Oh, and those HDXF tables look WAAAY better than the old. lol

 

I took a crack at it, and although I wasn't able to get a full WOT log after my revision, I was able to get a 2000 rpm sweep from ~3500-5500 rpm and... well... I'm dumbfounded by the results.

What I did was look at est VE Actual corrected and est VE Model corrected from my previous log... I took the difference between the two, which was just around 20% throughout the pull, and made corrections to all of the appropriate tables by subtracting the respective % based on HDFX weighting etc.

STFT on the log before modifying VE tables was +9.3ish to +6.2ish

Now here's where I get lost;

On the new Log, estimated VE Actual corrected literally stays withing 0.1% of 100.0% the entire 3500-5500rpm of the pull, but est VE Model is now much higher than it was before!?!? Here's the log if anyone cares to look

For those who don't I use this data point as a comparison between the two logs.

Pre-VE tune file: 4000RPM, VE Act = 91.6, VE Mod = 114.2, STFT = +9.38%
Post-VE tune file: 4000RPM, VE Act = 100, VE Mod = 135.37, STFT = +3.12%

Wait... what!? VE Act and VE Mod are now FURTHER apart than they were, BUT the STFT floats around between +3% and 0%, which seems damn near perfect to me.

Did I happen to luck-box my way into the almost-correct VE corrections? I must've missed that part of @matt@pandamotorworks lecture that day.

 

The HDFX weighting thing is blowing my mind. I admittedly have given very little research in to it, not really understanding why the HDFX weight is so dynamic and why it's not just a single table since you should be using the OP VCT.

 

That's precisely what I was thinking the first time I looked at them in ATR... Although I can understand that actually making optimum WOT power requires the cams to keep moving as the RPM increases.

 

Okay so I got a full pull in just now. The low end corrections are still there, but they are less, the midrange looks good, and up top I think I over shot the mark, so I dialed it back a bit.

I've noticed that since VE tuning changes the way the car calculates load and airflow that Load Actual is much higher up top than it was before, this is probably partially exaggerated by the fact that I overshot the mark, but needless to say, its affecting base timing because now it's going to be looking at a different part of the table for timing. It probably cost me some power because its showing more load than it had before, so its targeting higher cells on the BL tables that have lesser timing figures. Oops!

Live and learn I guess.

 

Moved the BL timing tables around a smudge to get them back into line with the new calculated load figures then did another pull on my way home from work today.

Fooled around with the VE tables some more because I had overshot the mark above 5500 RPM, and I hadn't done enough below 3500rpm. Still tinkering.

Also, anyone else running the Alpha firmware on their APv3 have their LTFT stop reporting? Mine won't move off of zero at all. Checked the maps and verified that Learning is enabled... I can only imagine its a quirk in the alpha code. @COBB ??

 

Ok, I'm getting it now. Wherever the cam timing ends up references the weighting table and that weighting table determines which ignition timing table to use out of the 15 presented and then also which VE table to reference out of the 15 presented. The tables will blend together based on weight %. So if you have 33% of 6, 33% of 7, and 33% of 9, the VE and ignition timing commanded will be a blend of all three tables.

With that said, it still seems like we aren't seeing the non-OP VCT tables to manipulate, but I suppose we really don't need to.

 

Pssh, we totally need them! GHOST CAM!
(Though turbo's usually hide a lot of loping sound)

I've yet to play with it, but it looks like you could also change the mapped points for HDFX. Of course it'd be a lot of work to redo the VE and Spark tables, but it looks possible at least.

 

Thats exactly how I interpret it, at least. As far as part throttle VCT, I'm not sure I'm exactly concerned with that. At least not until a different head or cams come into play. I don't see much reason to deviate from what is most likely a stock (or mildly altered by COBB) target of best economy for anything other than WOT.

 

You guys are getting most of it, August will blow your minds again. I've got a huge HDFX update in the works which will allow you to see the rest of the mess behind the scenes. The challenge is in the education, not in the implementation. I want to empower you all with the ability to understand this system and know exactly how to manipulate it.

Cheers,


Braden @ COBB
Sent from my iPad using Tapatalk

 

You guys are getting most of it, August will blow your minds again. I've got a huge HDFX update in the works which will allow you to see the rest of the mess behind the scenes. The challenge is in the education, not in the implementation. I want to empower you all with the ability to understand this system and know exactly how to manipulate it.

