WAI vs CAI test results

[MetroMPG]

i'll post the details later, but the summary is:

CAI (performed 5 bi-directional runs)
54.3 F avg air intake temp
79.12 km/gal (US) avg consumption

WAI (performed 4 bi-directional runs)
106.7 F avg air intake temp
79.31 km/gal (US) avg consumption

it's always disappointing to put this much work into something and not see any significant results.

in case you're wondering, that's a .2% (point two percent) difference of WAI over CAI. i haven't calculated margin of error yet, but i suspect it's within it, which would make the difference statistically insignificant.

details and methodology to come later on...

[Matt Timion]

That is disappointing. If you ever need help with calculating your margin of error, etc just send me the numbers and I'll run them for you really fast.

[DaX]

I honestly believe this is the ECU correcting for the warm air. This makes me want to hurry up and throw my Hondata in. One thing at a time though...

[SVOboy]

I honestly believe this is the ECU correcting for the warm air.

Honda ECUs do not do this, they lean the mixture based on hot air, though I'm not knowing much about obd2 gm stuff.

[MetroMPG]

but the ECU may also be controlling the timing advance based on intake air temp. in which case the gains in air temp are erased by less advance.

i can only view "instant" timing advance on the scangauge, though, so i can't really make any comment on what it was doing on average.

btw, for the CAI data:

0.18 = standard error (using the formula: (STDEV(F13:F21))/(SQRT(COUNT(F13:F21))) where F13:F21 are the cells containing the individual data for each run)

correct me if i'm wrong, matt, but standard deviation (a.k.a margin of error) is 2x standard error, or .36% in this case. which confirms my prediction that the .2% difference seen is statistically insignificant.

[Matt Timion]

You're better off just posting the results from all four runs if you have it. Post the miles and the MPG. We'll run a t-test or even a two way anova to determine if it was statistically significant. I can post all of the relevant information as well.

[SVOboy]

On the ECU topic, does obd2 correct timing for it? That would be seriously sucky!

[MetroMPG]

* figures are average of bi-directional runs (cancels effect of wind/grade)

* "test" road: is a nearly level, 6.6km stretch (13.2km round trip)

* speed was set at 89 km/h - cruise control was set once, cancelled with the brake pedal between runs; "resume" feature used for subsequent runs (cruise control consistency verfied by "avg speed" feature on scangauge)

* car was up to steady speed on cruise control when passing a fixed "start" or "stop" marker; scangauge "avg" logging was either reset (at "start") or recorded (at "stop" marker)

CAI (units: km/gal)

78.5
79.05
79.4
79.4
79.25

WAI:

78.9
80.2
78.35
79.8

[MetroMPG]

On the ECU topic, does obd2 correct timing for it?

yes, the level of timing advance is changing constantly, but i don't know what sensors it's using for input for the control.

it's a good point... when you get down to this kind of tinkering, you really need to know the precise details of how the ecu is going to respond.

it makse sense that temperature would factor in as one of the timing controls (but whether it's using coolant or air intake temp, or both, i don't know). it's certainly using throttle position in the formula.

throttle position also changes based on air temp. the hotter the intake air, the wider the throttle has to open to continue producing the same amount of power (thus reducing throttling or pumping losses - one of the several theoretical reasons WAI is more efficient). but the TPS is also a factor the ECU uses to determine a/f mixture and timing advance - a wider throttle would normally change both a/f and timing for the worse, mpg-wise.

it's entirely possible that there's a "sweet spot" intake temp which balances out the pros and cons of all these variables, and i just didn't find it today.

the only way to find it would be spending hours and hours on the road (this took me 3 hrs) testing across range of different IA temps... or simply having the proper info about how the ECU will respond to sensor input.

[Matt Timion]

<p align="center"><strong>Group Statistics</strong></p>
<table width="100%" border="1">
<tr>
<td>GROUP</td>
<td>N</td>
<td>Mean</td>
<td>Std. Deviation</td>
<td>Std. Error Mean</td>
</tr>
<tr>
<td>WAI</td>
<td>4</td>
<td>79.3125</td>
<td>.84101</td>
<td bgcolor="#FFFF00">.42050</td>
</tr>
<tr>
<td>CAI</td>
<td>5</td>
<td>79.1200</td>
<td>.37517</td>
<td bgcolor="#FFFF00">.16778</td>
</tr>
</table>
<p align="center"> <strong>Independent Samples Test</strong></p>
<table width="100%" border="1">
<tr>
<td rowspan="3">&nbsp;</td>
<td colspan="2"><div align="center">Levene's Test for Equality of Variances</div></td>
<td colspan="7"><div align="center">t-test for Equality of Means </div></td>
</tr>
<tr>
<td rowspan="2">F</td>
<td rowspan="2">Sig.</td>
<td rowspan="2">t</td>
<td rowspan="2">df </td>
<td rowspan="2"><div align="center">Sig. (2-tailed)</div></td>
<td rowspan="2"><div align="center">Mean Difference</div></td>
<td rowspan="2"><div align="center">Std. Error Difference</div></td>
<td colspan="2"><div align="center">95% Confidence Interval of the Difference </div></td>
</tr>
<tr>
<td><div align="center">Lower</div></td>
<td><div align="center">Upper</div></td>
</tr>
<tr>
<td>Equal variances assumed</td>
<td>6.372</td>
<td>.040</td>
<td>-.463</td>
<td>7 </td>
<td>.657</td>
<td bgcolor="#FFFF00">-.1925</td>
<td>.41545</td>
<td>-1.17488</td>
<td>.78988</td>
</tr>
<tr>
<td>Equal variances not assumed</td>
<td>&nbsp;</td>
<td>&nbsp;</td>
<td>-.425</td>
<td>3.956</td>
<td bgcolor="#FFFF00">.693</td>
<td>-.1925</td>
<td>.45274</td>
<td>-1.45504</td>
<td>1.07004 </td>
</tr>
</table>
<p><br>
<br>
<br>
Explanation:</p>
<p>In order for something to be statistically significant at the .05 level (that's what psychology uses, medical is often .01) the Mean Difference must be greater than 2 standard deviations of the error. In other words, error is always introduced into an experiment. The two numbers in yellow in the first table are the error. In order for two groups to be statistically different the difference of the means has to be more than TWO SEMs (std. error of the mean). In both cases the mean difference (.1925) was no where near 2 SEMs for either group.</p>
<p>The conclusion is that on MetroMPG's Firefly, there is only a 30.7% chance that the increase of gas mileage was due to the warm air intake and not something else.</p>
<p><br>
<br>
<br>
<br>
</p>

