[serfma]

I am running a 96 Mustang Cobra, 4.6L 4V, CDAN4/LLX3. I am painfully new to the tuning world. I am curious since I've been unable to find the answer of: What effects do long tubes have on how rich / lean the car will run? My issue: Running past 3k - 4k RPM in 1st (haven't tried other gears) results in black smoke. I've tried data logging at operating temp for idle since I've no wideband and it's looking like I'm overall running lean, seemingly more so on the driver side bank.

Mods:
- BBK Long Tubes
- No cats
- No rear o2's

(No wideband atm)


In my tune I have changed the following:


Should I increase my o2 transport delay? Force open loop at idle / below a load %?

Any help or advice will be appreciated :)

[ChsReb]

I'm not acquainted with your tuning software, but going back to the days of Carbs and Headers, I can say for sure Headers will lean things out. Is the Smog pump still there and working? Is EGR still there and working? The computer needs to know if things have changed.

Check your MAF. Lots of discussion here on changing one. I've got several videos on YouTube as to how I adjusted mine. I'd spend lots of time checking that. Now if your graphics are accurate, you're extremely lean down below and improving as load increases.

That screams MAF or ITHBMA to me. Check and dial in ITHBMA if you haven't already. Failure to do so creates a idle-cruise lean condition.

[cgrey8]

The best piece of advice I can give you is...umm get a wideband.

It appears you do have BE and EA so the next thing you could take a peek at is if your Injector Parameters look reasonable based on your datalogs. There's a writeup somewhere (possibly in the tech docs section) that talks about how to analyze your datalogs and tweak your Injector parameters. While it is possible to do this without a wideband, the results are far better and quicker with a wideband.

Once they are "close enough" then tweak the MAF curve to get it whipped into shape. However there's only so much you can do here without a wideband...but even more-so for MAF tuning than with Injector parameter tuning.

With those two dialed in, you shouldn't have rich problems.

Now to your point about long-tubes. What they do is accentuate scavenging of the exhaust from the cylinders at moderate to high loads. Combine them with a heavy-overlap cam and you could get scenarios where your exhaust appears lean to sensors due to unburned air getting into the exhaust, but shows black smoke out the tailpipe thus suggesting things are much richer than they otherwise would appear from datalogs. When using HEGOs, any significant unburned air getting into the exhaust will present as a lean condition causing your computer to adjust rich (during closed loop modes). Based on your KAMRFs adjusting for a detected-rich condition, I'd suspect this is NOT the problem you are fighting even if this problem actually exists. I suspect there's other problems with the tune right now that are worse than the affects of blow-thru due to a heavy overlap cam and long-tubes.

Here's where having a quality Wideband comes in. What QUALITY wideband controllers can do is tell you the ratio of air and fuel in your exhaust (burned and unburned). If the wideband controller is quality, it can compensate for the unburned air by also sampling the unburned fuel and reporting the ratio of what the air-to-fuel is in the exhaust. This is one of those areas where cheap wideband controllers are inadequate. Cheap widebands work fine for a mildly built engine where the vast majority of what's in the exhaust is burned air and fuel where the excess being measured is either an excess of oxygen and very little fuel (lean burn) OR an excess of fuel with very little oxygen (rich burn). But when you start introducing a significant portion of both unburned air and fuel, this is where they start not working well and their accuracy is questionable to down-right unreliable. So depending on how crazy your build is will dictate whether <$180 WB will be fine or whether you should not skimp and budget for a quality ~$400 WB controller.

This of course also assumes you don't have exhaust leaks. Exhaust leaks introduce oxygen into the tubes that did NOT go through the engine. So exhaust leaks will cause a leaner-than-actual measurements under conditions where the headers are creating a vacuum in the exhaust and literally are sucking in outside air past the leak. So if you have exhaust leaks, you do not have useful data from your exhaust sensors.

Those are my thoughts....

[034v]

HEGO Delay or FN1351

This is a delay in revs from the O2 sensor that the engine fueling will see that change.

For long tube headers and some turbo applications, these numbers should be raised.

Table is at 0x12B8E in CDAN4

[serfma]

HEGO Delay or FN1351

This is a delay in revs from the O2 sensor that the engine fueling will see that change.

For long tube headers and some turbo applications, these numbers should be raised.

Table is at 0x12B8E in CDAN4

When you say these numbers should be raised, I have no concept or baseline of what would be too high or too low of a change. Some reading around seems like people generally blanket 20% - 25% raise it across the board. My next question is, with it being raised, what am I looking for in terms of changes in KAMRFs (or elsewhere) to indicate whether it was a good or bad increase?

