Discussion on how bypass valves can affect LTFTs.
 

WARNING: This is pretty long and technical, so you have been forewarned!



Ok, to start off, I just wanted to start this discussion because this is a problem that has been bothering me for a while now, and I have (what I believe to be) a good hypothesis for why this is a problem (and how to fix it, of course).

Here's the problem: You add a BPV, and next thing you know, you're running way too rich and backfiring on shifts...WTF?

Here's my hypothesis: The BPV spring is too tight is not developing LTFTs where they need to be, which means that you need to adjust it so that it's looser.



Ok, to understand what I mean, let's start off with a little background on how a stock/stock-like BPV works in our car.

BACKGROUND

Our BPV is a differential pressure valve which is arranged roughly like this:

http://img509.imageshack.us/img509/4204/bpvgy2.png

The air in this picture flows from right to left as the engine runs, and the BPV closes by using the spring in the picture to push in the direction that the double arrows point. This pushes the piston against the hole leading to the Cold Pipe and keeps the air in there from coming into the valve and getting recirculated back into the intake. By manipulating the pressures at locations 1 and 2, you can force the spring to expand or compress on demand. For example, if you apply the same pressure to both 1 and 2, the valve will remain shut, because the spring is allowed to push the piston shut. Now, if you were change things such that the pressure at 1 is lower than the pressure at 2, the piston would be sucked upwards, opening up the passage from the Cold Pipe so that air can get recirculated back to the intake as pictured. Here is what happens in practice:

When the throttle is opened, the pressures at locations 1 and 2 (numbered in red) are about the same, which means that the spring is allowed to keep the piston shut...preventing recirculation of air back to the intake. When the throttle shuts, however, a vacuum is formed in the Intake Manifold, while the Cold Pipe section is left at a higher pressure. This would suck the BPV piston upward, which would force the BPV open...allowing air to recirculate back to the intake. Translated into everyday terms: as the throttle shuts every time you shift, the 16 psi sitting in the cold pipe gets released and recirculated back into the intake in a PSSH!

Now that we have covered the fully open and fully closed throttle positions, the important part of this discussion comes into play: what happens at the various partial throttle positions?

At partial throttle, the pressures at locations 1 and 2 are different, but not THAT different. This means that there's a bit of suction upward on the piston. This difference in pressure increases as the throttle closes more and more...eventually pulling upward on the piston enough to overcome the force that the spring is pushing down on the piston with so that air starts recirculating back into the intake. This point is what I'll refer to as the Release Point of the BPV. A higher release point will refer to requiring a larger pressure difference between locations 1 and 2 before the valve pops open...whereas a lower release point will refer to having a lower pressure difference between locations 1 and 2 before the valve starts opening.

The way the Release Point of the BPV works implies that the stronger the spring in the BPV is, the more the throttle can close (higher release point) before the suction at location 1 becomes strong enough to force the valve to open and leak air back to the intake.

Ok...that's enough about BPVs for now...here's some info on fuel trims that you need to know:

Our car uses Short Term Fuel Trims (STFT) and Long Term Fuel Trims (LTFT) to adjust the AFRs to account for variations in the intake system. The reason this feature is in place is because when the ECU goes to decide how much fuel to add, it cannot trust its sensors 100%. When the ECU reads from the MAF sensor how much air has entered the system, it calculates how much fuel it needs to inject in order to achieve a commanded AFR...let's just say this is 14.7:1 at this moment. Well, what happens when it injects the amount it thinks it needs to inject, but then reads the wideband O2 sensor to find out that it came up lean and actually got a measured AFR of 15.5:1? Why was there this disparity? The answer is that sensors aren't perfect, and may be off by a bit. In order to account for this, the ECU uses what is known as a STFT to add/subtract an offset to the current amount of fuel that it's injecting to increase/decrease the measured AFR to match its commanded AFR. If the ECU keeps seeing a STFT of the same value, it saves this value into what's known as a LTFT. This LTFT is then applied to the commanded fuel injection amount to account for the inaccuracy of the sensors. After applying the LTFT, checks the wideband O2 sensor to see how close it got to its commanded AFR and decides if it still needs to add/subtract an additional STFT to achieve its commanded AFR. This allows the ECU to "learn" how this particular car works and apply fueling accurately. In a car that is operating normally, the ECU will take 20-50 miles (roughly...your mileage may vary) to learn the LTFTs that the car requires, and then run with nearly-zero STFTs most of the time. This, incidentally, is what is meant by allowing the ECU to "learn" after a reset.