[SVOboy]

Hmm, all I can say as a starting point is that the older honda ecus do not do that, I will poke around for you though.

[MetroMPG]

more notes on methodology & test conditions:

* an A-B-A-B test method was used: 3 CAI runs; then 3 WAI runs; then 2 more CAI; then one final WAI

* when switching between A and B, the car was driven for 10 minutes to allow the intake tract to either cool down or warm up to a steady temperature before returning to the "test course"

* i used about 6.9L of fuel during the test, so the car got lighter by that amount from run 1 through run 9 (plus the cool down/warm up drives_.

* the ambient conditions were as follows, based on info from an online weather station (http://www.ndbc.noaa.gov/station_page.php?station=supn6) located several km from the "test road" location.

time....wind-dir...wind...gust...temp-F

12pm......ssw.......10.....10......39
01pm......ssw........8......8......38.1
02pm......ssw........7......7......39.6

(wind data in knots; the test went from approx noon to 2pm.)

* the car was warmed up prior to the test: block heater 1.5 hrs & 30 minutes of mixed city/hwy driving

[MetroMPG]

Hmm, all I can say as a starting point is that the older honda ecus do not do that

sorry, don't do what, specifically?

matt: cool analysis. where do i get me some of that program?? i was just using the STDEV function in excel and a formula I found on a stats web site. it's close enough to the figure your program gave you for std error and std deviation for the CAI group.

[SVOboy]

Adjust timing based on anything besides knock, and then they don't even do that unless they are japanese.

[MetroMPG]

not even vacuum advance? surely timing is controlled by something dynamically.

my car doesn't have a knock sensor. and i'm running enough base timing advance that i have to be careful not to mat it at low rpm or it will ping.

[rh77]

With the WAI -- do you have a way to get the actual temperature of the air? I'd venture a guess that cold air is slipping into the mix somehow.

To answer the OBD2 question, my '98 Honda (Acura) ECU leans-out the mixture with the detection of warmer air, as shown from a datalogger. With the CAI on there, I was getting an average of -2 to -3 long-term fuel trim, whereas the WAI ranges in the -4 to -6 range. Economy was instantly noted to increase from around 26 to 31. Timing advance is all over the place, so I can't really report any changes in that department. There is a definite decrease in horsepower, especially at higher RPMs, as shown in 0-60 times. Merging onto the highway, if I have to floor it, seems like it's choked-off struggling to get to redline. Have you noticed a decrease in power?

RH77

[SVOboy]

Ah, well I took vac advance for granted. Timing is on a big table with the variables being load levels and rpms.

[Matt Timion]

Hmm, all I can say as a starting point is that the older honda ecus do not do that

sorry, don't do what, specifically?

matt: cool analysis. where do i get me some of that program?? i was just using the STDEV function in excel and a formula I found on a stats web site. it's close enough to the figure your program gave you for std error and std deviation for the CAI group.

I used SPSS to get my results. I was trained on SPSS and some other old powerful command line program that everyone still uses (can't remember the name now).

I might be able to swing you a copy of SPSS. I think I have one around here somewhere. I think it is generally used for behavioral research but any statistics can be run using it.

I was originally planning on programming all of this into the website and letting users simply input data and then have teh website spew out the results. I'm sure I could still do that, although it would take a while to do.

[Compaq888]

With the WAI -- do you have a way to get the actual temperature of the air? I'd venture a guess that cold air is slipping into the mix somehow.

To answer the OBD2 question, my '98 Honda (Acura) ECU leans-out the mixture with the detection of warmer air, as shown from a datalogger. With the CAI on there, I was getting an average of -2 to -3 long-term fuel trim, whereas the WAI ranges in the -4 to -6 range. Economy was instantly noted to increase from around 26 to 31. Timing advance is all over the place, so I can't really report any changes in that department. There is a definite decrease in horsepower, especially at higher RPMs, as shown in 0-60 times. Merging onto the highway, if I have to floor it, seems like it's choked-off struggling to get to redline. Have you noticed a decrease in power?

RH77

I have to quote on this. i used to have a WAI and I will tell you I lost a lot of low end with it. I didn't gain any mpg either. All that happened was I gained a lot of top end. I had an Injen WAI. My mpg stayed the same. The best mpg I'm getting is with the stock intake system. As for the top end I lost .4 in the 1/4 mile thanks to the Injen WAI. that is not .04, but .4, which is almost half a second.