With the exhaust leaks fixed, I am seeing around a 14% idle lean condition on the passenger bank which I am unsure if this is normal due to LT's or another exhaust leak somewhere, or possibly bad O2? As for my driver side bank I had a leak at the two small EGR tube vacuum lines that caused it to reach 25% lean, to which I've since plugged the lines and idled around 10 minutes and had it drop to 20% but am unsure if I need to go ahead and reset KAMRFs on that or if the ECU will with some time adjust down.

[skunk]

I would leave the delay alone for now...it just makes the HEGOs a bit sluggish to respond.

As stated...you need a wideband to know how much error you have between what the computer thinks is stoic and actual WB results otherwise you are flying blind.

If you wish to do what you can without a wideband....reset KAMs with every change you make and get your fuel close by adjusting injector parameters first then attack the MAF for fine tuning.

John

[serfma]

I would leave the delay alone for now...it just makes the HEGOs a bit sluggish to respond.

As stated...you need a wideband to know how much error you have between what the computer thinks is stoic and actual WB results otherwise you are flying blind.

If you wish to do what you can without a wideband....reset KAMs with every change you make and get your fuel close by adjusting injector parameters first then attack the MAF for fine tuning.

John

Hmm.. Yeah getting a wideband ASAP. I just wanted to rule out any possible issues before starting to tune anything. With driver side having a much higher % of leanness, could this be possibly a symptom of a faulty injector or two?

[skunk]

Once you have ruled out all possible intake and exhaust leaks, I would verify that your base mechanical idle is set correctly. If the throttle blade is open too much or too little the ISC routines will only be able to compensate so much. You are basically stock and the tune should be close as is. The black smoke could be due to the stored closed loop enrichment carrying into open loop.

A suggestion if its available in your strategy....turn off adaptive learning (Kamrf) and tune via LAMBSE in CL. It will speed up the process by giving you the instantaneous correction. Make sure KAMs are cleared when you do this or they will stay in the mix. Just remember, if LAMBSE is commanding lean you are rich by that percentage and vice versa.

John

[assasinator]

yes adaptive can be turned off in cdan4. you must do it for sanity until its close.

plug up all egr in ur headers. turn off egr in your toon.

you cant toon it with exhaust leaks. the pcm will wander endlessly. either get a pair of widebands, or take it to a dyno and get them into the tailpipes. aftermarket adjustable FPR and a gauge to see it.

have you cleaned you IMRCs? are they working? as others have said, you need to get your injector slopes to match your new conditions, then adjust the transfer function. if its mild you could just toon the maf.

you also need to take care with the load and the borderline spark table to avoid hurting it. if you have IMRCs then stocking timing is close, if they are gone, timing is not close.

nothing is simple with our early 4Vs.

[serfma]

yes adaptive can be turned off in cdan4. you must do it for sanity until its close.

plug up all egr in ur headers. turn off egr in your toon.

you cant toon it with exhaust leaks. the pcm will wander endlessly. either get a pair of widebands, or take it to a dyno and get them into the tailpipes. aftermarket adjustable FPR and a gauge to see it.

have you cleaned you IMRCs? are they working? as others have said, you need to get your injector slopes to match your new conditions, then adjust the transfer function. if its mild you could just toon the maf.

you also need to take care with the load and the borderline spark table to avoid hurting it. if you have IMRCs then stocking timing is close, if they are gone, timing is not close.

nothing is simple with our early 4Vs.

IMRCs are clean and working.

Thank you for all of the info thus far, it is very helpful. At the moment, bank 2 is showing more lean than bank 1 by roughly 6%, I assume this indicates a leak is still present and it needs to be almost identical to bank 1 or is there something I am unaware of that could cause uneven banks?

Was going to attach log but it's too large (5.4MB) to attach, so here's a link to it: https://onedrive.live.com/view.aspx?res ... -rIIbihOG4

I have everything capped off driver side, just had my exhaust welded up and all of the pinholes covered up from the last shop's shoddy job, reset kamrfs and this was a short idle log while I was in a drive thru already up to op temp. The tune differences compared to stock are the following:


To disable EGR, COT since at WOT a ton of fuel being dumped and I have zero cats, no rear o2's, and I hadn't reverted changes back to stock for HEGO transport delay as I've not had time to get to tuning and I wanted to make sure the differences between banks was or wasn't a big deal so I didn't waste hours into tuning just to find out I have to narrow down more exhaust leaks lol

Planned on getting a wideband, but now I'm wondering if I need to go ahead and get a second to cover both banks as I seen a lot of people using or recommending one.

