogging acceleration data at a particular gear and mpg burn rate. Doing anything else leaves the door open for misinterpretation due to the car's senstivity to environment and the lean-burn window.

Anyone know a cheap accelerometer that can datalog to a PC? Is a G-Tech pretty accurate? Ideally, I'd like to start at a rolling start (<5mph) and record the time from something like 10-30mph or so. That way, we can normalize out most of the environmental factors and make this more objective....

 

[Infidel]

Hi Infidel:

___My experience was a bit different then yours? I would bring her up w/out assist in 1, 2, and grab lean burn from 3 on up while still accelerating without assist as per my usual. I do not remember seeing less then 70 mpg on the FCD from dead cold over a 1 mile distance except when it was below 30 degrees F. The best I remember seeing using the above acceleration technique from dead cold to 50 mph was 102 something after 1.5 miles with little to no elevation assist and all on E10. That one was in the heat of summer of course … Besides having a pack that for all intents and purposes was brand new when I sold her, the packs longevity was surely not an issue using the above technique. I hope that helps?

___Good Luck

___Wayne R. Gerdes

Well, its just that looking at just the first mile or two seems to be very misleading, and highly suceptable to how the lean-burn segment is managed.

My experience with trying lean-burn acceleration yeilded fairly mediocre results (around 70-75 mpg city-ish driving). I think better performance can be had by a modified assist technique. The only evidence I currently have for this is today during my trip home, I went through a 5 mile segment during rush hour that had stop lights every .5 to 1 mile. With near-full assist in 2nd and 3rd, then lean-burn coast in 4th and 5th, I was able to actually increase my segment mpg from 80.3 to 82.5 during the 5 mile peice. I had about 60 miles on that segment at that point.

 

[Subterfuge3927]

Long story VERY short, ANY assist is reducing your max possible FE.

 

[Infidel]

Well, I'm finding that as I'm getting more experience with how to drive the insight efficiently, my technique is constantly changing, However:

I still think that background charging (or rather, supressing it) is the only reason to NOT use assist. Also, on my insight, background charging certainly appears to turn off before regenerative braking does.

So, unless my observations are false (which can be proven or debunked by anyone with MIMA installed), greater FE should always be able to be achieved on a stock insight by using assist during a full SoC as long as using such assist does not activate background charging.

I.e, I claim you lose FE whenever you attempt to use regenerative braking and the ECU blows you off.

 

[Subterfuge3927]

nope, the very mechanism in the ECM that engages the assist does so based on the overload of the gasoline engine, and its inability to accelerate at the demand of throttle. In otherwords, dude, your ridding your throttle too hard! (burning too much dino power!)