I am chasing a cold start/run issue. On initial start the engine runs rough but not a dead miss. A CEL will eventually set with random misfire and lean codes. Also STFT are +25 and do not vary at all.
One would think that I am looking at an air leak and I have repaired a few of those (vacuum lines and intake retorque). The challenge is that once the engine is at operation temperature >190 F all is well.
STFT and LTFT numbers are great. The numbers go from the hard +25 to STFT numbers that hover around 0 and LTFT that is also reasonable. It is as the opening of the thermostat changes the nature of engine operations. Ideas? (new coils, injectors, O2 sensors, plugs are 30k old)

 

Ideas:

1. the Jag doc list all the possible causes for each code; as I expect you know, misfires are but one possible cause

2. when a code flags, freeze frame data get stored to provide extra clues - I suggest you see what has been stored (such as RPM, LOAD, ECT, ...)

Are there any drivability issues when the engine's hot? Does it accelerate as it should (say 30-70) and so on? I'm fishing for any signs the cats might be suspect or the coils (although you changed them).

 

Sounds like a vacuum leak to me. Do a smoke test in case there is another.

 

Thanks for the replies. Once the car is warming to operating temperature it runs pretty well. I went to Austin and back the other day and it ran nicely at freeway speeds. At the end of one leg of the trip there did seem to be a loss of response at high loads but not overly serious.
Why would cats cause a lean start condition? Also, the specs above do not show a lean running condition and these show up right as the operating temp reaches ~190 F. It as though as switch was set.
Anyway, I have thought of a cold air leak but cannot find one searching with carb cleaner or starter fluid (watching STFT for change). I did find some issues using this method previously. My re-torque of the intake solved a number of issues and was found using starting fluid and watching trim levels.
I will look at the freeze frame once it sets a code again. It does not set a code right away, it usually takes of couple of operation cycles. Recently it has started with the random miss and then will move to the lean code.
Again, thanks for the thoughts. I have searched the forum until incipient blindness and as you know this is a common issue. Will keep trying.

 

That's what the law requires for many codes. (Generally makes sense, too.)

 

Need to look into how the closed loop vs open loop is different. something in the closed loop system has a vac leak.

 

Does it stumble and run lean while in open loop? For the first while it does not pay attention to O2 sensors while in open loop. When the sensors are warm the system goes into closed loop. Does it run differently then? At what temperature does the roughness go away? Does the coolant temperature match the ambient temperature before startup? Possibly running lean because the thermostat is reading hotter at ambient than it should. Do you have a set of numbers, when it is running lean/cpld, like the one you posted when hot?

 

Interesting thoughts, it is fairly quick to recover from the cold stumble and it might be a correlation between the open loop running and the roughness. It does close the loop quickly (within 10-15 seconds). When it is below operating temperature (<190 F) the STFT are at +25. Then after the thermostat opens things calm back down to normal as you see in the initial data. I will take a closer look at some of your questions and respond back. Thanks for the insight.

 

Here are the cold numbers, this is right after the loop closed.

 

First - congratulations on doing over 40,000 miles in less than a month! The first set of numbers shows CEL Off. The second shows CEL On and that it was on for 40,721.56 miles. As you said she is running pretty well when she gets warm.
Second - The numbers show over 40% extra fuel being added however the O2 sensors both still show massive lean condition, Since it idles close to perfect numbers when hot a leak could be eliminated as a cause because the leak would have to be fairly large and probably would not diminish almost completely when warm.
Third - The sensors seem to be fully functional when the vehicle has run a while. However all four seem way off base when closed loop first comes in to play. I do not know if these cars supply +ve or ground to the O2 heaters but if ground then these vehicles are known for that problem. Of course it could be the O2's heaters are going bad but they would all have to be failing at the same time at the same rate! Same goes for connectors and wiring to controller.
Fourth - It could be none of the above and I am way off base!

 

Those trims looked high until I saw the RPM and I haven't any numbers for my car so now just don't know, but do all those trims head near zero and the RPM drop within (say) 5mins idling?

In effect it's a bit like it's still on choke (if you know older cars) with the numbers posted.

(In closed loop terms it's managing but the engine is cold so it could stall, and to avoid that the plan is to run a bit rich, and wait for the engine & cats to get hotter.)

We've not quite the same engine but mine idles nearer 600-650 when warm and I think yours also should.