[serfma]

Ordering a wideband, coming in Wednesday!

When long term fuel trims are showing 20% (richening) is it fully compensating and thus not lean? Or will this need to be checked with a wideband? I've had a slight heat issue, 10-20f above other stock / near stock Cobras and constantly sits around 208f which is low speed fan on thus on 24/7.

[skunk]

You are still lean which is likely contributing to your heat issues. The ECU should not have to compensate that much and is likely pegging the limit of compensation it can provide.

Either reduce your injector slopes or increase your MAF curve by 10% as a start to see if things improve.

John

[serfma]

Thank you skunk.

To my understanding, 25% is the hard limit which throws a code past that amount, and since at idle / low load I am under that 25% threshold I assumed I wouldn't be running lean since it's compensating since it's not throwing a code / past 25%. Tomorrow my wideband arrives so I'll be able to measure AFR. So I must have a misunderstanding still of things, as I thought if it was showing anything below 25% and it's goal is to reach 14.6AFR, that it is reaching that amount, thus is not running lean. I have a ton of research and YouTube videos to watch to learn. Lol

[skunk]

You've got the basic understanding... look at lambse (short term correction) if its still commanding a rich mix in combination with Kamrf (stored correction) you are lean.

The problem with allowing such high stored correction is the ECU has to fight it to get the EGO to switch the other way. The stored correction is also carried into openloop which may cause an overly rich or lean condition depending on whats stored.

Best to turn off Kamrf and base your adjustments on what Lambse is commanding to get your closed loop fuel as close to stoich as possible. This way when you do turn on long term fuel correction....it will have minimal influence on your tune.

John

[serfma]

You've got the basic understanding... look at lambse (short term correction) if its still commanding a rich mix in combination with Kamrf (stored correction) you are lean.

The problem with allowing such high stored correction is the ECU has to fight it to get the EGO to switch the other way. The stored correction is also carried into openloop which may cause an overly rich or lean condition depending on whats stored.

Best to turn off Kamrf and base your adjustments on what Lambse is commanding to get your closed loop fuel as close to stoich as possible. This way when you do turn on long term fuel correction....it will have minimal influence on your tune.

John

I understand much better now!

Looking through BinaryEditor (CDAN4 / LLX3, I forget what terms were what) I can see what appears to be a low / high limit for fueling correction with adaptive learning? I'm unsure if this is correct or not, but I would assume it means the maximum AFR that adaptive learning can change by, which is up to 10% rich or lean?

[skunk]

It does appear that there is some sort of limit on how much LAMBSE (short term fuel trim...not adaptive learning) is allowed to swing. What is strange is the numbers don't correlate to the High/Low AF Ratio Permitted. Check your logs and see if LAMBSE ever commands higher than 16.xx or lower than 13.xx AFR to confirm your thinking.

It might be hard to see with a 20% correction which is equivalent to an 11.7 AFR from adaptive learning already in the mix keep you close to stoich. Either way....you need to address the lean running and get your tune much closer to stoich without compensation from adaptive.

John

[serfma]

Wideband all hooked up, connected into pin 47 which is EGR Vacuum Regulator Solenoid. CDAN4/LLX3 (96 Cobra) Can't figure out what EGR tag to use, got transfer funcs set up but they're either maxing out or at the very minimum. Can read the voltage at the signal line.

Edit: disregard, found a serial to USB cable

So -- I'm looking at tuning MAF curve using wideband with KAMRFs disabled to get as close to stoich as possible in closed loop conditions, correct? As far as open loop / WOT what's the suggestion there?

Aiming at closed loop / idle first so I can get a grasp of the concept. Any good write ups / videos please do let me know. Trying to research and learn as best as possible.

[cgrey8]

...So -- I'm looking at tuning MAF curve using wideband with KAMRFs disabled to get as close to stoich as possible in closed loop conditions, correct?...

That's the intent is to get Closed Loop conditions as close to stoich as possible. While that sounds simple, it has its challenges, the main one being dealing with transients. The AFR that's produced at stable state and the AFR that's produced as you transition from one load to another are not the same. And often the source of driveability problems are related to over-enleanment during transients. Being a little overly rich, no big deal. Being a little lean, and you get intermittent stumbling, bucking, or surging. It's particularly annoying when you are on the cusp of the engine's tolerance and what often causes the surging is one cylinder producing power despite the condition, and the next cylinder failing to produce power. Those rapid power/no-power conditions happening at the speed that cylinders fire causes the surging and bucking which is particularly annoying on a transition from idle to take-off. Part of what makes the experience so annoying is that this is happening at relatively low RPMs, so the time between cylinder fires is long enough that the rotational momentum of the engine's components is not nearly enough to smooth it out like it would be if this same phenomenon happened at higher RPMs.