[MetroMPG]

With the WAI -- do you have a way to get the actual temperature of the air?

yup, from the air intake temp sensor (inside the throttle body i think) - via scangauge.

To answer the OBD2 question, my '98 Honda (Acura) ECU leans-out the mixture with the detection of warmer air, as shown from a datalogger.

yeah, i'm starting to realize that to go really hardcore, a scangauge doesn't cut it. i need something to hook to the laptop to gather data for experiments, and the scangauge for normal driving. SG doesn't report fuel trim, so i can't report on those figures.

Have you noticed a decrease in power?

my car feels gutless at the best of times :)

i haven't stuck my foot in it to see if i notice any difference. this car doesn't encourage (or reward) that kind of driving. though i may have felt it bog down a little the other day when the AIT showed that it was really warm.

[MetroMPG]

I used SPSS to get my results.

i remember using that prog in a stats course i took.

don't worry about sending me a copy. you've confirmed that the key is to be able to figure out the standard deviation and standard error and from there you can figure the margin of error (2x standard error). if the difference between the 2 sets of data is smaller than the margin of error, it's not statistically significant.

i double checked excel, and it does give the same calculation as SPSS did for you (i made a typo when i reported the standard error was .18 further up this thread. i meant to round to .17 but accidentally made hit '8'. excel wasn't the problem, it was me.)

so this is good. it's nice to be using proper stats analysis on these tests. i'll go back to my aero tests and my k&n filter test and update them with proper margin of error info (i had it partly correct for the aero test, but not completely)

[Matt Timion]

so this is good. it's nice to be using proper stats analysis on these tests. i'll go back to my aero tests and my k&n filter test and update them with proper margin of error info (i had it partly correct for the aero test, but not completely)

Feel free to post results for anything you've done here. I would love to install wheel covers if it helps, as well as the K&N filter.

Now here is something I've been thinking about. We just saw that the WAI didn't help too much. Let's assume that something else you install doesn't help too much either (wheel skirts, etc.) statistically speaking do you think it might be possible for the two items combined to provide a significant improvement? My gut tells me no, but part of me thinks it might be possible.

okay, done rambling now.

[MetroMPG]

Let's assume that something else you install doesn't help too much either (wheel skirts, etc.)

i'm sure i've mentioned this already, but wheel skirts work. i've tested them (http://www.metrompg.com/posts/grille-blocking-part-2.htm) on my car and they improved mileage by 2.8%. under controlled conditions, like today.

http://www.metrompg.com/posts/photos/aero-front-3-4-trim.jpg

the k&n made no difference (http://www.metrompg.com/posts/air-filter-part-1.htm) to steady speed fuel economy. i was so disappointed (because i blew 80 bucks on it), i actually ran a second test just to make sure (once at 80 km/h, once at 95). nada. not statistically significant. if only i knew then what i know now..

[SVOboy]

the k&n made no difference to steady speed fuel economy. i was so disappointed (because i blew 80 bucks on it), i actually ran a second test just to make sure (once at 80 km/h, once at 95). nada. not statistically significant. if only i knew then what i know now..

And matt and I were gonna go out and buy them though...But I'm still doing WAI because I know the crx adjusts fuel based on IAT and not timing, so I figure I won't run into ecu errors.

[Matt Timion]

the k&n made no difference (http://www.metrompg.com/posts/air-filter-part-1.htm) to steady speed fuel economy. i was so disappointed (because i blew 80 bucks on it), i actually ran a second test just to make sure (once at 80 km/h, once at 95). nada. not statistically significant. if only i knew then what i know now..

Let's talk about cost then. How many regular air filters will you have to buy in order to break even on the K&N? Since they are good for life that has to account for SOMETHING, doesn't it?

Oh, and I think the K&N air filters are only like $40-50 for my car, so that's a plus.

[MetroMPG]

despite the test today, i think it's still reasonable to state that the WAI probably improves fuel economy on my car. just maybe not at constant speeds, as tested on a warmed vehicle.

but as matt pointed out elsewhere, WAI will cause the engine to warm up faster. warmed up engines get better mileage than cold ones. fact.

unfortunately, (1) i have no way to reliably measure this, and (2) it's probably just a tiny improvement anyway, hard to measure.

i went back into the insight forums tonight, and the primary reason they do it is because the car won't go into lean burn until the IAT sees a certain temp. everyone there claims WAI makes a measurable difference (though I never saw any talk of controlled tests - maybe the improvement is so obvious on an insight that they're not needed).

as for k&n, already figured it out: it will take 125,000 km for mine to pay for itself.

i actually considered selling it on the teamswift site, but kept it partly because it's reusable, and partly because my test was only for mpg at a constant speed (throttle opening). the filter may still offer performance and/or mpg benefits under variable throttle conditions (i.e. wider), which of course i can't test reliably.

having said that, k&n would probably not be very happy if this report (http://home.usadatanet.net/~jbplock/ISO5011/SPICER.htm) was widely distributed.

http://home.usadatanet.net/~jbplock/ISO5011/Spicer_files/image002.jpg

"Filter efficiency is a measure of the filters overall ability to capture dirt."

much more in that link.

[rh77]

the k&n made no difference (http://www.metrompg.com/posts/air-filter-part-1.htm) to steady speed fuel economy. i was so disappointed (because i blew 80 bucks on it), i actually ran a second test just to make sure (once at 80 km/h, once at 95). nada. not statistically significant. if only i knew then what i know now..

