Need some tips on a fluctuating a/f ratio

[CyFi6]

OK well some info on my car
7MGTE
MAFTPRO tuning - speed density conversion
550cc injectors
Walbro 255lph
2.5" intercooler piping/3" intercooler
full 3" exhaust

The problem I am having is that sometimes at idle, the a/f ratio will fluctuate from normal to lean and back (will go lean for maybe 5 seconds, then normal for 5 seconds, then back again). While doing this the RPMs drop and sometimes surge around 550-600 RPM, then as a/f ratio comes back to normal it will even out around 650-700RPMs, then repeat itself over and over. This usually happens when the car has just barely warmed up, but has done it to me fully warmed up as well.

I am monitoring A/F ratio with my NGK AFX wideband and when it goes lean it usually reaches around 15.7 or sometimes maxes the gauge out at 16.0. Idling well it idles right around 14.7.

At idle my vf reads about 2.5v steady. Even when the idle drops and it runs lean/rough, my vf does not change by much if any at all. This confuses me, though i don't know how fast vf normally responds or should respond. Checking o2 sensor voltage with my DMM (no oscope) it will read normal (.5v -.8v) and then the signal drops to mv when it runs lean and rough. It almost seems to go to low voltage before the engine runs rough though, so I am not 100% certain that the sensor is reacting to the lean condition rather than causing it.

Here is the odd thing: if I unplug my o2 sensor, my a/f ratio goes rich, down to about 12.0 and no more fluctuation or surging happens, just runs very rich. Now if I am not mistaken, with vf happy with the o2 plugged in, shouldn't the engine run close to 14.7 with the o2 disconnected?

The o2 is less than a year old and is a Denso sensor. I am not certain if it a MAFT pro tuning issue or something going on with the stock system, but I am inclined to believe it is not MAFTPRO related because there is no fluctuation with the stock o2 disconnected, and the MAFTPRO does not even monitor the stock o2. Also, the airflow frequency out of the MAFTPRO does not change much at all when the engine goes lean, and the airflow signal is the only thing the MAFTPRO can actually alter.

Just wondering if anyone can clear some things up that I am confused about, such as why my car reacts the way it does when unplugging the o2, why my vf doesnt respond to the change in o2 voltage etc and if anyone has any idea where I should begin my diagnosis. Thanks.

BTW only codes i have are
24 - Intake air temp (no more AFM so i believe this is normal)
41 - TPS (because of a feature I am using in the MAFT PRO that mimicks the TPS signal, this code is "normal". Because the stock ecu fires up just a little bit faster than the maftpro, the TCCS sees an open in the TPS circuit at key on)

 

[roadboy]

actually to be honest, my car does this as well, not the surging though... at idle i could see it go lean for a sec and go back to normal, wouldent know what the problem would be but i have a code 51 which was a real-time code

 

[jetjock]

From reading other threads I thought code 24 is normal with a Maft Pro but 41 is not. I may be wrong though.

 

[CyFi6]

I have a stored code 41 (tps code) and i suspect its due to the ecu powering up with the pro and beating the pro to the punch. Didn't even realize i had the code until i checked it before pulling the bottom end lol! never noticed a difference when driving and never had the check engine light on at all!

Working on a way to power the pro before the ecu to avoid this. To much other stuff do to though so it will be a while before i fix it..

Looks like Grim had the same thing. It is an interesting function that allows the ECU to see a WOT signal when certain parameters are met to avoid a lean part throttle boost...unfortunately it isnt perfected and it throws a code. I am assuming that niether of these codes when stored will actually affect the operation of the engine?

Also, I was under the suspicion that because of my 550's I was scaling back the airflow enough that the computer was seeing a much lower than normal airflow value at idle, but even after adding a ton of idle enrichment (getting the airflow signal back into its normal range) it would still fluctuate lean.

 

[jetjock]

Got it. I thought the power up thing caused a 24 based on a past discussion with another member who was trying to resolve it. My bad.

Not sure what the problem could be but fwiw the mixture on a stock system with the sensor unplugged should go to about .9 lambda. Slightly rich. Every Supra I've ever messed with did. Should not go as rich as your seeing because that's in violation of EPA criteria. They don't want catalyst efficiency reduced to such a level. Remember, it's all about emissions, why OBD was developed in the first place.

