Designing a header for fuel economy (for a Metro G10 1.0L)

[diesel_john]

I am thinking the opposite of the majority here. I am thinking thin tubing in an insulated box which circulates as much heat as possible back into the intake. Since power and econmy are generally inversely propotional. That should reduce the power until something melts. Heating the fuel and reducing the voltage to the fuel injectors might further reduce the power. More rad yet would be sleeve the cylinders for smaller pistons. Mods to keep the kompression ratio up would needed. I know i am getting off the subject but that because i'm thinking the problem is the thottle plate not being wide open.

[RobertISaar]

Originally Posted by aerohead

*EFI and CPU will automatically recalibrate ignition timing and mixture,which headers will tend to lean out.....

there's a terrible assumption.

Originally Posted by TimRogers

So, if you find that you "need" a length of 108", a length of 54" will also have good effect at your chosen RPM (and at double that RPM as well of course.) If you can make the length of the "midpipe" from the collector to the muffler that same 54", that will help to reinforce the pulses at your chosen RPM as well.

I was going to bring this up since no one else has... it would be nice to get primary lengths to work with the initial exhaust pulse, but incredibly impractical for most situations(as has been discovered), but secondary, tertiary, etc can also have some noticeable effect, with each reflected pulse getting weaker.

going to the tertiary pulse, that's a primary length of 36", which isn't too unreasonable. you'll end up with less scavenging effect, but material cost and packaging issues would be far less of a burden. you'll also get a side benefit of having a 2nd strength pulse at 4000 and even a full strength pulse at 8000 (not that I see that being a design goal or anything). that actually should round out your powerband pretty well, since you'll also have a 4th strength pulse at 1000 and even a 5th at 500.

in reality..... look at the factory manifold, it would be somewhat difficult to make a worse design. it is in no way made to take advantage of any exhaust tricks, just simply somewhat correctly sized and made of a cheap, easily formed material.

[freebeard]

Google won't tell me who said this (I think it was Steven Wright):

"Remember: If the traffic signal is set for 25, it's also set for 50"

[Daox]

I have to look under the car and get some measurements. The CAT is right behind the engine I think, and even with the header being ~36" long, it is going to force the CAT to be moved farther back. I'm pretty sure there is plenty of room to do that, but I want to be absolutely sure. Sadly, I have no garage space for the Metro right now... Gotta fix up that Jetta.

[93Cobra#2771]

Not sure if it's been mentioned or not, but O2 sensor placement is VERY important. You need to keep it as close to the original distance from the cylinder as possible.

If not, you run into issues with the a/f mix due to delays in processing switching from the O2's. Too far from the original and the computer will start chasing itself trying to correct for perceived rich/lean issues.

And the O2 needs to (hopefully) share a common pipe for all three cylinders, otherwise you are at the mercy of a single primary. If that cylinder runs different than the other two, then you are at the mercy of that single cylinder.

"Stepped" headers have also become popular, as they start out big and then step down a bit in size to keep exhaust flow moving...

[Daox]

That is an interesting point. However, I'm drastically reducing the cross sectional area which should increase flow velocity a fair amount. Therefore, moving the O2 sensor rearward (into the collector or right after it) should have a lesser or non-existant effect.

[93Cobra#2771]

It's all about how much you are increasing your velocity. If you double the velocity, then the 02 could theoretically be double the distance of current location.

However, we know the velocity isn't going to be doubled. I'd be shocked at a 5% gain. We are all assuming that the current manifold is a hindrance to exhaust flow. As low a performance motor as the metro is, it may not be much of a hindrance. However, it certainly couldn't hurt to clean up the flow a bit. It might be worthwhile to find an old manifold and have it bench flowed just to see what it can flow.

The other issue is heat - O2s require lots of it to function properly. Too far from the heat source, they don't stay hot enough, causing a false read for rich condition.

I only know this due to working on custom tunes for mustangs. Some year models ran into issues with long tube headers that moved the O2 sensor 18" past normal location. These cars ran into idle issues. It was fixed by tweaking a parameter in the EEC that was related to delay on readings from the 02.

Also, the 02 needs to be located on the top side of the pipe, or as close as possible to it. They are very sensitive to water.

And finally, when looking at the piping size and such, why not look at motorcycles in the same HP range? Yes, I know efficiency and such is different, but might be a good starting point. Specifically on cruiser bikes, not crotch rocket liter bikes of course.

[Shawn D.]

Originally Posted by 93Cobra#2771

"Stepped" headers have also become popular, as they start out big and then step down a bit in size to keep exhaust flow moving...

Actually, they increase in size at the step.

https://www.motorstate.com/TrendsInHeaderTechnology.htm

Another primary tube variation involves "step" headers. Used by drag and circle track racers alike to extract that last bit of horsepower, stepped primary tubes progressively increase in diameter, usually in 1/8" increments, from the flange to the collector. The "stepped" tubes help increase exhaust velocity for that last bit of added horsepower. Stepped headers are typically offered in single- and double-step versions; a single step header may have 1-7/8" to 2" primary tubes. A double step primary tube header may go from 1-7/8" to 2" to 2-1/8" between flange and collector.

[93Cobra#2771]

Whoopsie, that'll teach me to rely on memory...

[RobertISaar]

moving the O2 sensor further back will have a similar effect as an older O2 sensor in general.... a little more sensor lag may not even be noticeable depending on how tightly the calibration engineer setup the O2 sensor PI/PID corrections.

not sure of the polling rate in this exact application, but GM(perhaps Suzuki would be better for a reference here) liked using 80Hz update rates for the O2 sensor for a long time, started moving towards 160Hz near the transition to OBD2. assuming a 160Hz poll rate, that's an update every .00625 seconds. assuming a 250ft/s exhaust velocity, the exhaust stream should move ~1.5625ft in that amount of time. if still on an 80Hz loop, then 3.125ft.

the calibration already has code in it to slow down O2 correction response based on calculated/measured airflow, so with increased exhaust velocity and an equal amount of time between pulses going over the sensing element, you may actually end up with a tighter closed loop correction range as a result. with that much distance from the exhaust ports though, definitely run a heated O2 sensor(assuming you don't already).