Let's talk about cost then. How many regular air filters will you have to buy in order to break even on the K&N? Since they are good for life that has to account for SOMETHING, doesn't it?

Oh, and I think the K&N air filters are only like $40-50 for my car, so that's a plus.

If it's any consolation, I just washed and re-oiled mine with 50,000 miles on it, and it can go a LOT longer. Air flow is increased, and top end has to be improved, although, more air can mean more fuel consumption, unless your're at cruise, then a small percentage in economy can be realized. But the kicker is that fuel IN-efficient vehicles are tested for the results and benefit the most. Pleated filters can be restrictive at high air flow rates (larger displacement engines benefit as more air is moved).

RH77

RH77

[MetroMPG]

here's another thing...

in reading some more at insightcentral.net, i discovered the claim that using a WAI in warm conditions doesn't increase mpg above "normal"; it only helps raise winter mpg up closer to the summer values.

the temps yesterday were pretty mild - the coldest CAI measurement was 53.5F and ambient was 39F.

i think i should do a second test when it's much colder to see if it reveals a larger difference between WAI and CAI.

the problem with cold weather testing is it takes the car so long to get to normal operating temp. when it's really cold, long after the engine temp has stabilized, i can watch the mpg continue to climb steadily for half an hour or more until the transaxle, cv joints, tires, bearings, etc, all reach stable temps.

makes it more difficult to do a controlled test.

[Matt Timion]

MetroMPG,

There is a guy on insightcentral that built the coolest WAI I've ever seen. I think I posted a link to it once, but I'll do it again.

http://www.insightcentral.net/forum/viewtopic.php?t=1769

Basically it's a variable temperature intake. If the temperature drops below, say, 60 degrees a butterfly opens and lets the warm air into the intake. If it's a normal summer/spring day the butterfly remains closed and only lets the cold air in.

This would be an AWESOME "how to" if you could get it working.

[rh77]

"Filter efficiency is a measure of the filters overall ability to capture dirt."

I'm not sure about that. Filter efficiency is the amount of air resistance required to force a volume of air through the filter. It takes less force to get air through a K&N, but that doesn't mean that it won't pick up dust and dirt particles in its oil-laden sponge.

I have used K&Ns for many years, and it has its pluses and minuses. It's very efficient -- I like like that -- the engine takes less engergy on the intake stroke to draw air in (using the principle that the engine is basically a big air pump).

After 50,000 miles, I ran a q-tip and paper towel around the inside of the intake, after the filter. It's surgically clean, so it has to be doing its job. Now I don't come across fine, sandy particles, nor have I over-oiled the filter at the last cleaning (to ruin the MAF sensor, etc.) I have a conventional pleated filter in the TL, since I don't plan on keeping it very much longer, and it's also clean.

I think if you're moving a lot of air (like in a big-block Chevy), you'll probably see big differences in power or economy with a change to a low-resistance filter.

RH77

[MetroMPG]

Filter efficiency is the amount of air resistance required to force a volume of air through the filter.

you're right in that that's the "normal" use of the word in reference to filters. but "efficiency" can refer to how well any specific effect is achieved. (in the case of the graph above, it's dirt trapping ability.)

It takes less force to get air through a K&N, but that doesn't mean that it won't pick up dust and dirt particles in its oil-laden sponge.

you're right again, but it's logical that the more open space you have in the filter media, the more particles will get through. it's not an enormous difference, but that's what the test above showed. that said, i don't know whether their test was representative of normal urban driving (particulates entering the air intake). the test may have simulated the equivalent of a desert dust storm.

the engine takes less engergy on the intake stroke to draw air in (using the principle that the engine is basically a big air pump)

this is true mainly at wide open throttle. at lesser throttle openings, the primary restriction is at the throttle plate, and it far outweighs any normal filter restriction. so i don't think we'll really see mpg or performance gains from a high-flow filter under normal driving conditions (diesel engines possibly excepted) - unless "normal" for you is foot to the floor a lot.

theoretically, a high-flow filter and a hot air intake are a good efficiency combination, because the throttle will be open wider to ingest more of the less dense air. (unfortunately, the theory didn't play out on my car.)

i agree with you that the effects would probably be more noticeable the bigger the engine.

i'm glad you liked yours. i've used them in my past 3 cars. but i proved myself wrong in expecting they would help my mpg though.

[MetroMPG]

guy on insightcentral that built the coolest WAI I've ever seen

i saw that thread. pretty inventive, those insight folks.

i take it your car doesn't already have one, right matt? my 89 accord had a factory installed vacuum driven variable WAI/CAI. but it was also carburated (can never remember how to spell that).

i think carbs are much more finicky when cold than fuel injection, so they benefitted from the wai more - it was a driveability issue more than a fuel efficiency issue.

ever experienced carb icing? i did a few times in the honda as the wai "collector" shield over the exh. manifold rusted away. by the time i sold the car this summer, the collector was gone. just the hose remained.

my brother in law said his late 70's mitsubishi colt (also carb'd) had a variable wai/cai too.

[Matt Timion]

having said that, k&n would probably not be very happy if this report (http://home.usadatanet.net/~jbplock/ISO5011/SPICER.htm) was widely distributed.

http://home.usadatanet.net/~jbplock/ISO5011/Spicer_files/image002.jpg

"Filter efficiency is a measure of the filters overall ability to capture dirt."

much more in that link.

The deceiving thing about this report, as can be seen in the above image, is that the marginal difference between the best and worst air filter is 3.13%, which probably not only falls within a margin of error (or at least approaches it) but probably has no real world difference.