There are four distinct LTFTs that I have seen our ECU to use. One or more of these may just happen to turn out to be the same, but they are usually different from what I've seen (since flow through the intake system changes under increasing flow conditions/turbo operations). These fuel trims cover:

1. Idle
2. Engine breaking
3. Light Throttle
4. Heavy Throttle (used for WOT operation)

These LTFTs are applied at these particular times in conjunction with STFTs (except at WOT...explained in a sec).

Now, you ask: why bother with LTFTs when you already have STFTs? Well, the reason for this is that LTFTs allow you to get things right the first time without having to react in order to correct a lean condition, fixing the problem before it happens! Equally important, LTFTs allow us to safely run in what's known as "Open Loop".

Our fuel system has two operation modes: "Open Loop" and "Closed Loop". These operation modes are used for different reasons...Closed Loop is basically an emissions/fuel conservation mode whose main goal is to maintain an AFR of 14.7:1, which lowers emissions and increases fuel economy...at the cost of raw power generation. In this mode, the ECU monitors its sensors very carefully and uses its STFTs and LTFTs to maintain 14.7:1 as closely as it can (with slight dips to protect the engine from damage as load varies, of course...but it always settles to 14:7). Closed Loop is usually accompanied with obscenely high spark advance (as high as 50+ degrees) and other assorted efficiency optimizations. This is the mode that we run in most of the time in our car...unless you go heavy on the gas pedal, of course, which causes the car to go into what is known as: Open Loop is a mode whose main goal is to safely achieve high power output while controlling engine cylinder temps and avoiding knock. At this point, the ECU ignores its O2 sensors and simply runs off of preloaded tables to decide what AFRs, timings, etc. that it will command in order to raise power output without regard for fuel consumption and emissions.

This is where LTFTs are absolutely imperative: The sensors are every bit as untrustworthy in Open Loop as they are in Closed Loop. How can the ECU reasonably expect its commanded AFRs to be properly achieved if it's not monitoring the O2 sensors? The answer is that it applies the LTFTs (without using STFTs afterwards) it learned while running at heavy throttle. The assumption is, of course, that heavy throttle flow and sensor characteristics mostly mirror the flow and sensor characteristics of WOT operation. This system usually works decently well and gives the ECU roughly the AFRs that it commanded.


Ok, now that the background is out of the way, let's move on to the point:

THE POINT

During partial throttle operations, the ECU is calculating LTFTs at the same time the BPV is struggling with its Release Point. Since a stronger spring will adjust the Release Point higher, this can definitely affect the LTFTs, since a stronger spring will cause the valve to stay closed at a more and more closed throttle positions.

My hypothesis is that our ECU is programmed to expect a low release point, since the stock BPV has a relatively weak spring holding the valve shut. I think that this stock Release Point is expected to be DURING the operation/learning range of the heavy throttle LTFTs. This is the same LTFT that is used during WOT (Open Loop) operation. When you add an aftermarket BPV which raises the Release Point of the BPV such that it puts it out of the operation/learning range of the heavy throttle LTFTs, which means that the ECU will learn these LTFTs with a fully shut valve (vs the slightly leaking recirc of the stock weak spring), which means that more air would actually get into the engine, making the ECU calculate higher LTFTs, since it sees more air in the system than it's used to (making it add more fuel). This illustration shows what I mean by increasing release points putting themselves out of heavy throttle LTFT operation/learning range:

http://img79.imageshack.us/img79/306...sepointin9.png

The important part to notice is that at WOT, the BPV would have been fully shut in both the stock spring and stiffer spring case, so what just happened here is that we artificially introduced a higher LTFT that the ECU would use during Open Loop operations...making the measured WOT AFRs richer. This means that when you romp on your fun pedal, your car's going to run richer than you're used to, and then backfire when you shift because it's running so damn rich. This, of course, can be fixed by simply adjusting your valve to have a lower spring tension, or tuning the WOT AFRs to command a leaner AFR to begin with.

Ok, obviously, there's going to be doubters here, but before you start ripping into me, check out why I think this is true: I've been beta testing BPVs for a few months now, and I've been through a NUMBER of spring settings in these prototypes. This has given me a wide range of observed BPV release points. Toward the beginning of my testing, I went way too heavy on my spring settings on one of the prototypes and got crazy rich AFRs on WOT along with backfires. This was on a fully stock exhaust car, so I was rather weirded out. To troubleshoot this, I went back to the stock BPV (because the BPV was the only thing I had changed), and suddenly, my AFRs were back to normal after 20 miles of learning. I swapped back to the prototype valve, and the problem came back. I swapped the spring tension (lowered it) on the prototype valve, and what do you know: AFRs back to normal again! Ever since then, I've been extra careful in adjusting my spring tension settings on these valves (before, I only played around with them to tune for pewpew sounds). I've also been a constant advocate of adjusting spring tensions to anyone I see on the forums with similar problems.






