Say hello to the Mighty Mira! [ Archive] - GasSavers.org - Helping You Save at the Pump

Hi,

Looks like I found my niche here. I have long been interested in fuel economy, however few people were as obsessed or actually knew as much about it as I did, and now I found a forum where people not only know more but have implemented what they know! Very impressive!

And what a torturous path to get here...googling undertray, to these (http://jime.homeip.net/maximase/Airflow%201/Under%20Car%20Airflow%201/article.html)articles (http://jime.homeip.net/maximase/Airflow%202/Under%20Car%20Airflow%202/article.html) on autospeed about how to DIY a very partial undertray, to phil knox, to gassavers.

I am an engineer (EE) and as nearly all engineers do, I love efficiency. And as such, I can't stand most of the automotive world. We are in a stupid arms race for height and weight, with no one better off in the end. But fortunately rising oil prices are waking people up in a big way. It is very possible to get Cds of 0.20-0.25 (or lower!) and we are stuck with 0.35.

Only a few more years before we get cheap commuting cars for the masses. We aren't there yet, but at least those who are in the know can insulate themselves from fuel dependence.

CAR

On to my car. have a 1993 Daihatsu Mira (http://specs.amayama.com/daihatsu/mira/1994_9/) that I purchased recently. I spent quite a bit of time looking and researching before I purchased it. My goal was to minimize both the expense of the car and the fuel costs, with an eye to make a daily driver that was as fuel efficient as possible.

I bought it mainly for the engine. It has a 660cc, EFI engine. I figure it would beat a 1 litre carburetted daihatsu charade (which was the main alternative at the same price point). The 660cc combined with the EFI should set me up with the best base possible.

Cursory tests reveal that I get about 6.5l/100km in city with normal driving (no engine off or anything). Not that bad.

I would prefer a more bubble shape, something that would lower the height and width at the top (e.g. lower seats, less van-like looking), however, this was the only thing on the market at the price range, especially with the EFI engine.

It also has a 5 speed gearbox, better than the 4 speed van version you can also buy (although heavier).

Surprisingly enough the manual says that it should weigh 1000kg, which I am very skeptical of. But maybe so. I might remove the back seats since I have no need for them (or their weight).

MODS

The very first thing I need to do is have some way of estimating instantaneous fuel economy. This sucks because I am in Australia, my car came out before OBD-II. That means I will either need to measure fuel consumption directly (not sure how), or I will have to figure out how to measure injector pulse width, etc etc.

I'm more a programmer than an electronics guy, so it will be interesting to get my hands dirty with this. I do have a rough idea (http://groups.google.com.au/group/rec.autos.tech/browse_thread/thread/6f0151e69c28a55e/a5de3a2632f8224e?lnk=st&q=DIY+trip+meter&rnum=10&hl=en#a5de3a2632f8224e) what to do, based on a newsgroup search:

Computing mileage on an EFI engine is as simple as totalizing the
injector ON time, totalizing distance traveled, multiply in a
conversion factor and display. Hardware needed is a Parallax BASIC
Stamp (about $50) and an LCD display board (about $90.) We use the
PWM input across an injector to grab the pulse widths. We use a bit
line to read the vehicle's VSS to get distance. We write a little
BASIC code to wrap it all together and viola! Mileage/trip meter.

There are several methods of calibrating the fuel injector flow.
The most accurate is to build my fuel injector flow bench and
actually calibrate the injectors.

(http://www.diy-efi.org/diy_efi/projects/inj_flow/inj_flow.html (http://www.google.com/url?sa=D&q=http://www.diy-efi.org/diy_efi/projects/inj_flow/inj_flow.html))

The next best way to do this is to hook up a pulse generator (or
even better a PC using my FIT software at the above URL) to the
injector, fire off the fuel pump, discharge 20,000 pulses or so and
catch the fuel in a graduated cylinder. The pulse width should be in
the 2 ms range. This must be done instead of just yanking the
injector open to measure the flow to account for the opening and
closing delays in the injector.

The crudest but still satisfactory method is to hook the Stamp up
with some code that will display the raw injector and VSS totals,
put an accurate amount of gas in the tank, drive until it's used up
(or fill the tank, drive, fill it again and record the amount), then
compute the volume per injector pulse by dividing the fuel used by
the total number of injector firings.

John

--
John De Armond So unfortunately that seems to be what I have to do first, and since there is no one here who seems to have gone this route, looks like it's up to me. (correct me if anyone has, I'd love to see a DIY of this).

After I have a base I can compare to and a means to compare with, I can then set about actually doing the various mods that actually improve the fuel economy.

So, I am starting with this:
http://i5.tinypic.com/1412wdz.jpg

Roughly 6.5l/100km city. Kerb weight 1000kg? At least 750kg, for sure. Cd 0.35?

Goal (UFE-III):
http://www.diseno-art.com/images/Daihatsu_UFE-III.jpg

The UFE-III (http://www.daihatsu.com/motorshow/tokyo05/ufe3/index.html) gets 1.3l/100km, has a Cd of 0.19, kerb weight of 440kg, engine size 660cc. Notice especially the design of the undertray right in front of the front tires - this should be relatively easy to copy.

I will proceed to make modifications on the order of price/increase FE %.

In the pipeline are:
-corflute undertray (you call this "coroplast")
-corflute deflectors in front of front wheel and rear wheel (side skirts)
-reversal camera and removal of external mirrors
-rear wheel skirts
-smooth wheel covers
-blocked grille
-either raised bonnet or some sort of deflector over the wiper blades.
-LRR tires (+40psi)
-possibly larger tires
-aftermarket cruise control

Possible
-antenna removal
-front wheel bubble faring
-caulked gaps
-shaved door handles (no clue how to do this)

I hadn't planned on doing all this stuff when I got it, especially the stuff that won't be reversable. However, seeing as the car was cheap and chances are that the mods will improve the value, I suppose I shouldn't worry. It's just that I would have liked to have bought a beater since then it wouldn't matter what it looked like and I could have done a half assed job on the asthetics.

Please give feedback and any suggestions etc. certainly wouldn't go astray.

Hey MM, welcome to Gassavers. It sounds like lots of work to build your own mileage computer, but certainly doable with your credentials. But before you go through all that work, you should look into the OzTrip mileage computer, made in Austrailia. Just google oztrip and you will find some articles about it. I emailed the manufacturer in February to order one and got this reply:

Hi Dan

sorry but we are no longer stocking this kit, the kit will be replaced with a new computer towards the end of this year...hopefully...

my appologies.

regards
Robert

You can contacto Robert at: rob@oztechnics.biz

I'm not sure if it will work on a 2 cylinder but you never know.

There is another option called a SuperMID. They will work on the Prius and many 4 cyl Hondas. Depending on the size of your injectors, it may work on your car. As long as they are larger than 115 cc the SuperMID should work to measure fuel consumed down to 0.001 Liters. If you car has an electronic VSS (vehicle speed signal) the SuperMID will also measure distance and calculate average fuel economy. It will also show instantaneous injector pulse width. What will not read correctly would be th instantaneous fuel economy because you have a two cylinder. For more details, just search this site for SuperMID, Super Mid or just MID as it has been referred to on this site.