In the graphic K&N appears to be a total flop... but the difference between the filters is so minimal I want to conclude that most of those filters (if not all) are exactly the same.

[MetroMPG]

can't argue with you on the way it is presented to exaggerate the results.

let's all say it together: there are lies, damn lies, and statistics.

[Matt Timion]

can't argue with you on the way it is presented to exaggerate the results.

let's all say it together: there are lies, damn lies, and statistics.


Reminds me of an old Cold War Era joke.

The US and USSR were going to compete in some sort of race. Like all other competitions USA vs. USSR was just a way to find out which country was better. Rocky 4 anyone?

So the race happened and the USA won.

The newspapers in America read "USA wins the race!"

The newspapers in the USSR read "USSR gets second place in race. US finishes second to last."

[MetroMPG]

Now here is something I've been thinking about. We just saw that the WAI didn't help too much. Let's assume that something else you install doesn't help too much either (wheel skirts, etc.) statistically speaking do you think it might be possible for the two items combined to provide a significant improvement? My gut tells me no, but part of me thinks it might be possible.

just re-reading this thread and now i understand what you were asking there.

my opinion: definitely possible.

e.g. - some of the aero mods i intend to try are unquestionably going to reduce drag, but will probably be too small individually (e.g. wheel discs, deflectors ahead of the tires) to show up above the statistical "noise" (margin of error).

so come warm weather, i intend to test some of them as a "set". e.g. compare no aero aids to this "set": wheel discs, front wheel arch gap fillers & ride height lowering. i don't think anyone would disagree that these items would only improve aerodynamics (and therefore mileage), but each item's individual contribution may be small enough to potentially get lost in the margin of error if tested on its own.

let's say the margin of error on this hypothetical test is the same as i got for the WAI runs - .38 km/gal

and i measure - individually - improvements of .25, .25 and .35 km/gal respectively for wheel discs, gap fillers and lowering.

from a purely statistical point of view, i would be forced to dismiss the individual results as statistically insignificant.

but if i measured the 3 aero mods as a set and saw a total gain of .85 km/gal, we could deem the set of improvements to be statistically valid.

is that what you meant?

[Matt Timion]

just re-reading this thread and now i understand what you were asking there.

my opinion: definitely possible.

e.g. - some of the aero mods i intend to try are unquestionably going to reduce drag, but will probably be too small individually (e.g. wheel discs, deflectors ahead of the tires) to show up above the statistical "noise" (margin of error).

so come warm weather, i intend to test some of them as a "set". e.g. compare no aero aids to this "set": wheel discs, front wheel arch gap fillers & ride height lowering. i don't think anyone would disagree that these items would only improve aerodynamics (and therefore mileage), but each item's individual contribution may be small enough to potentially get lost in the margin of error if tested on its own.

let's say the margin of error on this hypothetical test is the same as i got for the WAI runs - .38 km/gal

and i measure - individually - improvements of .25, .25 and .35 km/gal respectively for wheel discs, gap fillers and lowering.

from a purely statistical point of view, i would be forced to dismiss the individual results as statistically insignificant.

but if i measured the 3 aero mods as a set and saw a total gain of .85 km/gal, we could deem the set of improvements to be statistically valid.

is that what you meant?

This is what I meant, but there is another factor you have to consider. The margin of error is a direct function of the variability of the results.

In other words, the more varied the results are the greater the variabiilty and therefore the greater the margin of error. Unless your control and results have low variability I suspect that adding more items to the experiment will raise the margin of error on these experiments.

What I'm saying is that the margin of error might raise higher than .25 with the addition of more factors.

[MetroMPG]

i can't see how testing a set of items (vs an individual item) would increase the variability of the results (these aero mods specifically).

the variability comes not from the mods themselves (they're either going to work or they're not), but from outside factors: measurement errors (scangauge), variability in the cruise control, experimenter error (i don't know... maybe pressing the "resume" button too late so the car hasn't leveled off by the "start" marker...) wind gusts, changes in temp, etc.

---

this has nothing to do with the point made above, but just out of curiousity, i went back to look at margins of error on the other controlled runs i've done:

* .28 km/gal (N=4 bi-dir runs, mean = 68.85 km/gal) = 0.41%

* .39 km/gal (N=3, mean = 78.06) = 0.5%

* .35 km/gal (N=3, mean = 66.68) = 0.52%

that seems pretty damn good if i do say so myself - my range of error so far is around .5%, meaning i can statistically observe changes in mpg on my car at the level of .25 mpg (us) (hypothetical 50 mpg baseline) under controlled conditions.

[Matt Timion]

i can't see how testing a set of items (vs an individual item) would increase the variability of the results (these aero mods specifically).

the variability comes not from the mods themselves (they're either going to work or they're not), but from outside factors: measurement errors (scangauge), variability in the cruise control, experimenter error (i don't know... maybe pressing the "resume" button too late so the car hasn't leveled off by the "start" marker...) wind gusts, changes in temp, etc.

---

this has nothing to do with the point made above, but just out of curiousity, i went back to look at margins of error on the other controlled runs i've done:

* .28 km/gal (N=4 bi-dir runs, mean = 68.85 km/gal) = 0.41%

* .39 km/gal (N=3, mean = 78.06) = 0.5%

* .35 km/gal (N=3, mean = 66.68) = 0.52%

that seems pretty damn good if i do say so myself - my range of error so far is around .5%, meaning i can statistically observe changes in mpg on my car at the level of .25 mpg (us) (hypothetical 50 mpg baseline) under controlled conditions.

Agreed... you've managed to keep things very constant in your testing.

The only reason I say that testing a number of things will increase your variability is because the possibility exists for there to be more malfunctions. A screw can come loose and change an aero mod, etc. It's not likely, but the possibility exists.

[MetroMPG]

A screw can come loose and change an aero mod, etc. It's not likely, but the possibility exists.

a loose screw? no way - all my aero mods so far have been duct tape, duct tape, duct tape!

(i just hope the adhesive comes off the paint when i do up a proper grille block in the spring!)

[MetroMPG]

WAI update:

i'm still trying to understand why the test showed no significant improvement. to do that i need to know the role of the IAT sensor in fuel metering or timing on my car.

another guy on the teamswift site sent me scans from the factory service manual for this car & year, and IAT is not even mentioned in the sections on fuel metering and ignition timing. so the mystery deepens.

---

another guy on that forum set up WAI on his older geo (like jangeo's). he reported a 6.5% decrease in hwy mpg in 2 or 3 fill-ups - though he's doing tank-to-tank comparisons, so i'm not relying on his claims as experimentally valid.

---

all in all, it didn't work and i've spent a lot of time trying to figure out why with no luck yet. i keep intending to abandon this tack, cool off, and wait for warmer weather when i can do aero mods to the car which are NOT "possible" improvements, but actual improvements.

[JanGeo]

Ok maybe the reason it didn't work is because the air gets to that temperature by the time it reaches the cylinder anyway. You don't have a low temperature plastic intake manifold right? my xB has a plastic intake so the air charge stays colder. Now if it was really cold outside then the wai would probably help but you are heating the air into the cylinder which makes it harder to compress and reduces the thermal expansion ratio between pre and post combustion. BTY the 1994 Geo gets its air from the right fender - have to watch deep puddles. If you had fuel vaporization issues then the warmer air would help but you probably are running acetone right? You need an air temp sensor right at the valve and see if you are really heating the air significantly.

[rh77]

I think you have to understand the dyanmics of your ECU. I know for a fact that the warmer the incoming air into my engine, the leaner it runs -- I can post charts and graphs straight from the OBD-2 port if necessary.

On this tank of gas, I bottomed-out on a visit to a friend's house -- it always happens when I visit there -- the curb is at a steep incline into the driveway -- basically it ripped the WAI in two pieces as usual, and I started sucking-in cold night air (30-40 degrees, which translated to 60-degree air at the temp sensor, and the fuel trim was closer to zero). Completely killed my mileage on the first half of the tank. The next day, I re-routed the intake duct through a tight angle, directly into the airbox. Air temps were up, fuel trim went negative (lean), and mileage increased. Honda/Acura fuel maps like hot air to lean the mix. I can't vouch for other ECUs, though...

RH77

[MetroMPG]

Ok maybe the reason it didn't work is because the air gets to that temperature by the time it reaches the cylinder anyway.

i'm not so sure about that. try telling the millions of tuner car owners with CAI's bolted on that it makes no difference...

i'm pretty sure the air is moving too quickly through the intake tract to change temperature very much between the inlet and the intake valve.

you are heating the air into the cylinder which makes it harder to compress

actually, warmer air is easier to compress because it's less dense. expanding air on the other hand is harder to compress, but that's not what we're dealing with here.

Now if it was really cold outside then the wai would probably help

that's something i've considered. depending on the ECU, the WAI may help most below a certain temp threshold. the day i tested it, it was pretty mild out (39F).

<a href="http://www.metrompg.com/posts/wai-test.htm"><img src="http://metrompg.com/posts/photos/WAI-illust-s.jpg" align="right"></a>

and rh77, you're right: the key here is i need to know the role of my IAT sensor / ECU in all this. i need the factory service manual.

and i need to get a proper datalogger.

nevertheless, i've posted a page on my site with a full summary of my WAI experiment to date, with pics, for anyone interested: http://www.metrompg.com/posts/wai-test.htm

[rh77]

and i need to get a proper datalogger.

nevertheless, i've posted a page on my site with a full summary of my WAI experiment to date, with pics, for anyone interested: http://www.metrompg.com/posts/wai-test.htm

[/quote]

By the way, excellent detail on the subject of warm air at your site, I completely agree with your assumptions. But yes, I also agree -- the datalogger is an invaluable tool in using real data to analyze modifications. Due to the high integrity of your Scientific testing, I think you would definitely benefit from one (and for data nerds like yours truly, it's fun to gather data about your car). If you have a laptop, some freeware is available to run real-time logging -- all you need is a cable that plugs directly into your OBD port and PC (usually serial-to-USB). Otherwise, a flash-ROM chip can be used (which is what I have -- it stores up to 6 selected variables [2-mandatory: speed and if G-Forces on accel and stop have exceeded a preset limit] -- this is downloaded into the supplied program for charting on a PC. In addition, it pulls and clears Check-Engine codes).

RH77

[JanGeo]

It would depend upon the operating power level as too how long the air was in the intake manifold - low power levels would mean smaller slower air flow and more time for it to heat up. Also leaner fuel trim may be happening with warmer air to adjust to the fact that there are less molecules of oxygen in less dense air. Also too warm air means less density so to maintain the same power output you have to let more air in thus running at a lower intake vacuum higher volume which automatically would inject more fuel at a leaner trim net result is the SAME amount of fuel burned to get the same power out. REMEMBER it tries to burn the gas cleanly with the correct air fuel ratio. What water injection does it add some h2o which when heated by the ignition of the fuel makes steam which expands 1600 time adding more pressure to the cylinder. I used to run it in my 65 Rambler flathead 6 to break up the carbon deposits in the head squish zone to keep the pistons from hitting them. Helped with the knocking with the high compression I was running too.