Ok...sorry for the wall of text, but I figured I would get some discussion and awareness out there about what I think is a pretty interesting problem that a number of people have been running into. Please post constructive criticism about what you think of my thought process! :D

(saved for follow-up posts)

i think what you're saying is true, but ive never noticed it affecting LTFTs

i was testing a prototype valve, so its release points were definitely out there, but i have recommended this fix to a number of people running aftermarket valves getting this problem, and it's fixed the richer AFRs and backfires for quite a few of them...:D

great post sac... ive noticed this problem ever since i switched to the hks. even in recirc..im getting rich in shifts, and immedietly into decels.

i never ran into this problem with the forge or synapse. anway..im gonna go read this again.. good stuff.

hmmm.. I wonder if a difference can be seen the forge spring shim combos, time to do some logs. nice article.

See, about time Ming started acting Asian!

if you see problems with running too tight, it will be with the red spring + 2 shims. results will vary, of course, but red spring + 2 shims is waaaay tighter than it needs to be for partial throttle operations.

if you use the blue spring + 0/1/2 shims or red spring + 0 shims, the LTFTs seem to be pretty damn well behaved.

Great read, thanks for sharing your information.

This explains the pops I have been getting. I actually adjusted my BOV a bit looser when I got home tonight. I will test things out in the morning and see if I am getting different results with a lighter spring setting.

Great post! Also explains the fuel trims well too. Never did look at what they are or do but I am understanding the usage of them now. Thanks.

Nice work. Just for clarification though, your testing this theory outside the ECU's intial 20-50 mile learning curve right?

I might have missed it but there was a lot to read.

Big thumbs up on this post!

I'm getting really high LTFT's with my Cobb SRI/Forge BPV (blue with no shims) combo.

The odd thing is that my LTFT's were spot on with the stock airbox/Forge BPV and HKS SRI/Forge BPV combos.

Do you think The Yellow + 2 shims or stock valve might return my LTFT's with the Cobb SRI closer to normal?

the popping is probably because you're running richer than the ECU is commanding. one good way to check this for sure is to simply log the commanded AFR + actual AFR. if it's obvious that you're getting way richer AFRs than commanded, then it means that your heavy throttle LTFTs have been skewed too high.

what kind of BOV do you have? keep in mind that the looser spring thing should for push-piston style valves (where the spring is pushing the piston down to shut the valve)...but i do not have direct experience with pull-piston style valves (like the HKS). basically, if loosening the spring didn't help, then play around with it and tighten it instead. if you're still not sure, go back to stock and see if that fixed things...if it does, then you're just going to have to really carefully step through a whole range of tightness settings on your aftermarket BOV.

ya, i ran the valves for a large number of miles at the different settings. unfortunately, the ECU simply cannot unlearn (without tuning) the bad LTFTs, since it does not learn LTFTs during open loop (which is where things go to shit). when i changed the spring without resetting the ECU, it took about 10-20 miles before my LTFTs settled down to normal.

keep in mind that the LTFTs are used to adjust for your specific intake system. the fact that the cobb SRI kicked up your LTFTs isn't necessarily a bad thing, because each intake can have a different baseline LTFT that it should be sitting at. the important part is that you are actually running the commanded AFRs at WOT, where your ECU isn't paying attention to sensors. are you running richer than you expect or backfiring at shifts? a really easy way to figure out if the BPV is actually affecting things is to simply go back to the stock BPV and drive it around for 50 miles. if your LTFTs drop a lot, then you definitely want to go easier on the spring in your forge BPV. usually, though, the blue spring (with 0-2 shims) works very well for partial throttle operations.

i'm running the forge BPV (blue spring, no shims) + HKS SRI and get the occassional pop or two.

i installed an AP recently, and, with V1.03 93oct Stage1+SF my AFR's are reading ~14.7 at cruise and 11.8 at WOT. occassional cuts that might be due to the cold (15*F) but future logs will hopefully diagnose that issue.

my LTFT is generally +or- 8%. my most recent was -6% when on the gas, +8% when off. is this normal? i have read that give or take 8% is okay but obviously 0% is ideal.

+/- 8%? the LTFT should stay about constant after your ECU learns it. since there's 4 different LTFTs, you might see different LTFTs flash by on the screen as you vary your throttle position, but your heavy throttle LTFTs should stay about constant after it's been fully learned. if your heavy throttle LTFTs are varying from -8% to 8%, then something whack is going on with your HKS SRI. usually, it means that the air is too turbulent over the MAF sensor and is confusing the hell out of your ECU, because the thing it's counting on is that the fuel trims will stay about constant under similar throttle conditions.

your WOT AFRs look correct for an AP...have you checked your fuel pressures?