The other challenge that some people face is where adjustments to the MAF curve improve some RPM/Loads, but make other RPM/Loads worse. This is usually a sign that your Injector Parameters are off. Read more about that here:
https://eectuning.org/forums/viewtopic.php?f=17&t=9395

...As far as open loop / WOT what's the suggestion there?...

Most people find tuning WOT much easier than any other tuning IF you have a working and reliable wideband. RPMs are high. Transients are not nearly as significant here because the whole WOT experience is basically a huge transient. And WOT tends to be much more predictable and reproducible. So the philosophy here is OVER-compensate in the areas that are detrimental to begin with. Set the AFRs WAY too rich and work your way toward the desired AFRs.
Ditto for spark...run conservative spark values during the AFR tuning period...and then start gradually advancing to pick up more power.

N/A engines are far a bit more forgiving than boosted engines. Running too lean or too much advance on a high-boost engine can be a quick death.

...Aiming at closed loop / idle first so I can get a grasp of the concept...

Ironically, you are probably starting with the hardest part of tuning first. Although admittedly, tuning this area is going to be less damaging to the engine if you get it wrong. Most people "save" idle tuning to the end because it can be so challenging and in the meantime, set a high enough idle (either in the tune or with the idle screw on the throttle body) that you overcome the shortcomings of the tune.

Running an engine lean and/or overdoing spark advance aren't as critical to the engine under idle and most driving conditions. But what is detrimental is when the engine misbehaves in traffic. So if you are going to attempt to tune the engine on the road, do so on an open road with little to no traffic. As a novice tuner, I've had to push my truck out of traffic because the engine just died in the middle of traffic because something went awry with my tune and I couldn't fix it in the middle of the street. Don't attempt to change/update tunes while driving. Only do that in a parking lot somewhere when you are sitting still. Even if you have a person sitting in the passenger seat making the changes, don't attempt to make tune-writes while operating the vehicle.

[serfma]

Appreciate the time you took to write this! I'm definitely keeping to parking lots to change tunes and run changes on back roads, works perfect in Kentucky.

For injector slopes with my current situation of stock everything besides long tubes, deleted EGR and cats, would it be necessary? Seems almost as if it's a "change these if you change injectors only" thing but of course I have a feeling that I am indeed wrong. I did read over the injector slopes thread of yours, but I am sure I need a few more read overs of everything I've read from the beginning.

Is there any Discord / IRC that I can ask questions? Sometimes it's difficult to explain a question over a forum post, and I don't straightforward want my hand held or someone to do the work for me, I want to certainly learn but I am trying to grasp the whole concept better.

Now with that in mind -- at idle (just using as an example though you say it's the hardest which I completely understand why it would be now) and closed loop at all conditions we want to aim for 14.64 AFR or very close (within 3% - 5% both ways, 15.37AFR lean, 13.94AFR rich) at all times. After turning off long term trims and logging idle at op temp (200-210) I was watching AFR and seeing around 14.7 with LAMBSE1 at a 16.0AFR, LAMBSE2 at a 15.3AFR. (Going to swap my O2's then Injectors to see if I have an issue with either bank as for why one is reading much leaner although I've tested for vacuum leaks mercilessly.) I know that LAMBSE is the AFR the ECU is commanding, but is it demanding the fuel mixture to be that of 16AFR to reach the intended goal of 14.64, or what? Trying to get an understanding exactly the car/ECU's "thought process" to get a better idea.

Here's the trimmed down log of idling at operating temp while in a drive thru, if it helps any to explain if it's rich / lean and why or what to look for. Couple times of going into first gear to drive forward which I assume would be considered partial throttle / transient fueling that'd be applied there.
https://drive.google.com/file/d/1arHgIR ... sp=sharing

Also, what all "conditions" are there? So far what I've found is to be: idle, partial throttle, and WOT that I've found that people focus at tuning at a time independently of each other. Of course I believe idle / partial throttle is still "closed loop" and WOT is obviously open loop.

Thanks a ton!

[cgrey8]

...For injector slopes with my current situation of stock everything besides long tubes, deleted EGR and cats, would it be necessary? Seems almost as if it's a "change these if you change injectors only" thing but of course I have a feeling that I am indeed wrong...

This is a point of disagreement amongst the tuning community. Some people believe when you have injectors that have published info, you enter the published values and forget about it.

