Tweecer/Tuner guys, help with A/F at idle

Discussion in '1994 - 1995 Specific Tech' started by zenboy99, Jan 19, 2006.

  1. Just installed my wideband O2 sensor. At idle in closed loop after a 5 min warmup its reading 17-20:1

    Lambse values are at 11.7-12.0:1, the O2 volts are really low (0.03 sometimes stuck at 0.003) and the car stinks.

    I'm assuming that maybe the O2's and wideband are too far downstream with my Long tubes, but if they're cold shouldn't they be reading rich not lean? The wideband sensor is in the x-pipe directly after the collector.

    I'd love to drive it but I've got the car up on jack stands and have a few more things to do. When I rev the car up to 2000rpm (still in CL) the AF goes to ~15.5:1 and lambse reads ~13:1 wich seems about right so I'm assuming once I get the car good and hot driving along it should be accurate.

    If my A/F and lambse values are off at idle, do I just tune this thing by smell?

    Also, I don't have the noise filter capacitator in yet between the ground and wideband wire since I don't have a solderer, but that shouldn't effect the overall accuarcy of the reading.
  2. sounds like you need to work on your MAF transfer in the lower areas...

    also did you just un-hook your batt. if so you may need to wait for the eec to settle down a bit.

    also look at finals website and search his name...if your running larger than 24#inj. you may want to look into forcing CL at idle.
  3. I spent 2 weeks getting the MAF transfer right from idle to ~3.5 volts. I would datalog it and have EEC analyzer calc it over 15 different drives into work. 10 minutes of idling at each drive. Maybe my O2's are Bad?

    I know Grady and he's helped me alot. I have a crazy idle surge in Open loop that stops at closed loop so I'd like to keep using closed loop unless I can solve that problem.

    I'm just wondering why I'm reading lean when my car smells PIG rich, I mean I have to wash my clothes after I let it idle for awhile.
  4. Idle issues can be frustrating :bang:

    The smell can be misleading however.

    At idle, a bit lean is not gonna hurt anything
    I have found my car likes things a bit lean as opposed to a bit rich

    Speaking of rich ... If you got some surging at idle that is almost always a dead give away things are too rich.

    Lets look at two things you reported.
    1) WB values in a range of 17.00 - 20.00 to 1.00
    2) LAMBSE values in a range of 11.50 - 12.00 to 1.00

    1 & 2 go hand in hand with each other as they should.

    WB shows lean so LAMBSE (commanded ratio by pcm) shows fat to try and compensate in an effort to find 14.7 to 1 :)

    What are your KAMRF's doing at idle after ect's have gone normal?

    I'm with Greg about the maf points around idle ...............
    Juice em up a bit :D

    btw, if you suspect your O2's are cooling off a bit you can read a good bit about it on the Tweecer site and Daren talks a bit about it on his site at

  5. OK, the MAF was dirty, got some electronics cleaner and the signal is much better, but I'm still reading lean, about 17:1. So I'm going to have to go datalog some more MAF functions. This took my 6 hours to figure out. Some crap got sucked up into my MAF that I didnt' see.

    My question is though, how does the EEC calculate the Kam's? Does it use a specific ect? I've been idling for well over 1 hour so far in the last 2 days and the kams for idle haven't changed. It takes my car a good 10 minutes to get up to 180 degrees with the cold temps out.
  6. Might wanna do a pcm reset since you found that prob with the maf.

    You will have get to normal temps or close to them for you to start operating in CL which is needed for the K's to become active.

    As for how the K's come into play, it is the data from the O2's and the narrow band of operation they work in. If you move from that narrow band the adaptive starts to add or pull fuel with a change in pulse width.

    After a pcm reset you see K's at 1.00 even though the car is in CL. That is because the pcm is starting to learn from fresh. Over time you will see the K's start to move. When you have hit all the different points enough times the K's will stabilize. It is between those two times the K's will be most active and you can literally see the adaptive strategy at work if you watch your data log in real time.

    Not the best way to explain it Chad but I really don't know how to do it any better.

    It is just what I have seen in my dlogs :shrug:

  7. Is this with a modded MAFtransfer? If so you need to un-do your batt. to clear the KAMFR. It stands for Keep Alive Memory Fuel Ref.

    Your going to have to datalog that MAFtransfer and adj. it little by little until you get it as close to 1.0 (realy will get to be like .9998 or so). Its not going to change if you dont tune the low area of the MAF.

    Sorry if I come off :nono: or :shrug: ...just wanting to make sure your aware you need to datalog and tune the transfer to change the readings your getting. If your not making any changes then why would it change
  8. CL is coming on at 50 degrees ECT believe it or not. How much time do you need for KAM's to start to change?
  9. No worries buddy, I appreciate your help alot.