With my Cobb SRI/Forge combo I get +7.6 at cruise and +14.2 under heavy acceleration/WOT. I've already sent logs to Cobb and await their answer as to what could be going on.

My AFR's are fine. Also, no backfires and the knock warning isn't too frequent.

i am running a greddy rs bpv and was getting popping and 10.5 afr at 5500 rpm. i talked to sac about it on saturday and he recommended loosening it. i wont lie i thought he was crazy till he explained he why he thought the stock valve did what it did. so i loosened it one complete turn, well i did some logs last night and guess what i am getting 12.2 afr. i know some might think the man is krayzie but sac helped out where no one else could, i didnt even hear from christian at cobb. so if any of you have this problem with rich afr and popping at high rpm please try this. thanks sac you are my go to guy!

im runing hks in vta, currently with like 35 miles on the reflash my ltfts are at:

idle - +16.4%
cruising - -.8%_-2.3%
WOT - -.8%


Whats the good/bad in this?! :|

Everything looks good to me except for that idle. WTF!?

My idle isn't exactly stellar either... it sits at about +10.2 once warmed up.

ya it only has 35 miles on it i think... i dont remember i drove it 3 days ago 0_o


maybe just needs some more time? and if im at 16.4% at idle what is that telling me?

I am running a Turbosmart Compact BOV. I will check the AFR's this week and see what I can find out.

if you're not running rich, then the LTFTs are just where they need to be to account for the way the air flows in the SRI. if you want them adjusted closer to 0, christian can very easily take care of that for you with just a few changes. i had this same problem, and it took christian just one try to drop them from ~14% to <5%. keep in mind, that this is different from the BPV problem, because it simply means that your baseline LTFTs are rather high (which is nothing to panic about)...the BPV problem i'm referring to happens when your BPV artificially takes whatever baseline LTFT you should have been running at and then adds like 5-10% to it...causing you to run pig rich at WOT and bang bang out the tailpipe...=/

well, simply looking at the LTFTs doesn't necessarily tell us if you are having a problem. do you run richer than you expect at WOT? does your car backfire on shifts? if not, then i wouldn't worry about it, but i suspect that you DO backfire on shifts because you're venting to air (makes you run really freakin rich on shifts).


ya, if you're running hybrid mode on the turbosmart, you're almost guaranteed to run too rich on shifts (pops and backfires). VTA is guaranteed to run richer than BPV, because you're dumping metered air to the atmosphere...less air, same amount for fuel = rich as a pig. you can either leave it like this and make sure the spring tension is adjusted so that you at least run normally at WOT and deal with the fact that you will pop and backfire on shifts, or you can go conservative and put the valve in full recirc so that it runs as smoothly as stock (clean AFRs and no pops/backfires on shifts...not moreso than stock, anyways).

well, the LTFT does seem constant for a given operating condition, but those separate values are different. i'll have to log to figure out specific values, but i saw 8% when off the throttle/engine braking.

i've been wondering about the air turbulence with my SRI - BUT supposedly the air filter itself acts as a straightener. if my LTFT values remain steady then i won't worry as much about it.

i've looked a bit at the fuel pressure and it looks about right, holding above 1600PSI and hovering around 1700PSI at WOT. no decreases upon running WOT for a couple thousand RPM's. but i haven't done any logs yet, which i would probably need to diagnose the cut - boost, fuel pressure, possibly IAT, etc.

Sacrilicious, when Christian helped you was this via the Cobb AP? I am on the stock ECU and do not own an Accessport.

If I didn't have the Dashhawk, I wouldn't even be concerned with this problem as the car runs fantastics. Chris @ Cobb did say that my LTFT's are higher than he'd like to see though.

Thanks Sacrilicious, this is some great info.

ok, it's perfectly normal to have 4 distinctly different LTFTs for the 4 operating ranges. as long as each individual LTFT is pretty steady, then that's pretty darn normal. as long as your LTFTs are below 15%, you should be good to go.

here's a common misunderstanding: my LTFTs are 10.0%...that seems high...i must have a problem!

the truth to this matter is that all the LTFT means to you is that due to the way the air flows in your intake, the way fuel injectors respond to your ECU's command to inject X amount of fuel; the system has to compensate by Y% after all these sensor/actuator inaccuracies are taken into account to achieve the AFR that it wanted to get. as long as you allow the LTFTs to develop normally after an ECU reset, you will be able to safely do WOT pulls.