Good luck!

Welcome to the site! There really is a great deal to learn from everything that's posted. Hmm, I've been noticing something for the past few years. There are a bunch of people from Oz and Canada that are very interested and very knowledgable about FE. I have figured out why yet.:) Again, welcome.

I'm not sure if it will work on a 2 cylinder but you never know. It's actually 3 cylinder, btw.


There is another option called a SuperMID. They will work on the Prius and many 4 cyl Hondas. Depending on the size of your injectors, it may work on your car. As long as they are larger than 115 cc the SuperMID should work to measure fuel consumed down to 0.001 Liters. If you car has an electronic VSS (vehicle speed signal) the SuperMID will also measure distance and calculate average fuel economy. It will also show instantaneous injector pulse width. What will not read correctly would be th instantaneous fuel economy because you have a two cylinder. For more details, just search this site for SuperMID, Super Mid or just MID as it has been referred to on this site.

Good luck! Thanks, that looks interesting. Whatever I do, I want to get it up and working soon because I am anxious to start the other modifications.

Any links to buy it?

MM

Any links to buy it?

MM

I think that Matt Timion is working on some distribution agreement with Yoshi. Why don't you PM Matt for details on cost. Just be aware that it wasn't made specifically for your car so it may not work properly. You won't know for sure until you try it. Kinda risky. I would be happy to help you with setup and calibration if you do decide to buy one.

I think that Matt Timion is working on some distribution agreement with Yoshi. Why don't you PM Matt for details on cost. Just be aware that it wasn't made specifically for your car so it may not work properly. You won't know for sure until you try it. Kinda risky. I would be happy to help you with setup and calibration if you do decide to buy one.

I'd still like to iron out a few things before we bite the bullet and try to distribute it. I should email Yoshi about this. Perhaps it would be good to get a wish-list from the people already using it.

Well, I'm slowly figuring out where to go for what. Apparently service manuals are rare as hen's teeth, I've called pretty much EVERYWHERE. The UK site doesn't stock them either. But maybe I can figure it out by myself.

I checked the first injector, it has two wires connected to it, a black wire with yellow stripe, and a white wire with green stripe.

Also, a google of "vss ecu mira" led to this page (http://64.233.167.104/search?q=cache:brq-NjauHA4J:www.siemensvdo.com.au/files/502586_Vehicle%2520Speed%2520Pulse%2520Locator%252 02004.xls+vss+ecu+mira&hl=en&gl=au&ct=clnk&cd=3&client=firefox-a). Apparently google's cache of it is still good, though the original is nowhere to be found. The page is titled "Vehicle Speed Pulse Locator", which I suppose is what we want. According to that page, the 92-95 Mira (my car), has the Vehicle Speed Pulse:
Behind instrument cluster -bottom screw at speedo head
That's probably all I will need, then. Of course, I will need to check that it actually exists, and to remove the dash.

"Bother", said Pooh.

I suppose that's what I get for buying what is probably the smallest car with the smallest engine for the cheapest price in Australia.

I also tried my first FAS (Forced Auto Stop), or CODFISH, or whatever it is. Basically turning off the car, letting it coast down a hill in neutral, then turning the car on again, putting the car in gear, releasing the clutching and feeding some fuel via the accelerator again.

It works, how about that.

Well, at least you know where everything is, w00t w00t.

You could also check for the vss itself, should be coming off the differential area.

Well, at least you know where everything is, w00t w00t.

You could also check for the vss itself, should be coming off the differential area.
Hey, thanks for the advice!!! I need all the help I can get.

That's probably all I will need, then. Of course, I will need to check that it actually exists, and to remove the dash.

"Bother", said Pooh.


I think in theory you should be able to remove the cluster without removing the dash.

Oh, and that was a good link. I saved the info for future use if we ever start selling MPG computers here.

I think in theory you should be able to remove the cluster without removing the dash.
Yes... if the theory assumes that I am either 5 years old or have plastic arms 2 inches in diameter.

Actually Matt, you are right.

I just went down to buy an automotive multimeter (about $60 Aus, or $50 USD).

I unscrewed 2 screws on the cluster and 3 behind to access the cluster. There were two connectors in back. One which I assume is some sort of mechanical speed signal. I'm not sure what it was exactly, it had quite a bit of grease on it. I figured it was something mechanical and not electrical, as otherwise, why the grease? I managed to get that out. There was also a 16 pin (from memory) connector that plugged into the back that I managed to get out with a bit of work.

And so, with trepidation, I had a look at the screws on the back.
Behind instrument cluster -bottom screw at speedo head
Hmmm. Speedo... bottom screw... Was there a ground to attach in order to get a proper signal? There was! I hooked the black alligator clip to the GND signal, and the red alligator clip to the mysterious bottom screw. Dangling the wires through the back down near the peddles, I hooked it up to the multimeter.

Then I hooked the other stuff back up. The mechanical grease connector didn't feel right, and wasn't. When I started it up, I saw no increase in kph.

Stopping, I jammed it back in and it seemed like it seated properly. I was rather worried, because I thought it might have some mechanism whereby if you screwed with it, it wouldn't work again to prevent winding back the odomoter.

However, as I went faster this time, the speedometer appeared to be functional.

And what's more, as I turned the multimeter to "RPM", I got a signal that went higher as I went faster, went slower as I went slower, and did the same independent of gear.

I HAD FOUND THE VSS AND VERIFIED IT!!!!!

So now it's only a matter of finding a convenient injector signal. I can either buy a stamp myself to program, or buy the Supermid.

But no matter what, this sure beats having to mess around with mechanical vss signals. I can do several things now, measure l/100km on the fly and also attach a cruise control unit, both of which I have wanted to do.

Anyway, I am happy, that gives me a feeling of accomplishment.

welcome, how are your bump starts? i'm curious to hear your results. Although if your stats are correct then that's crappy power to weight ratio :p but i bet its easy to park around town!

I can bump start almost every time (take it to mean clutch starts). However, still sometimes it starts with a bump, and it's a little less predictable than I'd like.

Is it better to change to a high gear for the rpms, or a low gear? I started off using a low gear, but now I think it's working better with a high gear.

I should probably stop it, because just reading the manual it says "Caution: A large amount of unburned gas introduced into the catalytic converter may overheat the converter, which may cause a fire. To prevent this and other damages, observe the followings; ... Do not turn off the engine while vehicle is moving. Do not push or pull start your vehicle."

So far no problems though.

BTW, here are the vehicle stats:
Length: 3.295m
Width: 1.395m
Height: 1.415m
Wheelbase: 2.28m
Fuel tank capacity: 32l
Gross Vehicle Weight: 1040kg.

Now, GVW is usually 300-400kg more than Kerb Weight (just figured out through googling (http://www.mitsubishi-fuso.com/en/technology/qanda/02_01.html?a2)). So kerb weight should be 640-740kg, which is right around what I want.

