Ive been trying to find info on any new Honda ecu and it seems that there are 2 manufactures of Honda ecus. Iko in japan and Keihin Indiana Precision Technology (IPT) appear to have designed the Honda ECUs. The new 2005 ecus are about 4 times faster than 1995 Obd1 and the newer ecus could be even faster. The 2005 ecu has 66000 lines of code to go through and over 470 pages in the fuel&emission manual. It also has 26 maps of fuel and ignition maps compared to 4 in 1995. Honda ecu might also be integrated by both ecu company's so the code is even more complicated to decode.

 

I had a 2007 Fit and from day one I used 91 octane!

 

I still trying to find the article but found this Hondata ECU Performance - Tech Review - Turbo Magazine

 

Steve,

I find it interesting that you concede a few percent improvement due to better timing with 93 octane fuel, yet you claim that while an autocrosser will notice the difference, a "daily driver" will not, no matter how heavy their foot is.

I think you should know that I was a "daily driver" when I first filled up with 93, on a lark. I was told by a buddy who did some casual drag racing (at a drag strip) that some guys fill up with super unleaded or even race fuel prior to dragging to get better times.

I went to the drag strip with a fresh tank of 93 just to see what kind of 1/4 mile time I could expect with an econobox with an automatic transmission (my first Fit). I didn't get the speed I expected from the car, but burning 93 octane fuel, I noticed a difference in my daily commute. I switched back to 87 for the next tank, just to make sure I hadn't imagined it, and was appalled at how pathetic the car felt, so I went back to 93, and haven't looked back since, except for last month, when I did some experimenting with my present MT Fit.

Now, at this point I would regurgitate the details from the Hondata link I posted probably 5 pages ago, but I get the impression you aren't interested in facts unless they fit your preconceived ideas, so if you really are being honest about this, I ask you to read the link yourself.

I don't know why you would think them to be wrong, since they work almost exclusively on Honda cars, including the Fit. (Remember, the Kraftwerks supercharger kit for the Fit comes with a Hondata system.) I also don't know what would give you the idea they are suggesting there is no torque gain from using the highest octane fuel used when the engine is tuned at the factory, as their graph clearly shows the MBT curve--the one where peak torque is made--has the MOST spark advance, so anything less than that will reduce power by definition. It's simple physics.

 

That article is waaaaaay over my head but to a layman like me, well laymen as in terms of cracking codes in the Honda ECU goes, it seems like the max perf is achieved already by using 87 octane and it doesn't adjust itself when it detects better fuel, right? Am I wrong or right in thinking this way? The one or two sentence answer to these questions would clear things up a bit....alot bit actually!

 

Your car will run on regular but it will not run in MBT (mean brake torque) which is the area the ecu makes the most power and runs the most efficient. I notice that the throttle body changes too so you dont have to push down on the pedal so much. I have a scan gauge and have seen this cruising 60-64 mph on the highway I am getting almost 60 mpg with premium. When I do a few full throttle passes the car then is tuned for more power and mileage drops but then driving economically it get better mpg again all with the timing not going past 45 degrees and 27 at full throttle. I hook my SG when I bought the car and only got 36-37 mpg with 2 tanks of regular. I put mid-grade in my mpg went up to 37-38.5 mpg and with premium I get 40 plus most of the time. Honda has the most advance ecu so you can run regular but your not making the most power and not getting the best mileage.

 

The best thing you could do, if you are actually interested in learning this is downloading ECMLink, Hondata is very similar. They both use the principles of actual airmass in grams per cylinder per revolution to determine load.

So the computer is using the Ideal Gas law, {PV=nRT} that you learned in highschool to manage the engine. Now the downside to this, which I run into often because Speed Density can be a cruel mistress, is any time you change something that effects your VE table, like turbo, intake manifold, exhaust, cams, charge piping, intercooler, blow off valve/compressor bypass valve etc. You are forced to recalibrate your VE table or your AFRs/Lambda maps will be all fubared, consequently so will your timing, and your idle will be very rough or surge till you figure it out.