That's because OP weighting is disabled from the factory. Doesn't need it stay that way

This guy... comes in here and just teases us!!!

HEY!! Wait a minute!! What about the fact that my LTFT is still stuck at 0 on the AP since the Alpha firmware!? UGH! He bamboozled me and then left me here all sweaty and alone, without an answer to my original question... Sigh.

 

Braden stop it, I might pee myself if you make any more promises

 

Ha! Can you post a screenshot of your breakpoints?


-Braden @ COBB
Sent using Tapatalk

 

Ha! Can you post a screenshot of your breakpoints? -Braden @ COBB Sent using Tapatalk

EDIT!: NEVERMIND! Figured out how! Woop!

 

Alright... So I monitored Adaptive Learn Status and I did see it go from 0 to 1. It stayed at 1 the entire time at idle and through some general average driving conditions. But after looking at all 250ish seconds of the log, LTFT didn't move off of Zero.

Is there a minimum amount of STFT that the Learn function wants to see before it adds it to LTFT? Because my STFT were all really slim, between -3 and +3 the whole time. Come to think of it, that doesn't sound like its enough to make the LTFT want to store it.

 

I can tell you back when I tuned EEC-V with SCT the common method was to only make adjustments to the MAF curve based on the LTFT. Most of the time there wasnt any more that 5% adjustment to make. I imagine if my LTFT was working I would get very small figures as well.

 

I've done a few more HDFX to get the STFT in line for a few different driving conditions. Idle, highway cruise, WOT...

It's getting close, very close in a lot of places.

Once I get it to where I like it, I may play with the fuel accel tables and try to get tip-in dialed in, although it's not bad where it is now.

Once I do that, then I'll focus on fine tuning the spark timing... Then I think it'll be time to strap it to the rollers and see what she's made of.


Also, lean burn at light load cruise is a success. I've stayed mostly out of the pedal this fill up with only a couple of WOT pulls and my average is over 29mpg. I assume if I keep myself in check for an entire tank of fuel that I would be well over 30 average.

 

I want to say I saw some min and max values when I last looked through but they were within logical reason. I'm not certain there would be much benefit to modifying these when the VE can be dialed in right enough to be handled by STFT.


-Braden @ COBB
Sent using Tapatalk

 

I want to say I saw some min and max values when I last looked through but they were within logical reason. I'm not certain there would be much benefit to modifying these when the VE can be dialed in right enough to be handled by STFT. -Braden @ COBB Sent using Tapatalk

I wasn't worrying about modifying them, I was just wondering what the conditions are that cause a value to be stored in LTFT instead of handled live by STFT.

 

Operation: Lean Burn seems to be a success, so far. The past two tanks of fuel I haven't been beating on the car. Calculated 30 flat MPG last tank and the dash is showing 30.2 so far on this tank. My regular fuel usage is about 50% highway, maybe more. This tank, average speed so far is about 35mph.


Update: went 257 miles and then got fuel, 34.7 mpg. The first 60 miles of that was averaging 23mpg on the dash (around town) and then I went on a long highway haul for the rest of it. Dash said 30.4 but the calculated mpg by miles/gallons was 34.7

For reference, my current fuel table is targeting 1.08 lambda (15.85:1 AFR) at light load cruise. I suppose I may be able to go even leaner. Maybe 1.09-1.1 lambda (16.02-16.17 AFR)? I figure the goal is to run it as lean as possible while still supporting smooth combustion and without getting misfires? Not sure what the limit of that is.

 

Wish we had real-time fuel mileage so I could mess around with lean cruise a little more. I run a lambda of 1.05 right now and noticed the timing is running off the MBT tables, so not really sure how much more timing would help.

 

I'm not sure if more timing would help but I figure the goal is probably as lean as you can while maintaining combustion stability? Hell I don't know.

 

That graph is very interesting, to say the least. And I appreciate your very thorough explanation.

Considering all of those factors, safety of the engine is my priority. Sure, more MPG is nice, but if I'm gonna have to bite my nails about if I can put it under boost immediately from cruise, it's not that important. Ideally, I would like to know what the limit is there.