From what I find, it's not underpowered at all. (Of course, it will be even better when I have shaved off 50% of the drag.)

A bit more about the VSS:

It looks like however the rpm is configured, it measures the vehicle's speed perfectly. I am using the digitor q 1585 multimeter, rpmx10 setting. It shows a speed of 25kph as 250, and keeps going. This means that according to the manual, if it were measuring rpm, a 250 signal means 2500rpm.

Hmmm. since 1kph is obviously 10 on the guage, or 100rpm, what signal is actually being sent? Well, since the rpm meter is designed to be connected via an inductive pickup to a spark plug lead, that means that at 1kph the metre will be sending 100 pulses per minute. Or 1.6667 pulses per second.

There are a bunch of other screws at the back of the instrument cluster. I have a mind to attach wires to all of them and see if I can reverse engineer them. Perhaps I will get lucky and find an RPM signal, an INJ signal, or a fuel guage signal (very likely).

Anyway, I'm very lucky that the VSS signal was so easy to interpret, not requiring any fancy calibration.

Good job with the VSS!

I bump start in the highest gear I can for the speed, makes it smoother.

Anyway, mehbe you'll whip something up in terms of instrumentation, I wish I had the know-how!

To start the car smoothly with the clutch, depending on speed its best to start in the higher gear so that engine revs won't soar too high when coming back alive again. Fuel economy pros use this technique to turn off their engine then putting the shifter into Neutral when acceleration is not needed, and then when a lower speed has been reached (say 30 mph or 50km/h) then one would attempt to restart the car with the clutch in 5th gear, since this method is much less harsh than turning on the engine with the key. One would have the clutch disengaged (clutch pedal in) and then from Neutral slide it into 5th. Then as you let rengage the clutch (slowly releasing to the friction point) you will feel it gripping the engine. Immediately let off the clutch pedal till the engine starts firing its cylinders and then right away press the clutch again to avoid aftershock. If done correctly within the second, the clutch should absorb any NVH (noise, vibration, harshness) and engine revs should rise to about 1100 rpms. If engine fails to start try a lower gear, but remember the lower the gear, the more clutch you need to absorb the greater aftershock. Engine revs could be higher during a bump start like 1500 rpms or 2000 although you want to keep these revs as low as possible in order to reduce wear and tear on the engine (rev too low and the engine won't come back on, one must know which gear to start with so that NVH is controllable for all speeds... in this case its harder to control as your momentum decreases). As the revs fall and settle to idle, one should rev match into the appropiate gear and then accelerate their way to desired speed (usualy 40mph or 70km/h before areodynamic losses are prevalant) and then repeat for maximum fuel economy. I suggest refining this technique before attempting to use in real world driving situations as if fallen to the wrong hands can be deadly (like speeding). A master with this technique can recall the engine within half a second and start accelerating as soon as that second is done with no NVH to either the driver or the passengers. And the mileage increases are reported to beat EPA highway estimates in the CITY. If one hasn't noticed why not just coast in Neutral? Simply because as your momentum decreases, the same amount of fuel used to keep the engine running won't be worth it as opposed to coasting with the engine off and mileage is still increased even at very low speeds. There is no payback for fuel coasting with the engine off as opposed to coasting with the engine on in Neutral. In Neutral one has to be travelling at higher speeds to gain the mileage benifits of coasting (since when one is in gear, engine braking prevents the car from travelling far with just momentum), but with the engine off that rule is broken and mileage soars upward at any speeds. Its just a matter of learning how to turn on one's engine when they need it the most as well as maintaining safety, passenger comfort and not overdraining your 12 volt battery. something to helkp with your bump starts ^ higher gear the smoother, i go as low as 2nd when my speeds are low but i try to make it a high gear.

Well, I had another look at my instrument cluster, this time in more detail. And got some pictures...

As you can see, there is a 22 pin connector in addition to the mechanical odometer cable (or whatever it is). I reverse engineered the pins as best I could with the new multimeter and some educated guesswork judging by which conductors lead to which lights or went to connectors behind which guages.

http://i6.tinypic.com/14kxqav.jpg

http://i6.tinypic.com/14kxqmd.jpg

http://i6.tinypic.com/14kxt8i.jpg

Another great bonus - I was tracing the route that the injector wires take, traced them through the firewall, and lo and behold, I found the ECU! It sits right under the passenger side of dash, and is trivial to access (although I need to get a proper socket set to get the bolts off; my socket screwdriver won't cut it.).

This means that no longer do I have to remove the whole dashboard to get at what I need to.

At the moment I'm a bit awestruck at all the things I get to play with: a VSS, INJ signal, and hopefully an RPM signal (which I can get inductively from the spark plug anyway).

That means I can make:
-l/100km guage (Need VSS, INJ)
-digital speedometer (Need VSS)
-digital tachometer (Need RPM)
-digital temperature signal (useful for grille testing) (Need Temp signal)
-digital tank guage (also how many km left to travel, worst case). (INJ)

All of which should be relatively easy once I understand how to setup the basic stamp and the LCD screens. This looks helpful (http://electronics.howstuffworks.com/microcontroller3.htm).

So, my basic plan of attack is now:
1) Get basic stamp, LCD, board, etc, everything needed for final project.
2) Pry off ECU to get a good look at it, see if there is any schematic on the back, and see how to easily detach the various wire bundles for splicing. (Need socket set).
3) Find RPM signal for tachometer.
4) Use a PC oscilloscope through soundcard setup to analyze the various signals, such as this one (http://alfcasa.homedns.org/forum/).
5) Splice various wires such that I have easy access to every signal I need.
6) Start programming the stamp and testing out the basics of the device, troubleshooting until finished.

I think that's it.

All my quick replies get lost...

Anyway, this is worth retyping...

Good job with the cluster, I had to do a bit of that and it was a pain in the side.

Also, you should have a tach signal coming off your distributor and going to the ecu to control the function of the engine wrt timing/fuel, so you shouldn't need to use a spark wire.

Also, you should have a tach signal coming off your distributor and going to the ecu to control the function of the engine wrt timing/fuel, so you shouldn't need to use a spark wire. Hey, thanks!

I got the basic stamp, will see if I can find the wire you are talking about tomorrow.

Hint: Before posting, hit CTRL A, CTRL C.

I dunno if this is the industry standard...but try solid blue, :p

Well done Mighty Mira. I wish I could do something so productive.

Well done Mighty Mira. I wish I could do something so productive.
Try putting a teggy cluster in your car, it's as much fun, :p

I took some pictures of the ECU. Here is one (after I opened it:

http://i6.tinypic.com/14y0kfb.jpg

The various i/o pins are as follows (with my guesses for what they signify in brackets):
Top 12 pin plug:
IG (ignition)
E02
#10
FC
ISC
VSV1
E03
E01
AMP
W
VF
BAT (battery)

+B1
SPD (speed)
IDL
N+
THW
PIM
VC
E1
+B2
A/T
AC
PSW
N-
T
THA
OX (With an overbar over the "O" - could this be an oxygen sensor?)
E2

I also note that there is a group of three sequentially numbered labels - E01, E02, E03. I'm guessing that these correspond to individual cylinders.