If your injector deadtimes are wrong, the car might not start once it gets to temperature and the fuel enrichment requirements as well as the pooled fuel from injector spray before shutdown will be changing your perceived AFRs causing you to run super lean or rich while cranking and keeping you from firing up. But once you do get your settings figured out, the car will be so much more consistent from day to day. The trick is to try and use, when possible the stock MAF/MAS sensor to populate your VE table. This will save you potentially a couple weeks of swearing and banging your head on the dash. Ask me how I know


Everything I have tried to explain about internal ECU operation is confirmed by that article, and the Hondata Knock article posted earlier. So if you go back and read all my posts, these tables and that article will make a lot more sense. I do use acronyms like VE (volumetric efficiency or pumping losses) and IDCs (Injector Duty Cycles) which are how long the injector is staying open and terms like dead time which is the injector pulsewidth, or the delay between sending the signal to start spraying and when fuel actually comes out. So if there are any phrases, abbreviations like MaxOct, or acronyms like MAP, UEGO etc. please let me know and I would be more than happy to help, PM me email me whatever, if you are cool and need tuning help I will give you my cell # and I will try and walk you through over the phone if I have the time between work school and my own wheeled money pits. If you have questions, ask!


If you look around in this shot you may recognize some of those terms:


If you are paying attention you can see that his TPS @ 104% is not calibrated and he is showing only 5.3* of spark advance so he is seeing lots of knock and the car is running very rich for safety because of this. The Front O2 reading .78v is pretty rich assuming ~600*C EGTs, and the rear O2 which is probably 36" from the engine is showing even richer at .72v, because the car is running a conservative tune for safety, which would correlate with his mere 5* of WOT timing at >6000rpm.

The IATs are in the 40*C range, which is fine, but getting warm. Because he is at WOT (TPS @104% tells us that) and pulling a full 1.2" of Hg it would suggest he has a very restrictive intake system.

At least with his IDCs in the mid 60s hes got one healthy thing going for him! Max acceptable IDC by most tuners who care about their cars and customers is ~80-85% IDCs

If you recall I mentioned earlier how ECUs pull timing and richen the mixture to be safe when fuel quality or atmospheric conditions are poor?

This is a great example of a MinOct table tune:

5* peak timing and .78v lambda @ 6200rpm and 104%TPS (WOT.) The engine is trying to save it's self.

Hondata stage IV allows you to use basic Speed Density and tune through your lap top.

In the timing maps in the article for instance:

The y-axis is load in terms of pressure
The x-axis is engine speed
The z-axis is timing.



Load is ultimately calculated in speed density in internal ECU units which are often arbitrary and are different from car to car. For instance like a common set of values is a unitless scale from 0.33 LoadFactor at idle and 30kPa to 3.55 LoadFactor under max boost allowed in the tables which in this case is almost 300kPa, not bad for a mostly stock ECU from a non turbo car! As you get used to tracking the datalogs and watching what cells are being used at specific engine behaviour things will make more sense. Till you experience it first hand this will remain a nebulous concept and no amount of reading can really break it down unfortunately. All you can do is jump in and hold your breath till you surface again. It is one of the best ways to learn how to tune though. But time consuming. And as I learned you would be best to have a calibrated and streetable daily driver if you are going to tackle speed density for the first time.

If you are in the position to use after market MAP, IAT and UEGO sensors you will learn that placement is critical.

Your MAP needs to be after that throttle body because it MUST register boost and vacuum. The IAT has to be no farther from the engine than in the upper intercooler piping or throttle body elbow. If the IAT is placed in the manifold it needs a phenolic spacer to prevent heat soaking skewing readings. In the case of the Wideband/Narrowband UEGOs completely opposite what you will read on the internet. AEM offers .pdf files to explain how to read, interpret, and properly datalog it's 5v and 1.2v output inline sensors and hooking up the UEGO gauge controllers to your ECU for logging from there as well. There are also tables explaining the relationship between lambda voltage and AFR scales of different fuels.

For instance:

Gasoline with 10% ethanol - Stoich is 14.13:1AFR at lambda 1.0v
Ethanol, pure - Stoich is ~9.4:1AFR at lambda 1.0v
Methanol, pure - Stoich is ~6.x:1AFR at lambda 1.0v

Lambda if you haven't figure it out is more of an excess air (lean condition) sensor.

The curves in the rich areas we are concerned with for max performance are subject to wild fluctuations with change in EGTs. So at 400*C and lambda .7v could be 9:1AFR on gasoline, and at 800*C it could be 11:1AFRs on gasoline.

For max power an NA motor should be looking for 12.5:1AFR gasoline at MBT, and a turbo motor would be more in the neighborhood of 11.3-11.7:1AFRs gasoline scale at MBT.

For best fuel economy, at the sacrifice of emissions cleanliness (excess NOx compounds) you would aim for 15.4:1AFRs gasoline at idle and up to 3000rpm and 100kPa (before boost/spool picks up on a turbo car.)


For instance, for quickest and most accurate feed back on exhaust composition you want your UEGO mounted in the downpipe less than 18" from the turbine housing of the turbo. Or within 12" of the exhaust collector on a NA motor.