 

[kneedragger85]

A code 24 is no bueno. The ECU will set a default IAT value of 130*F and give you some fun tuning. I ended up using diodes to power on the Maft-Pro before the ECU via the accessories, then continued through another diode off the ECU power wire to keep power while on/cranking to avoid this. You need the Maft-Pro to generate the "dummy" signal before the ECU turns on.

 

[CyFi6]

Thanks for the info. So I need to do the resistor mod to get my car to run right without storing the IAT code? What resistor ohm value do you recommend/what temperature should I have it default to? Besides storing a code, are there any negative aspects to leaving the TPS and MAFTPRO power circuit the way they are? Will the engine function normally with the TPS code stored?
Thanks.

EDIT: If i connected an actual thermistor located pre-turbo for the IAT, would that be an ideal solution, or would that make tuning more difficult?

 

[jetjock]

...A code 24 is no bueno. The ECU will set a default IAT value of 130*F ...

Not only is this incorrect but it makes no sense for the reason already mentioned.

 

[kneedragger85]

Not only is this incorrect but it makes no sense for the reason already mentioned.

I was way off from memory but was going off this for just having a code 24 as a result,

1) Fault in the Coolant Sensor or Air Intake Sensor.

Necessity of function: If an open or short occurs in either of these sensors the ECU will judge the temperature to be either below -50 C or above 139 C. As a result the air fuel ratio will become too rich or too lean which could lead to engine roughness or stalling. Ignition timing will also be retarded if the coolant sensor input is deemed too hot.

Action taken: Standard values are substituted. They are 80 C for coolant and 20 C for air. These values are typical for normal engine operation on an average day. This is code 22 and code 24.

 

[cyberdx16]

Thanks for the info. So I need to do the resistor mod to get my car to run right without storing the IAT code? What resistor ohm value do you recommend/what temperature should I have it default to? Besides storing a code, are there any negative aspects to leaving the TPS and MAFTPRO power circuit the way they are? Will the engine function normally with the TPS code stored?
Thanks.

EDIT: If i connected an actual thermistor located pre-turbo for the IAT, would that be an ideal solution, or would that make tuning more difficult?

With Map ecu 2 we us a 2.2k ohm resistor, I believe that is equilvant to 25 degrees C

 

[3p141592654]

Nitpicking, but the default IAT setting is 20C when there is a code 24 fault, and the minimum value the ECU can read is -40C. Since the GTE ECU takes the AFM pulse width and corrects it by both a HAC factor and a IAT factor, I would guess the MAFTPRO would want to disable those two inputs anyway, so a code 24 might be a good thing assuming the MAFTPRO adjusts for those itself.

The problem I see is that the air temp signal does get used for the basic correction to the "load" variable, cold start asynchronous injector duration, and more importantly it affects the ignition timing map lookup and the enrichment map lookups. So, running at 20C on a hot day you will be more susceptible to detonation because the IAT timing retard and extra fuel corrections won't be active, and that is on top of the problem that you have from scaling the air flow to match the 550s which is already resulting in too much advance. Anyway, such is the life of a piggyback kludge solution.

I checked the ECU code and I don't see anywhere that the code 24 flag is used to affect the way closed loop mixture is calculated. So getting back to the original problem, is it possible you have a misfire, because that will drop the idle and cause an artificial lean reading on the O2 sensor. If you have a 4 gas analyzer or can take it to a smog check place with one, then you could narrow down if it reall y is going lean, or is a misfire.

 

[CyFi6]

Well it does missfire when it goes lean, but I believe it misses because it goes lean, and not the other way around. When I disconnect the o2 sensor the idle stabilizes, but runs rich. The car often does have a low load missfire (that common idle bump), but I have had the car idling and missing and not reading lean on my wideband.

 

[3p141592654]

The idle fuel trim does have a little gizmo that causes a "timeout" for 5 seconds in short term fuel trim adjustment after a rich->lean transition with the car in idle mode (speed <2km/h, throttle closed, temp>67C, and FPU off).

So typically when rich at idle, fuel is pulled gradually (0.06% every 16ms) until the OX sensor trips over lean, then trim is held for 5 seconds, and then a big 4% jump in fuel is added, followed by adding more fuel at a very slow rate (0.03% every 16ms) until it trips rich. This repeats for 3 minutes, and then the 5 second timeout is disabled until the car goes out of idle mode and back.