My attitude is even if you have the numbers, what's the harm in affirming that they are behaving as advertised by checking? And if their behavior in your application is different than advertised, for whatever reason, then I wouldn't be afraid to change the values based on what datalogs are indicating the settings should be.

...Is there any Discord / IRC that I can ask questions? Sometimes it's difficult to explain a question over a forum post, and I don't straightforward want my hand held or someone to do the work for me, I want to certainly learn but I am trying to grasp the whole concept better...

This is an interesting question. And I suspect as more and more GenZers start getting into tuning, you'll see more of that. But with most of us being GenXers and older, we tend to prefer forum-based interfaces more.

...I know that LAMBSE is the AFR the ECU is commanding, but is it demanding the fuel mixture to be that of 16AFR to reach the intended goal of 14.64, or what? Trying to get an understanding exactly the car/ECU's "thought process" to get a better idea...

Closed Loop and Open Loop thinking are, in some ways, backwards. I'll start with Open Loop since it's easier to describe. While in Open Loop, LAMBSE is the commanded value. You measure what the actual AFR is and you, as the tuner, have to modify the tune so that the tune is actually attaining the LAMBSE value on the WB.

Now Closed Loop adds a wrinkle to this. LAMBSE is still the "commanded" AFR. However because the EEC is responding to sensor feedback, it is changing the LAMBSE to attain stoich as detected by the HEGOs. So if the engine is running lean, then it will adjust the LAMBSEs richer in order to enrich the actual AFR to get the mix closer to stoich. The LAMBSE is still the "commanded" AFR. It still represents the AFR the EEC believes should've been the AFR given the amount of fuel that was being delivered. And again, as the tuner, it's your job to change the tune to get the tune to deliver the correct amount of fuel to get the LAMBSEs closer to actual so that it isn't having to adjust so much to attain stoich.

...Also, what all "conditions" are there? So far what I've found is to be: idle, partial throttle, and WOT...

I'm sure these are up to interpretation and there's probably a few more than what I'm thinking of, but here are the "modes" I think of:

Startup conditions:
Cold Crank (starter turning over the engine)
Cold start (moments immediately after cold crank, always open loop)
Warm Crank
Warm Start (always open loop)
Hot Crank
Hot Start (always open loop)

It's generally not recommended that you go crazy on the engine during cold and warm running conditions. You really want the oil warmed up enough that it is capable of protecting the engine under high load. When oils are colder, they are thicker, and thus are more susceptible to fluidic sheer. When fluids sheer, they literally tear away from the surfaces they are on and thus expose the two moving surfaces to each other. Many people are not familiar with what fluidic sheer is. The best way I can explain it is next time you get a chance, chew a REALLY cheap piece of bubble gum until it turns to a hockey puck and hurts your jaw to bite down on it. If you pull on the gum slowly, it'll stretch. However if you grab it tightly and pull it with a huge amount of force, it literally breaks or snaps apart and the breaking surface looks much like sheered metal. Oil does something similar. When the surfaces that glide the oil move the oil too fast, the oil will tear much like the gum "tears" if you pull on it too fast. And obviously higher quality oils will resist tearing more than cheaper oils. But regardless, it's best to just be easy on engines until they are warmed up.


Relatively temperature stable conditions:
Idle in neutral
Idle in drive (most relevant to automatics)
DASPOT (drop from higher RPM down to Idle, and most significant in manuals)
Preposition (high-hold of Idle RPM while vehicle is in motion)

Transition from part throttle to decel
Decel
Transition from closed throttle to part throttle (aka tip-in)
Stable Light Load
Transient part throttle changes in load
Stable Cruise Load
Stable Acceleration Load (moderate to heavy acceleration, but not WOT)

WOT

Then if you are into drag racing, you have trans-brake staging at WOT but stock EECs don't do this well. You usually install external devices that do fuel or spark cutting to chatter-box the engine while in a staging lane.

Each one of these conditions are conditions that may give you trouble at some point or require you to recognize as the source of some driveability problem.

[serfma]

Shoot yeah I have a lot of learning to do. Lol. Tackling some other issues with the car that I think I need to squish before I even look to begin tuning. Lately I have had around 5k RPM+ a mixture of either completely losing all power, to which I tested while logging and found AFR jumping to 20+. Everything looks "normal" to me except I see VBAT dropping to 12V during a pull and I am curious what exactly VBAT is measuring voltage at, because while it reads 13V I will read 14V on a multimeter at the battery terminals. Any ideas?

[jsa]

Vbat is the voltage at the EEC internally.
Voltage drop in external wiring will reduce measured voltage below battery terminal voltage.