If any of you have any idea as to what these may portend, let me know. I think I will need to get an oscilloscope, hook it up to individual outputs, and play with the car in order to get a feel for how the outputs change and what those changes may signify.

p.s. The outputs I'm most interested in are as follows:
Speed
Injector (any injector)
Tachometer
Also, a BAP sensor would be invaluable (or if there is another way of figuring out the Barometric Absolute Pressure), because I want to be able to measure altitude on the fly, hopefully.

Does the ECU have a BAP sensor on it? I know honda put the BAP in the ecu, if that helps.

EDIT:

AC could be the a/c clutch control, that's something you couldnt try disconnecting or checking for power/ground by toggling the ac switch.
A/T could be the lockupsolenoid control, depending on what kind of lockup you had, but since this is in AUS and it's a 95 it's very possible you only would have the one pin controlling lock-up.
AMP Part of the alternator test circuit perhaps?
PSW Does power steering need some sort of electrical crap?
FC Possibly fan control based on ECTs
IDL Mehbe controlling the idle air control valve?

I'm guessing the e01 stuff are injector things. You could always try following the wires/checking for continuity in the coloration.

Thanks for the help,

I'm not sure if the ECU has a BAP sensor in it. Any idea where/what to look for?

AC - you mean Air Conditioning? If so, this car doesn't have A/C.

I wonder if AMP = MAP. If so, that's not going to really help me here because the car will never be at WOT hardly, and it needs the BAP sampled as much as it is sampling speed etc.

I'm just heading out to the car now to check up on the wiring continuity.


Does the ECU have a BAP sensor on it? I know honda put the BAP in the ecu, if that helps.

EDIT:

AC could be the a/c clutch control, that's something you couldnt try disconnecting or checking for power/ground by toggling the ac switch.
A/T could be the lockupsolenoid control, depending on what kind of lockup you had, but since this is in AUS and it's a 95 it's very possible you only would have the one pin controlling lock-up.
AMP Part of the alternator test circuit perhaps?
PSW Does power steering need some sort of electrical crap?
FC Possibly fan control based on ECTs
IDL Mehbe controlling the idle air control valve?

I'm guessing the e01 stuff are injector things. You could always try following the wires/checking for continuity in the coloration.

Most of the things you can test by looking at the signal that goes through them. I should take a picture of my BAP...but I'm a little tired at the moment and I think my batteries are dead, but anyway, it looks like a plastic brick with a little hole in the side, :p Anyway, even if the car doesn't have a/c if any of the cars did (or any of the cars that shared a similar ecu) did the ecu will have a space for the a/c, regardless of whether or not it is used, *shrug* Unless they were really picky about that.

Most of the things you can test by looking at the signal that goes through them. I should take a picture of my BAP...but I'm a little tired at the moment and I think my batteries are dead, but anyway, it looks like a plastic brick with a little hole in the side, :p Anyway, even if the car doesn't have a/c if any of the cars did (or any of the cars that shared a similar ecu) did the ecu will have a space for the a/c, regardless of whether or not it is used, *shrug* Unless they were really picky about that.
The following is a pic of my ECU. Let me know if anything resembles a BAP sensor, otherwise I will have to wait until I get an oscope to have a proper look at stuff.

(BTW car was actually manufactured in Sept 92.)

http://i5.tinypic.com/14y5zz9.jpg

BTW, I just went and copied down all the colors of the wires and the pinouts they correspond to. I will get that up shortly.

However, I did have a look at the distributor like you told me to. I am fairly new to engines, but I found the spark plug leads and followed them from the cylinders to a cylindrical thing on the end of the motor, with three heads pointed off at 60 degrees spacing. I figured that this is the distributor.

So, this has 4 wires coming off it - two come off the end and are white/cream, and black/red, both with two silver bars. They are larger wires, and furthest from the engine.

Now closer to the engine are two smaller wires, signal wires obviously. They are blue/yellow and blue/green. I traced them to the N+ and N- in the ECU. I'm guessing that these are going to give rpm. I can probably test that with my multimeter now.

ALso, my injectors are hard to get at. I can look at the first one, which has white/green and black/yellow coloring. The next one is impossible to get at properly, and the only thing I can do is to lift up the shielding slightly. They too look white/green black/yellow. ????

But when I have a look at the ECU wires, there is +B1 and +B2 showing black/yellow, both are thick wires. Also, there is a white/green wire, that is marked #10. Or if I am mistaken, there is also E01 and E02 which are both white, black and thick.

I have a feeling that I may need to install an aftermarket BAP sensor on the car. I'll have to do some research to figure out how much resolution these devices have.

My basic purpose with this is to enable the FE meter to track changes in elevation.

Damn that ECU is bare as hell! Anyway, no BAP, what do you need it for?

Only 4 wires on the dizzy! Geez.

Black/yellow is the power the the injectors, you'll want the other wire, which is a switched ground, for injector pulse signal.

The rpm wire should only be one wire, hmm, odd. I wonder what sensors are in the dizzy, if any. I wonder if you even have these sensors? They didn't seem to be on the ecu side of things.

Perplexing.

I've finally been able to figure out my highway mpg. Highway is 50mpg (4.7l/100km or 21.3km/l), City is 31.3mpg (7.5l/100km or 13km/l).

Note that the highway wasn't completely highway, but maybe 65-70%. I wouldn't really consider it to be a true highway run. Which means that perhaps unmodified, I might be able to get what?

Let's see, if I travel 45km at city and use 3.46l... and I travel 128km and use a total of 6l, then the highway section I used 2.54l and travelled 83km, giving me 3.06l/100km!

While I don't think it is quite that good, I would certainly be pleased if it was! That would be 70mpg.

I did pump the tyres up to 50psi though. :)

I just found myself an old oscilloscope that should do the trick. Now I will be able to examine the signals first hand, which should prove interesting. I may need someone else to drive the car though. :)

Well, at least for the VSS.

Ok, I have set up my oscilloscope to run off an inverter powered by the cigarette lighter. I did some cursory testing - the scope measures 12 volts coming off the battery, so that seems to be fine.

I looked at the VSS. The VSS triggers a pulse every time the wheels turn a certain amount. That pulse is about 4 Volts. The waveform looks like so (it returns to the same level as it started, excuse my hasty drawing) :

http://i1.tinypic.com/1zp6zyr.jpg

I had a brief look at some of the ECU pins. There is one that varies the pulse width compared to how hard I mash the throttle. It obviously has to be either the rpm, the TPS, or the INJ signal. I suppose the best way to test that is to get on the highway in 5th gear and mash the throttle. If it doesn't change, then obviously it's got to be RPM. If it does change, then it's either the INJ or the TPS.

I'm not sure how to distinguish the TPS from the INJ. Well, I imagine that the INJ will vary according to the RPM, it's just that the pulse width will change. Whereas I'd imagine that the TPS will not depend on RPM at all. Thus if we know a signal is either the TPS or RPM, then a change in RPM should tell the difference.

e.g.:
1. Go to highway speed in fifth gear. Step on the gas. If no change, signal is RPM. If signal changes, go to 2.