Only way to get more mpg out is to improve combustion efficiency/completeness, reduce friction, reduce pumping losses, reduce drag.

[SVOboy]

Also leaner fuel trim may be happening with warmer air to adjust to the fact that there are less molecules of oxygen in less dense air.

I thought this was the hole point, and the fact that you are killing your HP a little bit.

[JanGeo]

But killing your HP a little means you would go slower but you tep on the gas more and burn more fuel and allow more of the warmer air into the engine to get the required HP to go the speed you want and the net savings is ZERO! Unless you effect combustion efficency you don't gain anything and could reduce efficency if timing is retarded due to excessive air temps. Developing more MAX HP requires getting more air and fuel into the cylinder and cold air, cold fuel and a less restrictive air filter is the way to do that. Getting the engine to operate more efficently at light loads requires changing valve timing to reduce intake stroke loading on the piston and maximizing power stroke duration and exhost stroke loading as well as back pressure exhost pulse tuning. Think about the forces on the piston drawing air into the cylinder under vacuum, the compression stroke, the power stroke, and pushing the exhost gasses out after combustion. The ideal engine could be a mixed internal combustion - Sterling cycle configuration.

[joke]
How do you top a car . . . you tep on the brake.

[MetroMPG]

But killing your HP a little means you would go slower but you tep on the gas more and burn more fuel and allow more of the warmer air into the engine to get the required HP to go the speed you want and the net savings is ZERO!

and what about the argument that hot/less dense air reduces throttling losses? by stepping on the gas more, you're actually INcreasing the efficiency of the engine at a given RPM:

"The air is less dense, so you get less horsepower at the same throttle opening, thus, you have to open the throttle wider to let in more air and get the horsepower that you need. That increases the efficiency because one of the primary causes of the well-known part load inefficiency of gasoline engines is the throttle loss." - (url=http://lists.osourcery.com/pipermail/bricks/2004-October/008698.html]source[/url])

heating the air is analagous to reducing the displacement of the engine, forcing it to run at a wider, more efficient throttle opening (at light loads).

Getting the engine to operate more efficently at light loads requires changing valve timing to reduce intake stroke loading

doesn't reducing the density of the intake charge also effectively reduce intake stroke loading (a.k.a. energy required to pump air)?

[rh77]

I discovered a site that discusses air temperature and density with relation to fuel. Basically it is a lenghty discussion of Gasoline (http://www.faqs.org/faqs/autos/gasoline-faq/part1/preamble.html), but it did mention a few interesting points.

Taken from the site:

7.6 What is the effect of temperature and load?

Increasing the engine temperature, particularly the air-fuel charge
temperature, increases the tendency to knock. The Sensitivity of a fuel can
indicate how it is affected by charge temperature variations. Increasing
load increases both the engine temperature, and the end-gas pressure, thus
the likelihood of knock increases as load increases. Increasing the water
jacket temperature from 71C to 82C, increases the (R+M)/2 ONR by two [111].

Since I've increased the intake air temperature and water jacket temp, the vehicle may be compensating with timing adjustments to prevent knock (I have noted that at full-throttle, high-RPM runs, a slight knock is noted). I may have to increase the Octane rating to achieve optimal economy (my datalogger shows timing advance as a rapidly variable series of spikes, so I can't diagnose with that respect).

An increase in ambient air temperature of 5.6C increases the octane
requirement of an engine by 0.44 - 0.54 MON [27,38].

Hmm, more evidence to suggest that air temp effects power/consumption and, I guess, Octane -- which needs to be considered more thoroughly now.

Basically the site goes on to discuss air density (altitude) and temperature. Your answers lies here:

7.11 What is the effect of altitude?

The effect of increasing altitude may be nonlinear, with one study reporting
a decrease of the octane requirement of 1.4 RON/300m from sea level to 1800m
and 2.5 RON/300m from 1800m to 3600m [27]. Other studies report the octane
number requirement decreased by 1.0 - 1.9 RON/300m without specifying
altitude [38]. Modern engine management systems can accommodate this
adjustment, and in some recent studies, the octane number requirement was
reduced by 0.2 - 0.5 (R+M)/2 per 300m increase in altitude.
The larger reduction on older engines was due to:-
- reduced air density provides lower combustion temperature and pressure.
- fuel is metered according to air volume, consequently as density decreases
the stoichiometry moves to rich, with a lower octane number requirement.
- manifold vacuum controlled spark advance, and reduced manifold vacuum
results in less spark advance.

There you have it -- basic physics (as noted with older engines). Fuel consumption was based on air volume and temperature. Higher density and temperature = less fuel required for Stoichiometric fuel enrichment. So, with new engine management systems, the oxygen sensor is going to account for this and pump more air with a leaner mixture, equalling less fuel consumption.

So, yes, I have noticed that it takes more throttle input to get up to speed as compared with the CAI -- but it has trained me to use the decrease in power as a forced "slow-down". From a standing start, I accelerate to about 3000 RPM and lift my foot to get the car to shift to a higher gear, and use whatever pedal remains to get up to speed, more slowly, without pushing so far down as to downshift -- this is repeated until top gear with T/C lockup occurs. Generally with this procedure, the torque-converter locks up sooner, and acceleration commences easier with the added torque, despite the RPMS being in the 2000 RPM range, with is low for my little DOHC 1.8L, 4-cyl.