    I sucked up some dust crap that was lingering around the garage last week. So I still think the MAF is pretty close. I did some datalogs tonight and I think its pretty close.
  10. most say to drive around for like 10-20 min.

    I do a drive loop that is stoplight the hwy and loops back to my house. It has some hills, hill stops, a light that is ALWAYS backed up so its perfect and takes 15-20 for a round trip.

    then come back and run a datalog.

    Or I will do the loop and then datalog on my way out later that day or something.
  11. To put it another way, KAM values are calculated based on the A/F that the lambse is calling for. If the value is less then 1, then you're rich and its pulling fuel. If the value is more then 1, then you're lean and its adding fuel. The KAM values will continue to change until the EEC finds a happy medium around 13-15. This is why its important to clear your computer EVERY time you make a change because you don't want previously "learned" KAM values influencing your new tune.

    I've read some guys say that you want the tune to be dead nuts and not need any adaptive KAM strategy. With that being said, I feel the KAM numbers to be CRAP and that this method isn't right. I think the best indication is to clear the computer then fire it up, and pay real close attention to early LAMBSE values. This will tell you right away whether you are rich or lean. I personally don't rely on the KAM values to tell me anything, instead I trust the W/B and LAMBSE.

    Post up your TwEECer tune. If its ver. 1.30 I'll be happy to take a look at it.

  12. My question is how long it takes the KAM values to kick in. They've been at 1.00 all week idling
  13. I realy dunno...might mean the o2's are dead/not switching...when you datalog how are the two arrows on the o2's showing? Are they moving back and forth all the time or just DEAD stable.

    Preaty sure the stable means a dead o2 sensor.

    Might be wrong I have never seen what a dead o2 looks like while datalogging.

    Try driving it around alittle too.

    WB's are deff. the way to go to be certain about that stuff. Alot of people have gotten preaty close without it and EA though then hit up a dyno to be sure.
  14. As with most things, you can get them done several ways.
    Self tuning is really no different I reckon.
    So, I'll give you my take on CL tuning.

    To me the K's tell the whole story and I focus on them more than Lambse or WB values.

    The major thing the pcm is trying to accomplish during CL conditions is to give a ratio of 14.7 to 1 and it uses the data from the O2's and other sensors to determine how to arrive at that goal. If the sensors are in good working order then the whole deal is on auto pilot for the entire working range of CL conditions. Things should be pretty accurate cause we are only talking about a narrow band of operation which the O2's were designed to operate in.

    From idle to maximum conditions for CL operation covers a great deal of territory and if a few points here and there are not quite right ... well, that is what adaptive strategy is all about.

    Larger inj's, maf, or a number of different things can cause the adaptive to be outta whack.

    To bring it back in line you can adjust
    inj values
    maf curve
    batt offsets
    and more

    Take idle conditions for example

    1) All the airflow values like tb airflow, neurtal idle airflow, and such can throw the K's off.
    2) A set of pulleys will even throw the K's off cause the batt volts dip which throws your inj offsets off.
    3) I saw my K's change when I installed my home made cai.

    Lots of things might throw the adaptive outta whack but hey .....
    All you gotta do is look at the K's and they will tell you whats wrong :D

    IMHO, the most important thing about CL tuning is to have a maf curve that is close enough to give pretty stable conditions to start tuning from.

    You can then use your inj values like slopes, offsets, and such to get most of the CL conditions to a point that the adaptive is not working very much or the K's are hugging the line (1.00 value) as they say. You may have to make a slight adjustment to a maf point here and there for fine tuning.

    Two things that are critical for efficient CL tuning are
    1) pcm reset after a change
    2) Don't work with more than one method of adjustment at a time. By that I mean don't work with the slopes for a while and the maf curve for a while.

    As for the amount of time it takes the K's to change

    At normal ect's and after a pcm reset I would drive around for about 10 mins before I saw the K's start to change. At that point I made sure I put the car in conditions that covered all of the maf volt range from idle to 2.6 volts (my highest CL maf volt value) for 15 mins. I then started to dlog the next 15 mins for data to see what the K's were doing. As I got closer and closer to having the adaptive be correct, the amount of time from when I did a pcm reset to when I started a dlog needed to be increased to gather accurate data from what the K's were doing because they were now not as active.

    Hope this stuff makes sense :shrug: or helps some of you :D

    Its kinda hard for me to explain :bang:
    if it doesn't make sense, feel free to point out where I confused the issue and I will try to make it more clear :)

    I assure you, it took me a while to see what the pcm was doing and I still wonder at times if I'm on the right track :rlaugh:

  15. Try adjusting the exhuast pule delay table. It's supposed to help with the O2 problems with those LT headers.
  16. Greg

    Your asking about a dead O2 reminds me of something that happened a while back and I thought I would share it with you and the other members as the info may be of some help.