So, to see if this timeout is related to your problem, warm the car up by going for a drive, and then idle it and see if the problem goes away after 3 minutes.

On my car, at warm idle the rich mode is short (1.2s) and the lean mode is long (7.6s), mainly because of the 5 second timeout, and also the 50% slower trim correction.

 

[Blackdawg]

yeesh crazy stuff..i am honestly scared to go through this with my car and the maft..thanks for the info tho!

 

[CyFi6]

Bringing this back from the dead. My car still does this and it is so irritating. The biggest problem is it is not completely consistent. The problem worsens with cold weather. All through summer where air temps are over 95 deg f even at night, I never had a fluctuation of idle rpms. As soon as the weather started to cool off, the problem reappeared. My situation is basically the same as before.

Idling with 02 plugged in, car not completely warmed up (driving for about 5 to 10 mins), engine idles at ~700rpm A/F ratio right around 14.7, VF right around 3-3.5v, does this for a few seconds, then A/F gauge pegs lean (16.0:1) and idle drops down to ~500 and surges up and down between about 400 and 600 back and fourth, does this for a few seconds, then repeat the process (idles good ~700 14.7:1 etc for a few seconds, then surges a few seconds and repeat).

Idling with O2 unplugged, car idles solid and stead at ~650 to 700rpm, A/F ratio very rich (12.0-13.0:1) but steady, no major fluctuation and idle stays rock solid. Plug the O2 back in and the previous scenario begins again.

Once fully warmed up the problem goes away mostly to completely. Here is an old video I took of the problem so you can see what I am talking about. This was back when I had a shitty prosport A/F gauge (has since been upgraded) so ignore the "stepping" action of the gauge, that is not the actual representation of the A/F ratio.

http://smg.photobucket.com/albums/v313/cyfi/?action=view&current=IdleFluctuationVideo.mp4

This video is when it is very cold outside. Engine started from about 40deg F and driven for maybe 7 minutes, then let idle. I keep my car in a garage now so it doesn't get the same kind of cold start, so when it happens now it is less severe but very noticeable. So when this happens, I need to let the car sit for 3 mins and see if it continues?

 

[3p141592654]

Missed this. Yes, let it idle for 3 min (time it) and see if the problem is reduced or eliminated after that.

In the attached figure, you can see the fuel trim at idle with the long 5 second hold after switching lean, then a quick ramp up rich and then back to a long lean cycle. At 11:47 the engine goes off idle and the rich/lean fuel trim frequency becomes much quicker without the 5 second hold.

 

[CyFi6]

It seems that the problem lessens slowly over time, and there is no distinct change at the 3 minute point. After driving/throttling the problem returns then slowly diminishes over a few minutes but never completely goes away. Here are some more videos to show the problem. You can see the fluctuation reduce over time, but after I throttle it, it returns. Since every time I come to a stop I have just let go of the throttle, I see the problem whenever the engine is fairly cool and I have come to a stop.

This problem doesn't happen when the engine is stone cold. When the engine begins to idle down to about 700 from the cold start high idle, then it starts to happen. And like I was saying before, once I unplug the O2 sensor, the a/f ratio at idle goes rich, but stays rock solid and does not fluctuate.
http://smg.beta.photobucket.com/user/cyfi/media/3_zps752bb78f.mp4.html
http://smg.beta.photobucket.com/user/cyfi/media/2_zps4f6240fa.mp4.html
http://smg.beta.photobucket.com/user/cyfi/media/1_zpsb0d8bfb5.mp4.html

 

[CyFi6]

Bump?

 

[3p141592654]

I would suggest a secondary ignition analysis, to check for ignition anomalies and make sure all the injectors are behaving.

 

[CyFi6]

Went ahead and got a pattern for the injectors, here they are 1-6 Everything looks normal to me. These were at idle, nothing stands out at 2k rpm. Unfortunately I have no means to get a secondary pattern with the tools I have now, so I will have to find someone to do that for me.

I do have a question regarding the injector patters though, maybe you could help me out. I understand that these injectors are fired in pairs, so each pair of injectors has a common switched ground. Does this mean that when I scope for instance the number 1 cylinder injector, I am looking at the injector pattern for both #1 cylinder and #6 cylinder? How could I only focus on one injector at a time? Would I have to disable the other?