This is where you relearn the phrase "you gotta pay to play" because if you are running to lean or too rich you will be going through spark plugs and ~$75 Bosch/Denso UEGO sensors often. But in the end a good tune is worth it. Especially when you can confidently say you tuned it yourself!

While a dyno is nice, a test n tune day at a drag course or even a secluded stretch of highway can work, though any serious speeding for WOT tuning (which is the easiest by the way) should be done on a closed circuit or a dyno. I want to emphasize that most Hondas will be above legal highway speeds by the time you get into 3rd gear. So with that disclaimer out there, be smart your life and others are at risk, it is not just about tickets. Speed doesn't kill, it's the sudden stop.



A couple main points that make a slight difference between most standalones and the Hondata unit appear to be that you can only run up to 3 bar sensor or 29.4psi boost, and thats after you have both sets of 60 cell additions flashed to your ROM. Nor does it specify if we can adjust injector deadtimes based on battery voltage, it doesn't mention anything about hot or cold start enrichment routines, or cold start idle rpm, CoastingFuelCutOffset so you can tune out any clutch-in stalling by changing where the injectors turn on and off in relation to deceleration and throttle position sensor voltage output.

You can adjust your global fuel settings for different injectors and their respective deadtimes, but they did not say if you can tune your injectors individually based on the needs of each cylinder like some computers can. Or if there is a limit on how big of an injector the ecu can compensate for in relation to the stock size injectors.

It is important to point out that injector deadtimes are done in microseconds and not milliseconds like the article states. The table below closely resembles the laptop application on a full AEM or MegaSquirt style standalone ECU, like you would see in other cars.

This is a read out of user selected "displayed values" for comparison between data logs and parameter changes. Hondata will allow you to log up to 48 channels though, which is not bad. They did not specify if you can alter settings in real time with the engine on like you can in AEM/MegaSquirt/ECMLink, but I am not willing to rule that out just yet.

Below you have RPM, Speed, Intake Air Temp, Timing, Gear and Oxygen sensor volts.

Since the oxygen sensor only puts out a range of 0-1.2v we can deduce it is a Nernst Cell exhaust gas oxygen sensor, and the signal it puts out are direct lambda values. A Nernst Cell is often referred to as a Narrowband Oxygen sensor, as opposed to a Wideband like you here turbo guys talking about when they are discussing things like Air Fuel Ratios (AFRs) which are described in terms like stoich ratio for E10 gasoline is 14.13:1AFR, like when I was trying to explain gas scale AFRs earlier when commenting on rich vs lean mixtures.

A quick glance at this log should tell you that the driver was doing WOT tuning during capture. You can see 6 full 2-4 gear pulls going from ~50-100mph in the 2 minutes shown in the frame. You can also see the time jump back up as he coasts down in gear. Notice how the O2 sensor fluctuates between .20v and .85v, but it sits and stays at ~.80-85v during WOT. That means his ECU is doing what it should, though he should probably lean it out 10 or 15 points because he is just fouling his plugs and UEGOs running that rich.

Not to mention how much more power he would make if he could lean out and bring his timing back down to the 23-26* advance range. But for some reason his computer is running super conservative values, which could be attributed to high timing targets and the air temp never going below 90F and approaching ~120F in some places.




The last exception being the hexadecimal coding, but the Stage IV Hondata unit operates on the exact same principles with comparable flexibility to the ECMLink application I posted a couple pages back.

So if you are truly interested in learning what is going on, I will be more than happy to email you some logs and walk you through them. If there are any FitFreaks in the Chicago area that are sincerely interested I will take you for a spin in one of my DSMs so you can see what the Hondata Stage IV laptop application includes.

I am currently re-calibrating my ECU so you might get to see first hand ALL of the concepts I am discussing here, because I have just finished installing all new bigger 2.75" aluminum charge piping, MAP sensor (upgraded from 3.5bar to 5.2bar ) intercooler, FIC BlueMax 1450cc injectors, -8AN braided stainless steel fuel line and swapped out my stock intank pump for a walbro 190lph, (because the stocker couldn't keep up with my inline bosch -044 pump), and a new larger nozzle (675cc) for my meth kit

This table is a little more straight forward in that you can see on the X-axis that the table changes units from millibars to psi at ~1000mb. That is because they have added the full 120 cells for boosted applications. All NA tunes will never reach these cells.:



But before you get to any of that, and my little rant above is just the tip of the iceberg, you need to give the ECU as much configuration information as possible. This is something that gets overlooked somehow, maybe because people get excited when they see all the tabs and menus, graphs and pretty rainbow colored tables



Things like:
MAF/MAS type
MafComp for variances in MAF Raw Hz between individual sensors
MafClamp functions
Transmission gearing
Weight
Throttle body
TPS calibration
Idle Switch
FIAV
MAP/IAT/Baro sensor out put range
Narrowband simulation
Injector size
Base fuel pressure
Airflow smoothing
Open loop thresholds
AirTempweighting
STFT and o2 feedback
Load Scale
Injector Battery Adjust
Base tip in TIP adjust and in some cases many more.

Hondata looks a little more simple than some though, but at a cost of less flexibility.

EDIT: Now I am sure there are things I will have to spell check and clarify in the morning, because it has been a long day and I am quite a few beers into the evening at this point now.

 

As of right now my OBC is showing higher average and its been colder out, we'll see in a couple days when I fill up for the Dragon. That will give me my "day in, day out" useage mileage which has varied from 34.4/35.1/31.4 in the last 3 fill-ups.

 

In my 2008 Fit I got 39-42 mpg summer, Winter for the first 2 years I got 35mpg on the 3rd year winter I averaged 32/33 mpg. I dont know why it came down except it was a little colder and had more snow. Before I traded it in I got 45mpg using the Scan gauge and had a 3 year average of 36-37 with 3 winters/2 summers.
All using premium with mobil1

Winter will bring down mileage about 10-15 percent on average.

 

If it is the asphalt Dragon I think you are referring to, you will see the advantages to 93 v 87 as things become thoroughly heatsoaked. Your biggest difference should be low and mid range throttle response and torque.

Because if the ECU is already pulling timing off the top for higher than preferred CoolTemp and AITs the last thing you want is Knock Retard or running below the "no knock" values simply because you have crossed the threshold for one of the parameters that changes which section of the octane tables you are in.

The quickest way to find out if you were seeing knock retard (that is if you cant feel it either because it was a gradual degree or two at a time safety routine or just don't know how to identify it) is to perform the following sequence right after a pass:

Let it idle down to cool off for a minute or two depending on how hard and far you just drove (so you don't "coke" the oil in the crevices of the cylinder head) then shut the engine off for a minute and restart to reset your octane values to MaxOct. Then on the way out you should feel all the timing you lost on the last few miles coming back for the reverse trip. Or, if your fuel is really lousy, since the engine is still hot you jump back on the gas the first couple times and it can feel slower as you will now be pulling out several degrees at a time. Most manufactures allow the ECU take out as much 15* but it could be more or it could be less.

If you are in an area of the table, like ~750mbars load and 3500rpm there might be 35-40* spark advance for MBT prescribed by the MaxOct map, so if you have 15* yanked, you WILL notice.

I definitely respect and appreciate you taking the time and money to do this experiment, but I do not think you will see any definitive or repeatable results unless you can log all the data and run several days and several tanks worth of passes through the dragon.
Even then the atmospheric conditions will conspire against you to skew things day to day, even hour to hour.
Unfortunately in this sort of experiment the only real observations can be made with the butt dyno

Please be safe, the place is a blast even on a bad day, but it is still highway. There may be someone less skilled then you out there who could really **** things up for you.

 

I bet we waste a considerable bit of fuel with the extra clutch slipping and 2nd gear starts to avoid wheelspin, which is an even bigger waste of fuel.

Mobil 1 is great stuff, I personally use rotella t6 5w40 synthetic in all my gas street vehicles for Chicago winters. 20w50 in the truck year round before I sold her, and in the race car for the summer.

The factory 5w20 oil is great stuff, too.

 

I have read numerous post from people of the damned yankee persuasion about how their fuel mileage suffers in the winter and it is just the opposite here....I think it has a lot to do with the way things are spread out down here that requires us to not have to put our cars through as many cycles of starting the engine, warming it up driving to have it quickly cool down and repeat the process... I have spoken to other people down here about this and they say they get better mileage in the winter also.... When my car was still relatively stock I was getting 6% to 10% better mileage when in the winter but I have experienced -0 degree temperatures no more than twice in my life and 100+ highs sometimes for over a month straight... I think it has been 15 years since I have driven on ice.

 

Also the tire wear changes mileage too. My new car I am using what ever oil the dealer uses. I will ask what it is though. 294 was under construction when I drove in summer so the 45 mpg was when it was done but the SC helped because I lifted the foot more. I noticed you could drive 60 mph and let off a little and still hold 60 mph but get way better mpg. But also noticed regular gas didn't work that way, I usually had to give more gas.