2. Go to second gear. Put the throttle down a reasonable amount. As the rpms climb, note the signal. Does something increase frequency? Then it's INJ. If not, then it's TPS.

I'm not sure how to distinguish the TPS from the INJ.
My car's ECU / sensors stay active if I stall the engine and leave the key in the "run" position. If your car is the same, you could stall it intentionally (while stopped) and then observe sensor inputs with the engine off. IE. INJ / RPM will drop to 0, but TPS will continue to vary depending on pedal position.

My car's ECU / sensors stay active if I stall the engine and leave the key in the "run" position. If your car is the same, you could stall it intentionally (while stopped) and then observe sensor inputs with the engine off. IE. INJ / RPM will drop to 0, but TPS will continue to vary depending on pedal position.
Thanks for the hint!

I made some more progress. I tested a few more pins at random, and got some interesting ones.

I think I found an injector signal. It pulses when I put my foot on the gas, frequency in proportion to rpm, and the pulse width gets longer when I floor it. When I take the foot off the gas, the pulses stop until the engine slows down to the point where the accelerator is, and then resume. That one is labeled "#10".

I think I found the rpm signal, or at least, something that measures rpm. It is labelled "IG".

And I think I found the TPS, which rises up and down when I pump the gas. It looks to be a straight voltage signal, although there appears to be some noise in that one. It is labelled "PIM".

If it's truly the case that these measure all those things, I have all the inputs that I need.

Nice detective work, Mira...

And I think I found the TPS, which rises up and down when I pump the gas. It looks to be a straight voltage signal, although there appears to be some noise in that one. It is labelled "PIM".

Rises & falls with the engine on, or off? The MAP is also a straight voltage signal on my car which varies (with noise) with throttle position - but only with the engine on. Engine off, it's a constant voltage.

Happy hunting...

Mira,

Here is a really good PowerPoint presentation (runs right in IE) about sensors. You have to scroll through a bunch of slides about the educational facility first, but the info is excellent once you get to it.

http://facultyfiles.deanza.edu/gems/waltonjohn/sense.ppt

Nice detective work, Mira...



Rises & falls with the engine on, or off? The MAP is also a straight voltage signal on my car which varies (with noise) with throttle position - but only with the engine on. Engine off, it's a constant voltage.

Happy hunting...
Hmmm. Engine off, it's a flat line. Maybe I will need some more hunting. Should the TPS signal come directly off somewhere near the throttle?

Mira,

Here is a really good PowerPoint presentation (runs right in IE) about sensors. You have to scroll through a bunch of slides about the educational facility first, but the info is excellent once you get to it.

http://facultyfiles.deanza.edu/gems/waltonjohn/sense.ppt
Thanks! That was interesting. I hope that I will be able to tell the difference between the MAP sensor and the TPS.

Hmmm. Engine off, it's a flat line.
Just to clarify, engine off, or engine stalled with the key at "run"?

Should the TPS signal come directly off somewhere near the throttle?
Normally, yes. If I'm not mistaken, the TPS is a simple potentiometer, eg.:

http://www.sdsefi.com/pictures/tps.jpg

Well, I spent most of last night building a digital tachometer, from a kit (http://www.dse.com.au/cgi-bin/dse.storefront/44b78ace0c347eea2741c0a87f9c0782/Product/View/K3241). It was rather fun, actually. What was more amazing is that when I tested it, it worked straight away. When I hooked it to the wire marked "IG", the meter went up as the rpms climbed. Since I didn't have a stable 12V power supply at home, only in the car, I couldn't really calibrate it properly yet. But it's rough enough. I do have the oscilloscope to check frequency, however. And that told me that my car idles at around 1000rpm (25Hz).

I needed to change the timer. Although the catalogue states that it only works for 4-6 and 8 cylinder cars, it also works for 3 cylinder cars.

It supplies the following resistors for the different cars:
8 - 47k
6 - 56k
4 - 82k

There is a 50k pot to adjust, so I just soldered the 47k to the 82k to make a 130k resistor. Good enough. Calibrated the tachometer so that it idled at around 1000rpm, which is good enough.


I also had a bit more of a look with the other oscilloscope, which appears to have a better trigger. That #10 is definitely an injector signal, which means I'm in business for constructing an instantaneous FE meter.

And I'm pretty sure that the other thing is the MAF sensor, as it changes as the rpm changes.

Well, I finished the RPM meter, calibrated it properly and gave it a test run in my car. It has yet to find a proper home, that is the next step. But we are almost there.

Here it is:
http://img84.imageshack.us/img84/1/pic001rs2.jpg
http://img84.imageshack.us/img84/4936/pic002hp5.jpg
http://img84.imageshack.us/img84/1792/pic003aq8.jpg

So... now I know that my Mira idles at around 1000rpm, and hits 4200rpm in fifth gear at 100kph! That's not exactly "just ticking over". I wonder what the redline is? The speed goes up to 140kph on the speedo, but the engine starts to sound ugly at around 5000rpm (119kph).

It's going to be ripe for taller gearing/larger wheels. With the addition of aero mods so that the engine is not struggling at cruise, there should be lots of potential here.

What type of gearing modifications can you do transmission wise? How much did you spend on the digital tach, I need one for more precise stuff every so often and it'd be cute to make one, you know.

4200 is ridic, do you have any stats on the tranny?

What type of gearing modifications can you do transmission wise? Hmmm. It's a fwd, do they even have differentials? I'm really not sure what my options are here. I agree that 4200rpm is ridiculous. I will certainly find a way to tall out the gearing, just not sure how or what my options are. Any advice welcome.

I certainly won't be modifying it until my aero mods are done though, but it should give me some time to do some investigation.


How much did you spend on the digital tach, I need one for more precise stuff every so often and it'd be cute to make one, you know. Roughly $30 USD. Link here (http://www.dse.com.au/cgi-bin/dse.storefront/44c1ec760675a1bc273fc0a87f9c078e/Product/View/K3241). I suggest you'd search under google via "digital tachometer kit" for something a bit closer. I estimate I probably spent 5 hours soldering it.

BTW accuracy is only two significant figures. The last two digits are wired zero.
4200 is ridic, do you have any stats on the tranny? No clue. 4200 seems reasonable, as the only reason the engine doesn't sound high is because the road noise covers the engine at 100+kph. It does sound noisy at the same revs and lower speed. And the fact that it idles around 1000rpm, which seems reasonable.

I'd love to know what the redline of the vehicle is. I figure at least 5000rpm, but no idea above that. Tell you what though, having an instantaneous fuel economy figure is going to make it SO much easier to drive slower. I tend to get impatient and drive faster because I get bored. If I had a meter that would help me to optimize my driving habits for saving money, it would help so much. I want a reward for travelling at 90kph, damnit!

FWD cars have the diff built in to the transmission. Where the axles come out, that's the diff.

Your options are: taller tires and/or wheels & tires, or a different transmission.