    Long time ago I played with different kinds of ways to analyze the data from my dlogs.

    The playback thing in CalCon was just too limited to my way of thinking :(

    I got so much more from looking at the raw data in Excell when I quickly wanted to focus in on a small part of the dlog.

    I then discovered Log Analyst and quickly saw many advantages it had over Excell for seeing the big picture :)

    Just a few of its Advantages

    1) Column headers always in view when working with raw data
    2) Easy delete of raw data events before/after your desired range of data
    3) Graph allows 1 or more payload options to be displayed
    4) Graph allows a conversion display formula that can easily be applied to selected payload item .... such as hegos.
    5) Raw data & Graph comparison of two or more dlogs
    6) Many more still, but by now, you can see I really like LA :D

    Item 4 is where the dead O2 came into play and was quite helpful.

    When I started to self tune I loaded my hegos in LA and used the conversion formula to display the full 0 to 5 volt range of their operation. I quickly saw one hego was a bit more active than the other. Many of the other guys were talking about seeing similar results. It didn't take too long for us to discover that a difference, if not too great, was nothing unusual. After that discovery, I rarely include hegos in my payload. As long as they are active and switching, they can be considered good to go. Long time goes by and I don't even think about hegos


    I get a request for help from this guy and his data looks like adaptive is all hosed up big time. He reports of CEL's and then a noticable loss of power.
    First thing I say to him after looking at the data is I suspect one or both O2's may have a prob but he didn't have them included the dlog he sent me. He tells me has asked others for help and changed all kinds of stuff including the O2's.

    I tell him to send me another dlog with the hegos in the payload.

    They looked like 1/16" wavey lines on the screen :eek:
    Normal hegos will cover like 75% of the screen :rlaugh:

    Maybe 5% of his entire dlog had some hego activity but at the very most, the range of activity was like ...... 2" in display size :(

    I used the conversion formula to reduce the display size and the tiny wavey lines looked much larger which in turn reduced the values on the side of the display (xy axis thing).

    I worked with him for a while and kept telling him something was wrong with his hego output which kept the adaptive strategy from doing its thing. I finally told him to get the guys from the various tuning sites involved and they came to the same conclusion.

    He contacts me a good while later to report he had blown fuses to the heater circuits on both hegos. :eek: :( :bang:

    I just thought some might find that little story interesting :shrug:

  17. That is interesting...

    I realy kinda posted that about the o2's as I dont realy ever use the cal-con display other than while I am actually logn' and even then just glancing at it.

    I dump the log into EA almost as soon as I am done logn'. The "view log" or whatever in EA is like a spreadsheet format that can be filterd' to what your wanting to look at, and yes it is ALOT more detailed/precise(sp?) on what you can find that way. I realy found the playback on cal-con kinda is nice if you want to see a few things like what the sensors are doing as a whole when something is looking funky though I guess.

    I got EA like 3 days after buying the tweecer so I never realy got into the habbit of paying attention to cal-cons display. I remember something about the o2's "switching" going all over the range in a search when I first got the tweecer. I got the jist of the subject that it was normal and realy never paid attention to the display after that, just what was in EA's "view" spreadsheet deal.

    Very good all your tips are grady. Without looking at what are the actuall numbers in the MUCH slower "looking" format of the excell format you can realy miss ALOT of info.
  18. OK, my KAMS have changed for normal driving loads, but they're still at 1.00 at idle and my O2's are "dead" half the time at idle (values not changing). I'm still showing a lean condition so you'd think after over an hour idling lean (over 16:1 AF) the computer would start to adapt. Oh well.

    I found out that a few days ago my wife swept out the garage and swept the pile right by my MAF that was sitting on the ground:bang:

    She's a sweety, just trying to help me out, but thats what I get for leaving my toys on the floor.
  19. Chad

    I found that if you are pretty far off in your CL tune, the pcm won't allow CL to happen from time to time.

    At idle, you have no load on the car or anything like that soooooo

    Fatten that bad boy up a bit ... you can't hurt it

    At your idle maf curve point
    one above/below

    Add 5% and see what happens :D

    If no change add 5 more

    Don't forget to reset the pcm before the first and all other changes :nice:

    Let us know what happens :SNSign:


  20. Hey Adam

    You realize that you and I are doing the very same thing but kinda from a different direction :D

    We both will arrive at the same result because ......

    LAMBSE & KAMRF mirror each other.

    Where you look for what the LAMBSE is doing shortly after a pcm reset ....
    I do the same with the K's.

    IMHO, you have come a long way in a short time :nice: