I'm going to get a Civic HX and I'm thinking about maybe getting a megasquirt and tuning it for mpg. But I don't know if this even makes sense. It already is rated for 44mpg on the freeway but I want more since I'm going to be doing mobile work.

Has anyone attempted to tune for mpg? Again, this is for a Civic not an RX7. I don't have some crazy idea of tuning an RX7 for fuel economy...that's kind of a waste for a RWD sports car. And I don't really care about power. I'm asking here since megasquirts are the cheapest standalone computers around and they are dirt cheap when buying a used version.

What do you think?

 

i believe that resposes will sound like this... tune for power because thats when the engine is operating most efficeintly.

however this is what i believe..

your stock fuel injection system is probably going to be more fuel efficient.. it uses a mass airflow sensor to measure the volume of air coming in that way it can accuratly calculate the required fuel. the megasquirt typicaly uses a map sensor and manifold air temp to calculate the volume of air.. this is a little less accurate. thats why most production cars use a mass airflow sensor..

on top of that you would have to figure out how to control any other emissions/ fuel econmy systems on the vehicle.

just be happy with your 44mpg.. my DD - a stock S5 vert gets only half that economy!

 

Good info. I think most of the honda's use MAP sensors though.

 

Honda tuned your engine & car to its stock settings for a purpose. At these stock settings, Honda's engineers decided that this was the most efficiency they could safely recive out of this engine in this vehicle. If it was possible to get better fuel economy at that time with the technology they had, they would have dont it.

 

there are plenty of cars that arent as efficient as they could possibly be due to low quality sensors, like narrow-band O2. it is very possible to make the engine more efficient by trying to get it to run at optimal air/fuel ratio.

working on a 44MPG car would prob only net you a 2-3 MPG difference and prob would not be worth it in the end. other thing you can do is granny drive it to milk the mileage out of it. its more about driving style than technology that dictates how efficient the engine can be.

 

most people completly forget that when manufacturers tune they have to worry about emmisions also which emmissions and fuel ecno do not come hand and hand like most cars they run alittle rich in the lower rpm range to keep noxious gasses from accuring and let the unburnt fuel be recycled buy the cat. on this theroy alown there could be some fuel econo gained from tuning you self but you are going to have to be damn good and run it on a dyno for lots of hours so like someone else said you might just want to enjoy you econo and save your money and time for something fun. just my 2 cents.
eric

 

On top of emissions, manufacturers don't tune every single engine that rolls off the assembly line. They have one basic map that is good for emissions and safe to run the engine on, Honda just happens to build the engines and chassis for fuel economy.

Unless you are ultra-lucky, you can always retune a car to be more efficient, but as stated above, that might yield you a very very small increase.

All Civic's from '88 to at least 2000 (I haven't looked at the new ones) are MAP based. If Honda would have been able to get more efficiency out of a mass air setup, don't you think they would have done it?

 

^^no... hondas are stupid

 

just kiddin!

no for real tho. honda could have choosen to use a map sensor for many different reasons.. like cost efectiveness.. or to conserve space or for the fact that a MAF creates more air restriction then a MAP.. i just see that it is more accurate to measure the air flow with one sensor then to use two sensors and then try and calculate the airflow.. it is also very impractical to try and guess as to why something was engineered the way it is.. there are way to many things considered in that process.

for what J_J is looking for i would be willing to bet that he would not find it going to a megasquirt. now if he was interested in learning a hell of a lot of stuff then yeah squirt that ****...

 

There's an expression I always use. It goes like this: "It's hard to out Honda Honda". Meaning that in certain applications, it's going to be impossible for an individual to do better then Honda's engineers.

Consider that the Honda Insight uses a MAP sensor and gets double the mileage of the Civic HX being discussed in this topic. If Honda felt that a MAF was a more efficient way of measuring engine load, then it's certain they would have used one on the Insight. Keep in mind that in order to build a highly efficient engine, you must minimize all losses, including pumping losses. It stands to reason that a MAF requires slightly more pumping energy to draw air through it then measuring manifold pressure directly.

Now that said, it may be difficult to match Honda's stock tune with the Megasquirt as Honda will have a far more advanced ECU. The 'Squirt is great, but probably has 10% of the resolution of the Honda ECU. You're also going to have to control the accessories like VTec and EGR. Both are very important to Honda's lean-burn engines (not that the Civic uses a true lean burn engine like the Insight does, but there are similarities).

 

Good point. Cars with variable valve lift or timing dont use egr's though.

 

Not true at all. In fact my Insight has both these systems and has an EGR system as well...

 

your correct aaron its hard to out honda honda!.. thats probably why they are the forerunner to fuel economic standards.. however.. there has been a common belief that MAF method is more accurate then the speed density method. i even believe that the toyota prius uses a MAF instead of MAP.. however honda did choose the speed density method but their reasons are most likely more complex then any cares to know about.. part of which is probably that for their setup they were able to make it very fuel effcient.

 

Most OEMs are using observer-based engine control now.. Which is really necessary to get the kind of emissions and fuel economy that OEMs are getting now.

Basically you have a physics-based model of your engine running in the ECU, which is then used to actually control all the various things that need to be controlled... The accuracy of things like pulse-width, spark timing, etc... is not likely to be that much, if any better than the ms2 or ms2/extra code running on the ms2 (definitely better than ms1 though), but the model used to decided what pulse-width or advance to set is much more advanced... ms2/extra has the beginnings of something like this with EAE... which is a model-based accel enrichment algorithm. I've used it to get AFRs that barely move, even under conditions where load is all over the place.

Most of those modern OEM ECUs running the observer-based code run on VERY fast CPUs by embedded standards... 400MHz or faster 32-bit cpus as opposed to the MS2 (24MHz 16-bit), MS3 (40-50 MHz with a 80-100 MHz XGATE interrupt coprocessor), or MS1 (8-bit 8 MHz processor).

Eventually, looking past ms3 towards whatever comes out next after that, we'll be running a 32-bit 400MHz+ powerpc based processor, using observer-based control... people will be tuning their models instead of tuning using tables.

Interesting stuff....

Ken

 

^^ wow your pretty smart..

so where do you find interesting information like this?

 

MS3? how long until MS3?

 

Oh, I didnt know that. Thanks

 

I keep using the Insight as an example, and actually it's a very poor one. In the Insight, the LAF (Lean-burn Air Fuel) sensor has complete authority over the A/F ratio. Within a few seconds the car is in closed loop and even a slight change in the LAF reading will have a huge impact on the fuel curve. Of course, the Insight's engine operates at about 22:1 under lean burn, and I don't think I've ever seen it richer then 15:1 once in closed loop.

I assume most OEM systems are the same now, just less extreme?

 

I actually don't know what today's new cars are using, at least not on some of the newer direct injected engines..

I know that most port injected engines that I've seen are still using the more conventional 2 O2 sensor setup, which to me implies that they're not doing anything like what the Insight does, but without knowing for sure, I'd imagine that the Insight engine's ECU is also running an observer-based set of algorithms for controlling the engine.

To answer some of the other questions:

1) I got most of the information I'm using from Bruce Bowling, who got it from various sources that I eventually intend to read more closely, mostly from a set of SAE papers, etc... I think he posted a couple of links on msefi recently in the Hi-tech section as well.

2) It'll be a very long time before the MS3 comes out... it is not likely to be a "DIY" setup, at least not for building the ECU... wiring and expansion will be of course. It's going to be using a 40/80 MHz hcs12x (XGATE) cpu, or a 50/100 MHz depending on availability. James (jsmcortina on the ms forums) and I would like to see the ms3 be an extension of ms2/extra... but with sequential fuel and COP for up to 8 cylinders, and a bunch of I/O. That platform is still very up-in-the-air right now, James and I don't even expect to have hardware we can start developing on for another month or two, and it'll probably be a year after that before the hardware and software are to a point where they could be considered "ready"

Ken

 

I did some more digging, and I'm fairly certain that just about everyone is using some form of observer-based (or model-based) control on their ECUs as far as OEMs go. Pretty much none of the aftermarket standalones are doing this yet.

Basically I don't have any "proof" that they're using this, aside from Bruce says he asked around and dug in himself, and this is what he found. I'm going to try to find out what SAE papers he read though, and see if I can make heads or tails of them.

Sounds interesting anyway.

Ken

 

What exactly do you mean here?

Both MAP and MAF use several of sensors to calculate air flow. Are you talking about the necessity of using a air temp sensor with MAP? Because you do on MAF also.