Redline: to give you a ballpark - my 993 cc Suzuki 3-cyl (6 valves) rev limiter is 7K RPM, the Honda Beat 660 cc engine has a 7200 RPM cutoff, or a 9k RPM cutoff, depending on the model.

FWD cars have the diff built in to the transmission. Where the axles come out, that's the diff.

Your options are: taller tires and/or wheels & tires, or a different transmission.

Redline: to give you a ballpark - my 993 cc Suzuki 3-cyl (6 valves) rev limiter is 7K RPM, the Honda Beat 660 cc engine has a 7200 RPM cutoff, or a 9k RPM cutoff, depending on the model.
Ahah! So even though the engine sounds like it's going to explode any minute, there is actually no problem! Cool...

I'd still like to get it down to something at 3k or below for cruising.

As for the tyres, probably best to assess things as a wheel/transmission system. Since we may need to change both, and the options probably don't lie along a continuum, best to see what will give best combination of both rolling resistance and gearing.

Thanks for the help!

Well, I just found the throttle positioning sensor (TPS), although I have not yet been able to track it back to the ECU. I basically just followed the throttle wire to the engine, where conveniently it sits somewhere along the intake line.

I opened the housing, and saw that the bottom two wires (a white and a yellow with green stripe) connect when the throttle is pressed. This obviously sends the idle signal, something I found on the ECU, marked "IDL".

The other wire must give the position. It is yellow and red.

Edit: Hmmm. Maybe it isn't a potentiometer.

I had another look. It seems that the upper is permanently on +5 Volts, the lower is permanently on ground, and the middle is at 0V at idle and +5 Volts at WOT.

That's pretty damn lame, if I am correct. It means that I can't get a look at where the throttle is without grafting on some sort of aftermarket TPS. Damnit. I wonder why the engine doesn't need to know where exactly the throttle is?

google gave me this (http://www.zerotohundred.com/newforums/showthread.php?t=24894).

Polyxtroy: Cannot, missing a proper throttle sensor on all Mira's engine. Mira's throttle sensor operates on on/off setting only. Whereby a proper one's gives the ecu a range of trottle positions. E manage fuel maps are based on rpm/throttle positions/map sensor.. therefore rather untuneable on a L series ecu.. (By experience)

equin0x:i believe polyxtroy have experience with e-manage before he throw it away and replace with something else. L200's TPS is based on logic on/off only (either 5v or 0v)

John Siew: When your mechanic tells you that Daihatsu Mira L200 does not have a throttle sensor hence air/fuel tuning is not possible, they're probably crapping or they're not familiar with it.

The fact is..

The Daihatsu Mira DOES have a TPS (Throttle Position Sensor) but the TPS on a Mira is only an On/Off TPS (same like what we've studied in logic, 1 or 0). Meaning to say, 0 Volts when the throttle is off and 5 Volts when the throttle is on; even if you just touched it a tiny bit.

On a Potentiometered TPS has on/off feature as the Mira. Upon a little bit more throttle after "on", the voltage goes back to 0 Volts. Then as you progressively step on more throttle till full throttle, you should get about 4.5volts to 5volts. So wiring the throttle function in your Mira doesnt do anything as it does not have the progression of the potentiometer.

However, technically if you can fit a potentiometer type TPS to your L200 and still having the On/Off side hooked up to the ECU, you might be able to get the piggy back air/fuel controller to work, but you would probably have to replace the entire throttle body.


The disadvantage of this set up is that, if your vehicle is tuned for 14psi (1kpa or 1.0 bar) of boost, running less boost might be a little richer than required. So to be honest, not much of a big deal as fitting a potentiometer type TPS, you will get the air/fuel tuning to work, but the ECU would still have nothing to do with it.

To summarise that, myth said that air/fuel tuning is not possible on a L200 but the actual fact is, it is possible to be done! Apexi S-AFC is a good economical start. The EF-JL management system does not run a TPS and therefore, you can't tune your low throttle setting. However you can tune the high throttle at full wide open throttle, which may make your mixtures slightly richer if you are running lower boost than what it is tuned for. So basically, looks like if I want TPS I'd better attach my own aftermarket one.

That saved me a lot of work trying to find a TPS that wasn't there. In the meantime, I should be able to build myself a little FE meter or a coastdown tester.

Injector waveform:

http://i7.tinypic.com/211jbd3.jpg

This $20 1960s vintage scope is easily the best investment in test equipment I have ever made. Here it is sitting in my car plugged in from the inverter.

Hmm, smack on a honda tps, same upper/lower limits but with a continual blah blah.Do you have a MAF sensor?

This $20 1960s vintage scope is easily the best investment in test equipment I have ever made. Here it is sitting in my car plugged in from the inverter.
That is one phat ****ing oscilloscope.

That is one phat ****ing oscilloscope. Thank you. I knew another EE would appreciate it, especially one who actually was interested in that sort of stuff prior to getting into college. I wouldn't worry about getting a job post college... it's always the guys who actually do this stuff for fun who always end up readily employed.

It's funny, this is the first non-software EE stuff I've done since college, and I'm enjoying it. All 10Mhz of bandwidth, too! Should be fine for an automotive application.

I wish I had a pic of the banana plug to BNC connector I soldered up. Unfortunately this is the only photo I took with most of the scope in the picture.

The scope cost something like $21 AUD and I had to spend a couple bucks for the connectors and $30 for a new probe. I tried adjusting the probe, but I think it just has some residual waveforms showing through from the inverter in the car. Anyway, it's close enough for what I want to do. Ebay rocks.

To put it in perspective, I would have paid $128 for a new 10Mhz scope with about a third the visual area. Of course, this does take 3 minutes to warm up, but them's the breaks. But it does have excellent triggering, the other scope I bought from the same guy is 20Mhz, dual trace, but does not have as good a triggering. Maybe it's just my incompetence though. (I got the 20Mhz dual trace for $50 AUD.)

http://i7.tinypic.com/212uwzo.jpg

Do you have a MAF sensor? Pretty sure I do. I have a sensor that looks almost like the throttle, but after the rpm catches up with the throttle, the line backs down. Why?
Hmm, smack on a honda tps, same upper/lower limits but with a continual blah blah. Cool! I'll have to have a look. I am not really sure where I would get one from, any ideas? Should I expect it to just bolt on?

I have no idea about it bolting on, I'd have to see a comparison. Matt has some, I'm sure.

Anyway, the MAF would prolly explain the crappy tps, since the MAF will know well how much air is coming into the engine, so*shrug*

I have no idea about it bolting on, I'd have to see a comparison. Matt has some, I'm sure.

Anyway, the MAF would prolly explain the crappy tps, since the MAF will know well how much air is coming into the engine, so*shrug*
Hmmm. Does MAF * RPM = TPS? If so, I could simulate a TPS signal.

The MAF should operate outside of any other crap. I think it really exists only to replace the IAT.

It basically just measure the amount of air entering the engine by weight. So I could see how the ECU would sense idle/wot as seperate things and for the rest use MAF and RPM to calculate fuel trim. But mehbe that's all BS.