I wish there were a way to directly monitor internal engine part temps. We know coolant temp doesn't go up because actual fuel mass burned is so little. But it would be nice to know the valve and piston temps so we could see how long it takes to cool things back down after running lean for a good while.

Although my fuel tables are created mostly by my curiosity, I feel like I've done my best to keep them safe? I have my tables as such that it really only runs lean when load is right around 0.300-0.500 which is where I've determined almost all flat land cruising with very mild incline changes occurs, with or without AC compressor. I'm running only slightly lean at idle, about 15:1, and it quickly ramps into 14.7 when load goes to 0.600 which is well before boost comes in if I remember correctly.

I wonder if anyone has any information on time required to cool pistons and valves? I'm thinking that even if temps are higher and you go WOT, the amount of fuel added by accel tables for the time it takes until boost pressure builds might actually be a wide enough window to cool temps because the pistons and valves "shouldn't" actually be THAT much hotter due to the relatively low amount of fuel being burned. Again, this is all IMO. I have no idea if that's truly the case, or if there's any way to monitor or test it.

As far as emissions, according to that graph, it seems like best emissions would be somewhere in the 17.75:1 area LOL

 

I honestly don't think you'll have an issue at 15:1. My pro tune seemed to run around there, and I trust @Alex@Stratified to not endanger the car. Running 16:1? I don't know. It's just something to keep in mind. Perhaps Alex can chime in as I'm sure its something he messed with.

The 17.75:1 area actually is good for emissions without a catalytic converter. The downside is torque is really crap because the temperature is about the same as stoich in the cylinder, but with 15-20% less gas.

With a catalytic converter though, the HC and CO emissions go quite a ways down, as it essentially combines the CO and HC to make water and carbon dioxide. Modern three-way converters also reduce NOx emissions too, but not well enough to do well with lean burn , so stoich + cat is waaay better than 17.75 with no cat.

OEMs have experimented with lean burn quite a bit as part of their mandated increase in fuel economy, and some older cars (Honda Insight types) used it extensively. They had specially designed catalytic converters to handle the super NOx and extended periods of heat. We'll probably see its return sometime soon, but for a Focus ST, I don't think it's something that's really going to come out as a huge benefit.

Of course, running WOT gets hot too, its just with lean burn you're doing it all the time

 

This is what I don't get about the whole equation... Fuel has a specific energy capacity per mass, injecting less fuel doesn't seem like it can chemically create more heat. Can it? Maybe we're misunderstanding the concept.

I understand that richer than 14.7 we get cooler temps because a certain amount of unburned fuel remains to soak up some heat and be carried out through exhaust. I understand that as we go for, let's say 12:1 to 14.7:1, we are able to burn more of that fuel as we reach optimal mixture and therefore are creating more heat. But I think we're misunderstanding what happens PAST 14.7:1

I think when we say "lean", we generally mean "leaner than best power, under WOT" instead of "leaner than stoich." In that case, I can understand how we're creating more heat because we're burning a higher percentage of the fuel per cycle.

But what about when we go leaner than stoich?...

Imagine this (at least this is how imagine it);

Instead of thinking of this as a question of more or less fuel, let's imagine it as a question of more or less air.

Let's say we have 1gram of fuel.

1 gram of fuel we add 14.7 grams of air, we're burning all the fuel with all the air and releasing X amount of heat.

If we remove some air, we create a "rich" environment and are unable to burn all of the fuel, we burn all the air but not all the fuel, leaving us only to be able to release a portion of that 1 gram of fuel's BTU, meaning less heat.

But if we ADD air to that 1 gram creating a "lean" environment, we still only burn the original 1 gram of fuel, releasing the same X amount of heat, except we just have left over, unused air that pass along with it out the exhaust.

Does this make sense? Or am I just dumb today?

I'm thinking that stoich has to be the point of highest heat because it's the point where we burn all of the fuel.

What I'm trying to say here is, yes, under WOT, when target is 11.8:1 or whatever, leaner than THAT target is nearer to stoich, it's burning a higher percentage of the fuel, and therefore will make more heat and richer than THAT will make less heat because it has residual fuel charge to absorb heat during evaporation.

