Rob W.

iTrader: (8)

In my quest to figure out why my car idles so lean (17.5-18) after transition to closed loop operation, I thought I'd look at how the fuel trim works. Can someone give me a basic summary? Searching turned up nothing.


On a separate note, I did a boost leak test this weekend and it was tight as a drum up to ~30psi. I now plan to read the service manual to see how to diagnose a bad upper O2 sensor.

MalibuJack

iTrader: (5)

Fuel trims are a % of adjustment.. In a standard OBD-II logger, the short term trim is equivalent to the O2 feedback in Evoscan..

Long term trims are an average of adjustment.. There are actually 3 long term trims, only one seems significant and logged in a OBD-II scanner.. I have never seen one of the three adjust, only two of them.. They take some time to update though.


Trim percentages are an adjustment from the base map.. Thats why getting your base map hidden under closed loop fueling, and your injector scalings and latency in check..

The closer you get your injector scaling and latency the less adjustment you need to make to the base map to keep things in check.. However adjustment is almost always necessary..

If your base map (low load columns) is 14.7 according to the map, and the car idles at 17.1, the fuel trims are going to be outside the region of adjustment, so it will idle lean..

There are several ways to make the adjustment, and all of them have different purposes..

But I'll keep this focused on adjustment for open loop (cold start)

if your trims are over 25% off (in OBD-II) or 50% off in EvoScan, you need to add or remove fuel accordingly in the base map.. The closer you get to getting the adjustment in the correct range, the less the fuel trims will adjust, and the less likely you'll get issues..

This needs to be done at low load idle, low load low speed driving, and low load highway cruising.. Since long term trims reflect an average of all types of driving conditions and idle.. You need to spend time to get it right..

The closer your base map is to idling and cruising at 14.7 AFR's, the less adjustment it makes, the more likely the car will run better.

There are some things that will affect your ability to do this well however..

Your ISCV adjustments can affect your idle and cruise adjustments, since the ICSV will open varying amounts under different loads, if its open too much at idle, and it takes some time to return to a setting that idles well, it will throw off your trims.. Hence why you spend time adjusting this.. even your BISS (base idle set screw, or air bypass) settings can affect it..

If you find your car idles fine for some time, and your trims are okay, then every so often the idle goes real lean, or real rich, its due to increased load (Alternator load due to fans kicking in, etc) This would be one of the places you look, since there is a target idle, and the ISCV value makes assumptions of a good starting point for that value.. If its wacked, it will occasionally reset to that value...

Hope this explains a few things..

The O2 sensor will either read or it wont.. If the Voltage values are in range, then its okay.. if they read a low voltage, or high voltage ALL the time, even when decelerating, then you have a bad O2 sensor..

Rob W.

iTrader: (8)

Thanks for the advice, Jack. You've given me a lot to look into.

What is the range (I assume it's voltage) for the NBO2 in EvoScan? I see on decel, idle, and WOT, the range is always between 0 and 1. Sound fine?

I noticed that on my HO Fuel map, the upper left cell (500rpm, 10% load), my tuner set it at 15.7. My buddy's stock EVO 8 is set at 13.3. I'm not sure why the tuner would ever set it there, but that's the first thing I plan to fix! I didn't pay attention to this before because I thought these open loop maps didn't affect the idle AFR target.

In a long idle, my all my Fuel trims are reading right around 100 ("Long" never changes, as you indicated). What is the scale here? I mean, at 100% is the ECU maxed out trying to add fuel? Or is 100% ideal (so 150% would be '50% over', etc.)?

Thanks again,

Rob

MalibuJack

iTrader: (5)

Not 100% sure.. The trims in Evoscan have always been something I was confused by, so I typically use a different tool to log that data.

MalibuJack

iTrader: (5)

Its a range from 0.0v to 1.0v Unfortunately if its rounding the data you won't see anything but 0 or 1..

On my logs it shows a voltage of .8 to .95 and sometimes a value of 0 on decel.. so I think you found your problem..

chmodlf

iTrader: (19)

If it is worth anything...Big Al says that the Fuel Trims in EvoScan are off.

Rob W.

iTrader: (8)

Actually, I think the O2 is working. Maybe I described it wrong. I get 0 on decel, and various fractions between 0 and 1 the rest of the time. (0.95 on WOT)

I just reflashed that low load AFR cell back to 13.3, and it did NOT change anything, unfortunately.

MalibuJack

iTrader: (5)

Yeah, in .8 they do seem to be wrong.. but for those of us familiar with a regular scan tool, their definitely wrong..

I think that was fixed in 0.9's data.xml file..

Mad_SB

iTrader: (8)

So, is the problem only at idle? Is the parththrottle cruize afr around 14.7?? I ask because when I put ebay 02 on, i got a leak at the down pipe where it bolts to the o2 housing. The leak is after the front 02 sensor and only seems to affect the wideband at idle. Front 02 seems happy, part throttle cruize the wideband reads 14.6-14.8 so I'm fairly sure the problem is the leak in my case. Waiting on a couple gaskets to show up so I can seal it up.

Rob W.

iTrader: (8)

So, nobody has used fuel trim in EvoScan yet? If so, can you help me with the correlation?

mad: Yep, exactly. Part throttle cruising is a perfect 14.7. I know all the bolts are tight, but I'm wondering if it could be the flex joint in the DP. The DP is new, but I guess this might be considered normal for a flex joint??