where a large baseline LTFT at WOT can be scary: if you have a baseline LTFT of 15%, and you go and reset your ECU and then you don't bother taking the time to relearning your LTFTs before WOTing the shit outta your car, you will now be missing 15% of the necessary fuel to achieve a commanded AFR of say, 12.0:1. if THAT is the case, then you will run way leaner than 12.0:1 and probably end up in like the upper 12s or even 13s...that is pretty scary if you keep blasting it like that...=/ but, if you allow the LTFT to learn itself back up to 15% after an ECU reset, then it will still run perfectly normally and get you that 12.0:1 AFR that your ECU commanded.

long story short: it's nice to be able to tune your LTFTs all down to 0%, because that allows you to WOT without relearning fuel trims after an ECU reset, but even if your baseline LTFTs are rather high (10-15%), as long as you take the time to relearn the LTFTs after an ECU reset, you will have no problems running your car normally.

Mmmmmmmmmmm, Sacrilicious......mmmmmaaaaaaaaalllllgghhhhhh





Good post Sac!!! :D

sticky?

done a while ago

and no i dont backfire on shifts.. i havent driven the car much but it didn't when i ran 1-3.

hrrm...sounds like you might be lucky enough not to have issues then? :D

all this makes sense.... NICE WRITE UP SACC.........i ran the Turbosmart in Hybrid setting...excessive pops/small backfies all the time in vacuum i hated it, made the car sound like it was running bad..in boost generally ok with pops...in recirc it ran well all pops were gone in vacuum out of boost, but if it is a little too loose it will not hold boost as well so needed adjustment often to keep it right also needs to be oiled,,,sold it and put stocker back on...actually liked how car ran on stocker, quiet very little noise and very little gurgles and NO backfires...but i think mine might of leaked a little at very high boost....put HKS SSQ on. running 100%recirc...very quiet in vacuum....but sometimes when hittin it hard it would backfire in between shifts.....still figuring it out but overall the backfire issue has me, questioning its place(if i like it...not decided yet)on the car...the blow off is nice but for sure the car is experiencing some gas...its farting too much...and the HKS has no adjustment....i quess it can be tuned to run a wee bit leaner to clear this up???? or i will buy valve #4 or put the stocker back on...

Sac, I appreciate your efforts with this but I think your logic is flawed. You're basically describing a boost leak that happens all of the time. The air is being recirculated in your theory so it's already been metered by the MAF. It recircs right back in to the turbo.

Also, the stock bpv is a pos plastic hunk of shit and there is no way the ecu actually uses it to accurately calculate fueling. Sorry man, but your theory just doesn't hold.

i dont know all i know is what he suggested, loosening the screw fixed the afr i was getting that no one else could fix. and the car no longer pops at all like 0. so whatever he is its definately in the right direction or just flat out right. i dont know i just know he fixed me.

yes, it's being recirculated, but by diverting the air like that back to the turbo, you end up pulling air from the intake manifold, because what you're doing is delaying how soon that air gets into the intake manifold. just because you recirculated that air back into the turbo inlet, does not mean that the air gets into the intake manifold in time to run the AFRs that your expected.

another way to think about it is that the valve is moving compressed air back to a point before it was compressed (requiring it to be recompressed by the turbo), which would stand to reason that you're actually pulling enough air from getting into the intake manifold...corrupting what you thought you knew using the MAF sensor readings.

like i said in my post, dada, it just does not make sense that i have gone back and forth between spring settings and essentially toggled this rich/backfire problem on and off just by changing out the spring of the valve. i understand that this air is already metered by the MAF, but when you play tricks with the system by pulling compressed air and shooting it back into the pre-compression area (turbo inlet), you are definitely keeping air from reaching the engine cylinder when your ECU calculated it would be there...making the resulting combustion richer than you intended.

Put an HKS on and fahgeddaboudit.

haha...i can't argue with your there...the vast majority of people who use the SSQV v2 have no issues (assuming recirc, of course), but you might want to talk to randy about the HKS problems he's been having for the longest time...it runs he richer on shifts and decels than his forge or synapse...but i think that's just cause the valve is opening differently than the ECU expects due to the pull piston design...=/


EDIT: i just thought about this again, and randy's problem can't be as simple as a valve that's not opening fast enough on shifts and decel...that wouldn't make it run rich...=/

Now that I finally have the aem wideband in my car, I saw that I'm running rich as fuck.

Cruising: ~14.5
Half: ~12.5
WOT: 10-10.5

I'm still running the stock bpv with a SRI and stock exhaust. Will an aftermarket one help out in this situation, or make me run even more rich?