I hope this helps -- at least I now realize that I might have to go up a grade in octane to optimize efficiency with the added heat, especially this summer when humidity increases.

On a side note, water injection is noted to actually increase fuel consumption as it cools the charge and results in more fuel required to achieve Stoich.

RH77

[JanGeo]

Ok that all sounds reasonable and yeah lowering the intake vacuum with wider throttle helps but all the timing implications could and seems to make it a lot more complicated.

The water injection that I needed was to prevent carbon buildup in the head - I controlled when it was injected. The engine was running with 195psi compression in the 12 to 1 compression ratio or higher so it needed a little cooling. Actually I used to get ice forming on the intake manifold - couple of times the throttle locked up on the ice.

[rh77]

The water injection that I needed was to prevent carbon buildup in the head - I controlled when it was injected. The engine was running with 195psi compression in the 12 to 1 compression ratio or higher so it needed a little cooling. Actually I used to get ice forming on the intake manifold - couple of times the throttle locked up on the ice.

That kinda sounds dangerous. Before you joined, there was a discussion thread on the best way to clean away the carbon deposits in the system. We found that water works well, but Seafoam works pretty well too. The thread. (http://www.gassavers.org/forum_topic/opinion_best_fuel_injector_intake_cleaner.html)

[MetroMPG]

more on the IAT vs efficiency mystery...

just heard back from a fellow metro owner with a factory service manual for the 1998 1 liter car:

The PCM uses the IAT sensor in order to calculate incoming air density. The PCM uses air density information in order to adjust fuel delivery and ignition timing.

so that confirms the role of the sensor in the Metro: the IAT DOES control fuel trim (for air density), and it's likely doing just what i want it to as a result of the WAI.

however, the PCM is also adjusting timing based on the IAT. generally, the timing is advanced with decreasing air density:

from http://www.stealth316.com/2-ignitionsystem.htm: (though it describes the Mitsubishi 3000GT and Dodge Stealth)...

Intake Air Temperature Correction
When the air temperature is low, for example in cold weather, ignition timing is delayed. Colder air means denser air-fuel mixtures, which both combust faster and can be more prone to knock, and so ignition must be delayed. When the air temperature is high the ignition is also delayed because hotter air-fuel mixtures are more prone to knock.

so it appears possible (likely?) that there's a "bell curve" effect for efficiency vs. IAT.

timing is initially delayed from a cold start, then advances as the IAT warms, and then it's delayed again as the air gets "hotter".

lately, i've been seeing IAT around 110-120F at around 35F ambient. i may adjust the shroud on the manifold to drop that down to 70-80, where the SVOboy's AIT/fuel trim plot shows a 0 correction.

[diamondlarry]

timing is initially delayed from a cold start, then advances as the IAT warms, and then it's delayed again as the air gets "hotter".

In light of that staement, I may have to start experimenting with different resistor values in my IAT sensor mod. Right now the ECU "sees" 242-247 degrees F.

[krousdb]

Yep, it took all of 5 minutes. And the best part was that I had what I needed laying around. Last year I bought 3 of those accordian type 1 1/2" tubes from NAPA for my Prius WAI. I had one left over so I used that.

Starting from my air filter box, the existing CAI goes straight down, then does a 180 and comes straight up. Then there is a tee shaped piece. On one end it is closed with a resonance chanber. The other end pulls air from befind the passenger side headlight. The first thing I did was to yank it off and turn it 180 and popped it back on, press fit, no fasteners. This puts the air intake right at the distributor. Thats better, but not close enough to the exhaust manifold. So I popped the tee back off and shoved one end of the accordian tube where the tee was and liberally applied duct tape. I routed the other end between the exhaust manifold and the engine block/transmission. One wire tie secures the tube to the upper radiator hose. Viola. Done. Hope see some difference.

Honda didn't have a 195F thermostat so I will try pep boys. Thermostat and recirculating ebh next.

There you go.

[SVOboy]

Post up some pictures, y0! I am especially interested to see the grille block on the del sol.

[krousdb]

Post up some pictures, y0! I am especially interested to see the grille block on the del sol.

http://img429.imageshack.us/img429/1643/dcp59821dl.jpg (http://imageshack.us)
Grille Block from cardboard wrapped in duct tape and painted. Fastened in place by wire ties. Plug is for 200W magnetic heater on the oil pan.

http://img407.imageshack.us/img407/4861/dcp59835nc.jpg (http://imageshack.us)

WAI, connection to existing CAI where you see the duct tape.


http://img407.imageshack.us/img407/6028/dcp59843jc.jpg (http://imageshack.us)

WAI ends up between the block and exhaust manifold. All held in place by wire tie to upper radiator hose.

[krousdb]

I just got back from running errands, about 20 miles in all. Post office, a few auto parts stores and honda dealership looking for the 195F thermostat. No luck on the thermostat, all were listed as 172 or 180F. Anyone know where I can get a 195F?

Anyway, with the new WAI, after the engine warms up, it idles lower (600rpm) and a bit smoother that before. I assume that the WAI is working. Maybe the ecu enables a lower idle speed when the AIT is warmer. Before, the lowest I had seen it idle was 650-700rpm.

[SVOboy]

Autozone near me claims to keep a 195 in stock for my car, I imagine it's the same for yours but if you want I could check on yours and send it out with your ecu.

[rh77]

Nevermind, figured it out. I guess I just need to actually read the post...

RH77