BUT, at light load cruise, when 14.7 is the target, richer still burns cooler because it's burning less fuel than stoich. BUT leaner than that is no longer nearer to stoich, it's equally as far from stoich, just in the opposite direction. So unless we're burning a higher mass of fuel, we won't make any more heat than we did before.


Therefore, lean burn for MPG... FTW?

 

On a simple physics scale, you wouldn't be far off. But engines are complicated beasts.

Consider this, in your hypothetical world: We have 14.7g of oxygen, and 1.1g of fuel. So our lambda is now roughly 0.9.
In your world, only 1g of fuel will burn, releasing the energy of 1g of fuel, which is exactly the same as it was at stoich. But you and I both know we will get more power at 0.9.

Why?

The answer is the real world isn't perfect. I can have 14.7g of oxygen and 1g of gasoline in a chamber, but unless the mix is 100% homogeneous and perfect, it won't burn perfectly. As a matter of fact, even if it was, it wouldn't burn perfectly!

The primary reason for this is fuel mixtures take time to burn, and once ignited, it takes the flame time to travel. A pressure wave builds up with the flame, and this causes very slight imbalances.
A slightly richer mixture actually ends up burning more fuel because this flame front travels quicker and burns more evenly. Essentially, our richer mixture is packed closer together, so the flame can "hop" from molecule to molecule quicker. Our flame speed goes up, our energy released goes up, and the pressure in the cylinder goes up, which means more torque, which means more power.

With 1g, we might burn 0.95g. With 1.1g, we might burn 1.0g. We get more overall energy from the richer mixture, but we have slightly more unburned fuel (1.0g vs 0.5g). (Hence the HC emissions go up, and CO emissions rise as we run out of oxygen in our mixture!)

That left over fuel though does double duty! Whilst being a "bad" emission, it also takes some of the heat in the cylinder with it out into the exhaust. This keeps our cylinder temps cooler.
We can of course push this so far, to the point that all of the molecules "smother" the flame front out. This gives us our rich misfire.

So now, lets go in the opposite way!

If we get leaner than stoich, say 0.9g, we start to separate the molecules, slowing down the flame front. Our hydrocarbon emissions start to drop, as we burn almost all of the fuel. Say 0.899g now. Almost all the hydrocarbons paired up with some oxygens to get their party on, so CO emissions are lower, we now have much more CO2.

But what about the heat Bugasu, where does that come from? The answer to this one is yet again: time. A lean mixture burns slower. The flame front lasts longer. So while there is less heat overall (less was burnt!) the time it is in the cylinder is longer, and we don't have any excess fuel to absorb the heat! Heat transfer laws are simple. If I touch a hot object for a very short time, I don't get burned. If I stick my hand in it though, and give it some time, the heat transfers into me and burns me! The same thing happens with your cylinders and pistons. The rich mixture burned hotter, releasing more energy and giving us more pressure and power, but it lasted for a short time, and there was a lot of left over fuel between it and the cylinders to absorb the heat. In a lean mixture, we don't get those benefits, and the cylinders are exposed to the heat for a longer period of time, with nothing between the heat and the metal.
So our cylinder and piston temperatures rise.

Because of our excess amount of oxygen, we also have to remember that with our 14.7g of oxygen we added, because we got so much nitrogen in the atmosphere, we added roughly 55g of Nitrogen into our cylinder! Now because we didn't burn all of the oxygen, and we got the extra heat laying around not getting absorbed, nitrogen and oxygen start to become friends to a higher extent, and we increase NOx emissions.

When we kept going richer, we smothered the reaction. The opposite happens as we go leaner, we just don't have enough for the spark to ignite, or it ignites but the flame front cant spread evenly. This causes the lean misfires.



We can go more in depth, but I think this is a thorough enough run-down.

I'm not saying your car will explode from heat damage if you run lean. Just that you should keep it in mind when doing it. Don't push it too far!
Also, this only applies to gasoline. Diesel, ethanol, kerosene, etc all have different properties and change some of what is discussed here.

If you got any more questions, let me know!

 

Cars don't use lean cruise because of emissions. There's nothing wrong with it. I target 1.05 and it works well for me. I used it on my 2004 GTO and was getting 28mph on the highway with no ill effects. Past 1.05 lambda the car did begin to misfire, but we aren't driving a GTO, we are driving a Focus ST.