 

Northern winters are colder requiring longer warm ups, and higher loads (heater,lights defroster,wipers) so more fuel being used, but compare to hot summers where the air is less dense but more fuel is used to control engine temps. Thats why when it cools off your mpg goes up because it leans out.

 

I'd say you are spot on with everything there.

 

You make a couple excellent points.

In that respect, I'd say you were lucky for living in a warmer part of the country, but when it gets cold I can always put more layers on.

Plus forced induction loves cold weather! Everything becomes so much more efficient and around here at least the air dries out considerably.

You're right though, Texas is a big place. Thats why they let you, or at least used to let you guys do 80mph day time limits in West Texas.

What Silver Bullet pointed out about running a more conservative, richer tune in the heat would probably explain that. Though you are also running a supercharger, you are probably having both a little timing pulled and a lot of fuel dumped in, which unfortunately means higher EGTs and worse emissions if you are worried about that sort of thing. But not too bad because you are already getting a more complete burn with the S/C and 93.

 

I was wrong on a couple points and would like to come clean and hopefully clear some confusion (if only my own).

After reading conversations on a site the big-boys frequent I found that it's a common misconception that higher octane burns more slowly (slower flame front). The difference in burn rate between different octane isn't significant; the only difference is the knock resistance of the fuel. (nods to bullet: you were closer to "the truth" than I was, however higher octane gas doesn't burn "faster.") This makes my statement that the ideal timing is different for different fuels with different octane ratings wrong.

Also scratch (nods to scratch) posted the most relevant link with his hondata page showing the tables for knock retard (here it is again). At the time he posted this I was thinking it was either one curve or the other; I didn't read it well; but with my misconception that higher octane required different timing to produce the best power, I was drawn to the wrong conclusion.

But this better understanding doesn't lead me to change my opinion that a car designed for regular fuel will perform noticeably better on premium.

For the ECU to retard timing, and degrade power as a result, it has to detect knock.

Even if knock is detected, the ECU returns to normal timing for best power and efficiency.

I'd like to think that the Honda engineers, with clever combustion chamber design, cooling, and valve timing, have reduced the incidence of knock even while increasing the compression ratio of the engine. Also conditions conducive to knock have to exist (i.e. low RPM and more throttle; hot operating conditions, both ambient air temperatures and engine temp; heavy loading, pulling a trailer with your mother-in-law inside). Engine knock could occur even with higher octane fuel under these conditions (less likely) and the timing would be retarded.

The only question that remains is how quickly does the ECU return to "normal" instead of "retarded" timing? If it "takes several tanks" then once detecting knock, the performance is degraded for around a thousand miles. I don't think Honda would have designed their cars this way. Other posts on that site indicate the ECU advances timing each revolution back to normal until the next time knock is detected and the cycle begins again. Using this logic the retard would only be in effect for a few seconds, if that long.

So, unless you're driving at extremes, (autocrossing? lugging the engine in a misguided attempt to save gas?) there isn't an advantage to more octane in a car designed for regular gas. Under normal conditions the timing map delivers the same power whether the fuel is premium or regular. YMMV especially if you've modified your cars to use forced induction.

And no; premium fuel doesn't have more energy than regular. It has less due to the properties of the anti-knock compounds. But I think the difference in energy content is insignificant.

No; premium doesn't burn cleaner than regular. At least since new EPA regulations requiring low sulfur in all fuels, and high detergents after about 2004.

No; premium won't give you more MPG in a car that runs well on regular gas.

Sources:
Gasoline FAQ (note, this was written about 15 years ago, when sulfur and detergent requirements were less or non-existent. Regular today is not your dad's regular)

Engineering Tips , engine and fuel engineering forum (you need to register to search).

 

I use 87 octane on my GE8.

 

Steve, if you bothered to read my long post up thread it would clear up the rest of your misconceptions.

There are no opinions involved, that is purely how these ECUs work. Obfuscate all you want, but this is reality.

If you want to reset your Octane value, shut down, pull the battery, touch the terminals to reset the ECU, put the battery back in and start the car. You are now at MaxOct.

Also:

"For the ECU to retard timing, and degrade power as a result, it has to detect knock."

Had you taken the time to peruse what we have outlined for you, you would know this is not true and why. There are other statements that could be clarified for you but frankly, you are not worth the time or effort because you are willfully ignorant and don't want to learn.

 

I stopped paying attention to your posts after you stated 15-20HP gain in a Fit for a couple bucks of premium gas. But please continue.