Hi all, glad to have found this group!
Being short on cash this Xmas, I chose to drive my 1992 Mira L200 from Alice Springs to Adelaide, (Usually around 1600 km).
Temperatures were (low for here) but around 37-39 degrees. So I pulled out the air-con radiator before leaving (otherwise the car would have overheated). I split the trip over 2 mornings, leaving at 4 am daily. I left with 2 full jerrycans to avoid the exhorbitant fuel costs mid-way.
I travelled while outside temperatures were between 23-34 degrees, spraying myself with water from a bottle to keep cool, keeping windows closed but interior fan on. I was pretty comfortable, other than a paralysed foot on the accelerator to keep a steady 110 Kph.
Well, I can say that the average consumption was 5.1 l/100. The only time it was lower was between 3:30 and 5:30 am, when it was dark and I drove at 70-80 to avoid kangaroos: then I did a refill and had managed 4.6l/100.

Now the trip is over, I have already put larger tyres on my Mira:
replacing the 145/70R12s with 155/80R12s gives an 8.5% increase in radius.

I bought second-hand tryres, testing them in case the tyres scrubbed or the vehicle mis-handled. Fine all round; even better feel to the car. (I will possibly have to add larger shocks to cope with the extra 1 kg/wheel unsprung weight).
The change to gearing is dramatic; I no longer shift to 5th gear in city traffic at 60Kph as was intended for a shopping-car Mira(?). I can now do as low as 70 kph in 5th, as long as there's no hint at an uphill drive.

I'm now preparing my trip back. I've fitted a Uniden GPS301 (picked up for $100) just so I get a correct speedo reading. I've got a cruise control ready to squeeze under the bonnet for the next 1600 kms of straight road. (There is a real risk of deep vein thrombosis (http://www.fda.gov/fdac/features/2004/604_vein.html) when you cannot move your feet around for 10 hours at at stretch.
I am watching these posts intently for advice on sensible modifications.

kind regards,

Steve Peters

Well, I've found the time to post the photos of all the modifications. The next post will have a discussion. Here goes:

http://img253.imageshack.us/img253/3839/00047640x480nh9.th.jpg (http://img253.imageshack.us/my.php?image=00047640x480nh9.jpg)

http://img402.imageshack.us/img402/7111/00049640x480pv9.th.jpg (http://img402.imageshack.us/my.php?image=00049640x480pv9.jpg)

http://img402.imageshack.us/img402/6031/00051640x480rd6.th.jpg (http://img402.imageshack.us/my.php?image=00051640x480rd6.jpg)

http://img402.imageshack.us/img402/6096/00052640x480hj1.th.jpg (http://img402.imageshack.us/my.php?image=00052640x480hj1.jpg)

http://img253.imageshack.us/img253/7214/00053640x480fe4.th.jpg (http://img253.imageshack.us/my.php?image=00053640x480fe4.jpg)

http://img402.imageshack.us/img402/6606/pb090023640x480cz9.th.jpg (http://img402.imageshack.us/my.php?image=pb090023640x480cz9.jpg)

http://img402.imageshack.us/img402/7492/pb090026640x480eg6.th.jpg (http://img402.imageshack.us/my.php?image=pb090026640x480eg6.jpg)

http://img210.imageshack.us/img210/3286/pb090027640x480bw7.th.jpg (http://img210.imageshack.us/my.php?image=pb090027640x480bw7.jpg)

http://img402.imageshack.us/img402/9766/pb090028640x480qy9.th.jpg (http://img402.imageshack.us/my.php?image=pb090028640x480qy9.jpg)

http://img253.imageshack.us/img253/9035/pb090029640x480bq3.th.jpg (http://img253.imageshack.us/my.php?image=pb090029640x480bq3.jpg)

Ok, so in total the current modifications are as follows:
grille blocking (with approx 200mm * 40mm hole for radiator)
- everyone wonders about that. The fan goes on sometimes when traffic is stopped on really hot days. It never overheats. The thing to realize is that the resistance of a typical radiator is very high. Imagine a cluster of tiny straws, for example. Note that they will not flow as well as a single pipe with the same cross sectional area as the sum of all those little straws. The same applies here. Although if I really had a problem, I'd make sure to completely duct the little airhole so that all air had to pass through the radiator.

Air holes on the front of car are primarily for the benefit of the lay public, who appreciate the stylistic aspects of lots of air holes.

front air deflector
The first picture shows this. Basically just a rectangle of aluminium (0.6mm from memory) bent around, so that from the front it sits just in front of the wheel. Took less than an hour (but more than that to test, because I originally made them too long and they scraped).

partial boattail
-it's about as long as you can go and still pass Australian regulations for rear license plate visibility, which must have 45 degrees to the vertical unobstructed.
-I used some sort of aluminium from the local hardware store for the frame, some bolts to bolt it together, and pvc film (0.3mm from memory) taped around with industrial strength clear tape. It flaps a bit in the breeze. You can't secure via drilling for this thickness, you must use tape. But it works. The first time I used thicker stuff, and thought it looked marginally uglier (also aerodynamically uglier), because I used 3 sheets, one for top and two for sides. The heavier 0.9mm stuff I used the first time hung down because it was heavy, meaning I had too sharp an angle between the top of the car and the boattail.

In total, took the better part of a day.

full 0.6mm aluminium undertray
-this took about 4 hours work, and someone helping for about 2 hours of it. Needed jack stands, and rested all four wheels on the corner of a wooden pallet.
Rattles a bit like an airplane at idle (sometimes), next time I take it off I'm putting flame proof insulation to damp the vibrations down. It stretches from the front to the back, just cut out where the suspension is. The reason I used aluminium and not corflute is because my exhaust system zigzags all under the car and it was a false economy to cut it out and patch it with aluminium over the heat sensitive end - far, far, far too much work. As it is, it's two sheets screwed together. I've screwed it into the chassis in a couple places as well as under both sides. I've duct taped the two sides together (for smoothing reasons, not structural of course). If it leaks oil, I might remove some of the duct tape.

rear fender skirts
-These are a single piece of aluminium, screwed using small galvanized sheet metal screws. Note they attach to the mudguard at the rear, flare out quickly and take a slow angle back (much like a typical airfoil). No bracing other than the natural stiffness of the aluminium.

rear deflectors
-self explanatory. Held on by two screws.

removed left side external mirror
-In Australia, if you have an internal mirror you only need a right external mirror, automobile manufacturers stick the other one on for symmetry. So I removed it and fashioned something in it's place with some rubber, some aluminium and some bolts.

faux racing rims
These are the cheapest 12 inch wheel covers I could find. The same pvc film is screwed all around, then heat-gunned a bit as an attempt to shrink it. In hindsight, should have used the tape I used for the boattail. May do if I can be bothered.

In future, I plan to cover up the gaps in the front under the bonnet with expanding foam, and do the front wheel skirts. Front wheel skirts I plan to make with aluminium and fish pond rubber (very similar to inner tube material, but flat). Basically the aluminium strip will extend virtually to the front of the bumper, flare out a little, and also around the back. Probably about 100mm long in total, if that. Hardly enroaching into the wheel well.

