AFR Problem??
 

The Car: subaru legacy b4 be5
The Engine: Suabru Wrx STI ver5 jap import 2.0L
Mods:
-Upgraded ignition from b4
-MAF delete
-MAP Sensor
-Heat exchange sensor
-Cone Filter
-Straight pipe decat exhaust
-Syvecs s6 pnp ecu
-Equal length headers
-Aftermarket downpipe
-1 after turbo (upper downpipe) wideband lambda sensor.


Issue im having is seen in with the Lambda Reading...
The AFR seems to go eeextremely lean but during this lean period the car seems to fuel fine with no knock and no crazy over fueling, full boost, responsive etc.


I'm wondering if any one can make more sesne of the problem is it possibley a leak around the turbo area? I also today heard a noise when letting off boost similar to a kind of chattering noise quite faint, but reminisent of the sound i heard when my turbo to intercooler pipe was loosely connected. I'm wondering am i right to think this is a leak near the turbo possibly in the downpipe to turbo gasket? hence the lean AFR reading? any input appriciated.


FYI: got the car mapped around 2 years back and is running 310 on a dynapack dyno, soon after getting it back my manifold needed replacing and needless to say i replaced alot of other stuff too and only getting it ready for road now and noticed this issue. still not crazy familar with the syvecs software so have uploaded a photo.


as its an on road map its done over time with a seperate graph for rpm rather than run along the rpm.

Welcome to scooby net gaz
Hopefully someone will come and help you out

Doesn't make much sense. Most wide bands read between 7:1 and 22:1 so don't see how you can see over 22:1. I don't know syvecs but I'm thinking that you've possibly selected the wrong AFR graph. Not sure what it means by Lam1(AFR(P))... the (P) is confusing me... Is there any other AFR graphs like Lam1(AFR(V))?

so lam1 is just lambda 1 which is the wideband lambda its taking the afr from after the turbo on the top of the downpipe... and afr(p) just means its showing afr relevant to petrol engines... there is also the lambda voltage reading but this is running in tandem with the lambda1 final reading. as far as i can tell so far there is no afr reading on the syvecs, just a case of changing the lambda to afr petrol. lambda voltage Wasn't logged in these runs.... the afr also fluctuates a bit at idle going from 13 to 15 afr constantly up and down and will head towards rich more often than not.... if thats adds anything to it.

Oh is that all the P means lol fair enough. The fluctuating at idle would be because its running closed loop lambda control at that point wouldn't it. Its the weirdness going on where the values go well over 20:1 that doesn't make any sense. Logging the lambda raw volts would help determine if its a bad sensor, interference or if something in the setup/ecu is playing up but I wouldn't run in it unless you can be 100% its not actually running lean.

Well you were right... Contacted the buddy that mapped his own hawkeye using syvecs... He said not to ask why... But you have to ignore that lam1 reading... Go to clltarget and that gives you the afr and you can then change to lambda units if you want... So I did this and it is indeed running pretty normal and as expected everywhere and will hit 14.7 when held at low revs and then at idle 14.0... Is the 14.0 normal at idle? Or should this be hitting a little leaner?

at idle you want to be hitting high 14, low 15 but also keep an eye on how much timing your allowing to hold idle as you don't want to run too much there.

The NTK sensor (assuming that's what you have on syvecs) will read from about 8 to 30:1 AFR on petrol.(AFR is dependent on fuel type, lambda is always lambda hence the P for petrol vs Diesel or LPG Gas which has a different stoichiometric ratio).

Are you using the full wideband control if so how is your finalfuel1 and lambdaerror look?

You should be able to save/export the log then anyone with the sview app can read through them.

How far back from the turbo is your wideband placed? Sometimes depending on placement you can get a rough reading (assuming it's not just an exhaust leak and it's sucking air through). Whereabouts in ireland are you, if your nearby i'd be happy to take a look some time.

also just reading your last post there sorry the CLLtarget is the fuel target and LAM1 is the actual fuel reading so you should consider both and adjust until the lambda readings match your target value so the correction doesn't have to swing so much to keep things in line. You can log the correction value and apply average figures to the fuel table over time to get it running a little better.

If you only go by the closed loop lambda target then your actual fuel reading could be quite different which could result in being too lean/rich.

Giving it a final look over today in work... For leaks etc... And going to log the trip home. Will get the full log uploaded this evening as that will be the easiest! You'll have a better idea of what's happening :)