I'd actually rather richen it up a bit to maybe 1.02 or 1.03 and run more timing.

 

Okay. Had to read that a couple times to grasp it. But it makes good sense. I also like the holding your hand to a flame analogy for time exposed to heat.

Isn't there something to be said for the actual amount of heat that things are being exposed to? Im sure there is, maybe hard to calculate though.


I understand there's less buffer of insulating gas and little to no unburned fuel to absorb heat... But isn't there a fairly large difference between total eat generated during medium-heavy load on the engine and light load cruise?

To continue the hand analogy, isn't it similar to holding your hand for a short time, wearing a glove, over a blowtorch, vs. holding a bare hand for a longer time over a bic lighter? There's just more heat with the blowtorch than there is with the bic, so even being unprotected isn't getting as hot as a covered hand to a blowtorch...

I suppose I wish I just had actual data on this. I feel I grasp the concepts fairly well. I just don't know how much of a difference it actually is.

 

On a car with a well designed piston, combustion chamber and very well metered fuel like our ST, running it a little off stoich does not disturb combustion stability. An older car you'd have a hard time even running stoich! Stoich is not maximum power, it is simply least emissions.

 

To my knowledge the chart bugasu posted is incorrect. I did a lot of research on the whole "green" movement and CO2 and I remember an almost identical chart but the CO2 levels peaked at stoich and were significantly lower on either side while all other gasses that are actually dangerous were at there lowest levels at stoich. Pretty odd but not really the focus of this topic. I also remember when i was researching lean cruise that above 1.1 lamda there is a significant loss in torque and increase in heat. The heat lends its self well to MPG but the torque loss was too great. Its all about the opposite of making power. Hot air, fuel heated to near vaporization but not so high that it can vapor lock a line, etc.

 

My chart is correct. It references CO (Carbon Monoxide) and not CO2 (Carbon Dioxide). Carbon monoxide levels rise when there's more fuel than oxygen, as the combustion is "incomplete". CO2 would follow a curve similar to what you're saying.

Stig, I wouldn't worry about where you're running right now. It was more of a food for thought thing. 15:1 is fine. I'm sure 16:1 would work great on our engines as well. Just keep heat in mind so you don't suddenly get toasty.

 

I use the base Cobb cruise targets 14.99ish and get about 28mpg

34 is pretty impressive though.. Would be curious what you'd see solid highway as I usually get about 32 on the highway

 

I use the base Cobb cruise targets 14.99ish and get about 28mpg 34 is pretty impressive though.. Would be curious what you'd see solid highway as I usually get about 32 on the highway

That 34 was about 200 miles of highway and 50-60 on the street if I remember correctly.

Id love to just cruise it all highway to see, is also like to have the self control to stay out of the boost for an entire tank of fuel. LoL

 

and from my understanding, going leaner or richer than stoich actually is less heat

EDIT: Here is one source among many http://www.newton.dep.anl.gov/askasci/eng99/eng99611.htm

 

and from my understanding, going leaner or richer than stoich actually is less heat EDIT: Here is one source among many http://www.newton.dep.anl.gov/askasci/eng99/eng99611.htm

I actually read that same thread there once before. I think what we are all guilty of doing is saying "lean makes more heat" because we are so used to referring to a WOT situation where leaner does make more heat because it's closer to stoich. We mistakingly apply that to cruise, forgetting that leaner in that situation means further from stoich

 

I used to reference a lot of MG carb tuning documents that are really well tested and documented. I remember the suggestion from those being not to go too lean because the head didnt handle the increased heat from day to day driving. The same documents showed exactly why header wraps were an awful idea because they removed the exhaust manifolds ability to dissipate heat and ultimately caused increased head temps.This of course is on a carbed engine so its limitations apply and to my memory there was no reference to any of the "lean heat" being wot or cruise. Thermal dynamics and efficiency should say that the less optimal the ratio, the less energy produced. Less heat. and to my knowledge for gasoline 12.5 is the peak power ratio and 14.7 is peak efficiency ratio. Pulling 10% fuel without losing 10% torque while keeping stable combustion should be winning for MPG

 

From what I understand, 14.7 isn't the peak efficiency ratio, it's the peak emissions ratio.