I figure if I make the flexible material as big as possible (i.e. roughly the shape and size of the wheel well) then it will bend easily and not rub excessively or ride up.

Ok, so how does it go? I've noticed vastly improved ability to coast. Take a relatively shallow hill, and I can go in neutral down it and maintain a speed of about 80kph. Steeper hills get closer to 90kph.

And that's pretty good, for a fairly light boxy car. The wake in the rain is also vastly reduced.

The pvc boattail flaps a bit at speed, but I can't hear it. My only concern is as to whether the aerodynamics are messed up at all. I might restretch it a bit more and retape it, but I can't really be bothered.

When I'm travelling at 100kph, my subjective feel is that the engine is only working to do the excessive 4000rpm (or whatever it is) required of it, the load of actually working against the wind is fairly minimal. I can of course put larger tyres on it as the desert mira has done. I think that would be a good option.

The other option is pulse and glide. It's funny, I don't think I was all there when I bought the car. The idea was that with a tiny EFI motor of 660cc, roughly half the size of the civic, that it should be more efficient at partial loads such as when I modified it. I didn't realize it didn't bump start too well though - in the manual it warns not to, something to do with the catalytic converter. It seems to take some effort to bump start, and I don't like using up the vacuum in braking without thinking, so that I get the heart stopping no brakes feeling with another car looming up ahead of me.

The other idea was that it was very light (600 something kg) and also had a small cross sectional area, so that was a good starting point.

Anyway, pulse and glide. Now I realize that having a small engine WILL be an advantage, and I might not even need to get bigger wheels. A car in idle should expend a roughly proportional amount of fuel with respect to engine size, so a small engine should get very good results from a neutral, engine-on glide. It might also be an idea to adjust the idle to the minimal value such that it won't conk out. So, if I'm pulsing and gliding, the engine will be in a relatively efficient range when working, and when not working, the bigger tyres won't be helping anyway.

Anyway, couple a car that coasts very well, an engine that doesn't use much fuel in idle at all, and since I haven't been pulsing and gliding, I think I've got some better fuel figures to post. Especially considering that an engine forced to overcome friction at 4000rpm should be much more efficient at 700rpm or so.

All in all, it has been an interesting experiment. With a bit of paint, the car would look a little less science projectish. I'll probably get that done after I make the front skirts. And it is a good learning experience on a very cheap car, before I get something like a prius and give it the full aero treatment.

I'd like to thank basjoos for providing a bunch of inspiration, and a commute that made this all necessary. Thanks also those who have been logging bike miles such as svoboy, I've probably clocked up about 10km trips to local stores so far that would have been done in the car previously.

One thing I would warn, when working with pvc film, start in the middle, pull to the other side, move a bit down, pulling as you go. In this way you don't tend to get ripples. The best way to get ripples is to secure the corners first and then do the middle - it will either be too loose or have ripples that way.

One other thing I had planned was to do subtractive testing rather than additive testing. i.e. Do everything I know should work, then do A-B-A testing by removing the modification, leaving all other modifications on.

Doing this I'd expect to see more of a difference, reducing the error. That's because you are comparing Cd to (Cd + x). The smaller the Cd value is, the larger [(Cd + x)/Cd - 1] should be.

If aussie regulations regarding licence plate visibility is your only concern, then why don't you build a boattail onto your hatch that would cover the entire back of the car and would flip up with the hatch when you open it. Just mount your license and rear lights into the back of the boattail. Design it right and you could also considerably increase your interior storage space.

If aussie regulations regarding licence plate visibility is your only concern, then why don't you build a boattail onto your hatch that would cover the entire back of the car and would flip up with the hatch when you open it. Just mount your license and rear lights into the back of the boattail. Design it right and you could also considerably increase your interior storage space.
That's what I would do if license plate visibility were my only concerns. Remember, I live in the city and my commute is at least half city (it's quite impressive that I'm achieving better than Australia's version of EPA rating for highway as well as city, for city I'm bettering it by about 40%.).

Other concerns are
-ability to park it easily in the city - one advantage of a kei car I've come to really appreciate.
-time/money calculation - the investment could easily be a week's worth of work.
-worries about lift generated by such a boattail. It's a light car, and the torque generated might be significant.
-crosswinds removing a lot of the benefit. As it is now, there is minimal area on the sides for a crosswind to affect
-I want to have things, once painted, eventually look, if not factory, at least competent aftermarket manufacturer-like.

My car is less an experiment in what can be achieved in fuel economy with no expense spared, than a synthesis of the practical and low hanging fruit in terms of payback periods and consumer acceptance. It is also a way to learn how to do this sort of things, hands on, in a car that is virtually a throwaway item. Considering the car cost about $4000k AUD, that's a year's depreciation with most other cars.

If I eventually decide on an experiment that is more "balls to the wall", I will probably be picking a car that is either already mostly optimized for drag such as the insight or the prius, having the bonus of safe, easy and zero fuel used in EOC. That way I can just do grille blocking, mirror removal, undertray, and front skirts. The only problem is that these cars already have the very tip of the rear facing horizontal, so it's not easy to extend via a boattail. In fact, to make it look professional would be a multiple of the price of any fuel saved.

Or: I will be picking a ute/convertible/something etc, that has the combination of overall low height, that height very close to the front and the back exceptionally low, in order to minimize total length. Think ford capri.

So thinking on this a bit more, I'd like to go further than I have in order to approach ultra low fuel consumption. But...
1. It must be similar in practicality to the parent vehicle.
2. It should conceivably be an investment with a fairly low payback period, 3 years at most.
3. Aesthetically, it should look somewhat professional.

One thing that must be investigated is if similar results to a boattail can be had by increasing the rear angle (shortening the boattail) and tripping the boundary layer into turbulence. If so, that might enlarge the range of vehicles that could be modified.

Practically, that means two things.

1. The vehicle must be modified to have as small a drag figure as possible.

2. The engine must have miniscule idle consumption at worst, or so-called mild hybrid capability for everything except stop start driving. Of course, regenerative braking is great with city driving, but a lot of that can be minimized with enough attention on stop lights.

The choice, as I see it, comes down to:
1. Accept the rear end of an insight or prius as given, seeing as modifying the rear as I would like is in the too hard basket, and just modify the things I can.
2. Go nuts with a capri, using a mild P&G with the relatively small engine it has.
3. Enlarge the circle of cars I'd be willing to look at a bit, and nut out all the hairy details of an aftermarket mild-hybrid conversion. Frankly, that looks difficult and expensive to get it to even the stage where it switches on and off and doesn't harm the engine when the throttle is below a certain point.

I suppose if I can start a trend or make it fashionable to have all the things that are currently considered stylistically ugly or unecessary (such as fender skirts, grille blocking, LHS mirror missing, etc), then maybe the auto makers will make something more extreme in terms of what the rear end suction is doing than the prius or insight.

Or maybe someone like JanGeo could help me out with option 3. Frankly, I need to measure my own car's idle fuel burn and evaluate the maths of P&G before I even seriously consider things.