This weekend, I'm going to try a test where I wrap the flex joint in something (tin foil, probably) to try to seal it, then see what the idle AFR is. If it's the joint, maybe I'll just leave it alone. I mean, since I get good WBO2 readings at WOT, is it really going to hurt anything?

l2r99gst

iTrader: (2)

I used fuel trims extensively in DSMLink, but I have not yet in EvoScan. Fuel trims really have an adjustment percentage positive and negative. In the DSM ECU, there was a range from +12.5% to -12.5% for the LTFT.

Looking at the data.xml in EvoScan, it looks like he has is set to a scale from 0 - 200, so I would venture to guess that 100 means you are fine, needing 0 adjustment. But, I don't think anyone has verified whether the trims are reading correctly or not yet.

If this is how the trims are setup in EvoScan and you are reading right around 100 during a long idle and steady cruise, then you should be fine. If this is the case and your wideband is reading lean at only idle, it is most likely a leak at the DP.

For example, I think I have a small leak at the downpipe myself after installing my exhaust. At cruise my wideband shows stoich, right around 14.7. At idle, it shows about 15.2. I was thinking about getting studs for the O2 housing/DP connection to create a better seal, possibly, but I never got around to doing that yet.

Jack pretty much answered your questions about fuel trims. The only thing that I have to add is that there is a separate LTFT for idle and for cruise. They are referred to as LTFT low and LTFT mid, usually. There is also a value for just plain old LTFT in the ECU....this is the particular LTFT that the ECU is using at the moment. For instance, I think most OBD-II loggers will just show you the plain old LTFT, so that you don't have to log both LTFT low and LTFT mid. Plain old LTFT will show you LTFT low at idle and LTFT mid at cruise. DSMLink allowed us to see all of them....I'm not sure how EvoScan is setup at the moment. But, it is good to know the LTFTs for idle and cruise, as they are different and they can be used separately to pinpoint issues at idle, etc. These were used a lot with DSMLink to dial in the deadtimes of injectors, since the IPW at idle is about half that of cruise and would affect each LTFT (low and mid) differently.

Anyway, I think I'm just rambling now.


Eric

MalibuJack

iTrader: (5)

There is fuel trim low and fuel trim mid.. They do read 100 when things are perfect.. In my case its always leaning things out slightly so its 83 or so..

You can cruise at 14.7, when you come to a stop, if it goes rich, then the cruise trim may be a bit leaner than at idle..

Thats the one thing that is really missing at the moment, and that is a true injector latency and injector dead time value.. I think I have found areas in the rom where that might be them, but I'm not sure..

When Al says the trims are off, basically its because their different than how a regular scan tool reads them.. They are actually accurate on a scale of 0-200 if 100 is the 0% trim.. I am planning on using a scale of -50 to +50 for my trim calculations... That should make it more similar to the OBD-II trim values.. (I do think its actually -25 to +25 since I've never seen a bad trim number go over that..)

l2r99gst

iTrader: (2)

MalibuJack,

When you are cruising the ECU is using the LTFT mid. There is a switchover point at idle so that the ECU uses the LTFT Lo. They are related, but separate trims. There is that small moment in time where they switch, which may cause a momentary rich or lean condition, which is what, I believe, you are referring to above.

I can't tell if you are agreeing with me or disagreeing with me based on your wording and/or my wording.

Usually, though, it is the fueling at the switchover from closed loop to open loop that is dramatically affected when fuel trims are off, causing either a rough rich transition or a rough lean transition to open loop.


Eric

MalibuJack

iTrader: (5)

If assuming the zero fuel trim in a regular scantool, equals 128
and the fuel trims don't go over +/- 25% I'm assuming it has to be 128

I am going to use this..

int(.1961*x)-25

basically this should put the range of values to be 50, subtracting 25 should put the midpoint at 0 where a regular scantool would read..

You can't really do that in evoscan, so you may have a slight rounding issue.. but it works without the int()

in Evoscan it would probably be .1961x-25

I'm going to have to log the raw value for the trims.. but if the midpoint is 127 or 128, then .1961 should scale it as correctly as possible to be useful..

FWIW I derived it by dividing 50/255 and it gave me .1961 which works if the midpoint is 128

If someone gets to do this before I do.. Remove the trim formulas from the O2feedback trim, and the Trim_Low Trim_Mid and Trim_high.. For O2Feedback it should go to 127 or 128 at WOT (or open loop fueling)

Trim_high always reads 100 and unmodified it should read whatever the 0 point value is.. it should be 127 or 128... If it is, then you can use the formula I posted and it should get you the trim you'd expect to see.

Why didn't I give you a range of 50%? Well, I have never seen it go more than -25 or +25 on a regular scan tool before it reset the trims to zero and threw a CEL code..

MalibuJack

iTrader: (5)

I was actually agreeing with you.. Typically I try to give as much data as possible behind my comments that it isn't always clear..

And yes, that transition point between the two trims is where I see it..

of course the jump from closed loop to open loop where it can transition lean or rich, is what typically causes people P0300 code.. so getting your trims in check will reduce or eliminate that..