 

40k per month LOL. I think that is total miles with CEL on. Which is kind of sad, 40k out of 120k is not a good percentage.
When it starts it stumbles and runs rough for maybe 10-30 seconds as it finds its way to a smooth, high idle. From there it will go closed loop, start to warm and the idle progressively reduces to about 650 rpm. From there it runs smooth enough but the STFT are pegged at 25. If I drive it the STFT will reduce to the 10's or teens but goes right back to 25 at idle.
The magic happens at about 190 F (probably about when the thermostat opens) and all the trims and O2 readings go to normal. The car then runs smoothly although I will say that when it is hot (having run for 15 minutes or more and especially after a longer drive) it does run poorly under load at low rpms (1500 - 2000).
Everything is newish, O2 sensors are not more than a few months old, coils and plugs are two to three years old but I have pulled to check and to clean connectors and the plugs look clean (but quite white for the most part). I used to reset the CEL every time I would mess with the car but this time I am letting it stay on. The emission control system is not yet being monitored as it has not run the proper number of cycles. I am waiting for that to see if there is an indication or fault code there.
Sure appreciate the help, I have been through the mind games with this and it is frustrating. Love the car, this, not so much.

 

Once you've stopped resetting things (for several drives) & once the engine & cats are properly hot if all your idle trims are within about 5 of zero all is pretty much OK, but you should not have stumbles/etc. They point to air leaks, usually. Check they occur even if a/c is fully off.

 

Your added info points to a relatively small leak in intake or exhaust system. However, and this is where I stumble, the initial hot numbers do not show any indication of a leak!
I always think of the exhaust as being under pressure constantly however the pulses actually go negative and can suck in air making the O2 sensors think there is a lean condition. Again, where I stumble with that is, both are reacting the same so it would have to be at a common junction before the cats. I don't think there is one, so we are down to leaks after MAP.
My OBD reader won't let me see the number one O2 sensors so can not run a test to verify any theory. The wide band O2 sensors give completely different output values and are actually interactive with the car's system. Reading voltage output does not work with them.
If you run a test, starting when completely warmed up, you should see the ma values of the O2 reading changing with revs - if there is a leak. Otherwise the ma reading should always be close to the value seen in your original post - around zero.
With a leak, as the revs are increased the values of ma and STFT will be pegged, if the leak is large enough. When the leak becomes less significant, in the increased air flow at higher revs, the added fuel will start to make a less lean condition. Both the ma value and STFT, and most likely LTFT, will decrease. As the rpm drops back the values should reverse their change and head to adding fuel again.
If this happens the search for a leak will have to continue.
As a matter of clarification, for anyone using this for educational purposes - The ma reading is actually what the car's system is providing to try and maintain the O2 voltage at the supplied 3 volts + the O2's output of +.3v. This is why one cannot tap into a 4 wire wide range O2 since the voltage will mostly be corrected to the voltage of 3.3v by the car. The ma reading compensates for the O2 current output to get the voltage back where it belongs as reference. Changes in readings are the opposite direction to a standard O2 output.
Anyone with a better understanding than I have please jump in and correct the above long winded comment.

 

You can use a cheap fog machine about $40 from ebay and tape a hose from it to your intake tube and run a smoke test this way. I would do it when your engine is cold since that's when your problem happens. It's possible that you have gasket somewhere that is worn enough to cause a leak but swell up enough at warm to plug itself.
Secondly, when was the lastime time your plugs and coils have been replaced? Plugs or coils that are on their way out could start the misfire when cold but get better when hot, happened to me.

 

I replaced the intake manifold and plenum gaskets. I also noticed a pinched o-ring on the vacuum connector at the rear of the intake going to the brake booster. I think it was the plenum gaskets but hard to say. The manifold came off either way.
Cold start stumble before closed loop - computer could not compensate for vacuum leak.
Closed loop - engine smooths but STFT at +25% until warm.
Once warm fuel trims settled down and sometimes right around 0%.

Still have a rough acceleration issue at part throttle but will deal with that in another thread.
Thank you all for the insight, just good to have the discussion.

 

2004 XK8, new plugs,COPs, O2s, MAP etc. RPM 1,300 on a cold start. After 30 seconds RPMs start down and in a minute or so 650 RPM. STFT is 25 both banks. Increasing RPM will bring down the RPM so vacuum leak. When car reaches temperature STFT and LTFT are fine. Car runs great. If I did have a scanner you would never know STFT are high on a cold start. It seems to me there is a vacuum leak when cold but expansion closes it up. What could it be? The intake manifold has new gaskets torqued in sequence. I think the plastic manifold is suspect. I just had a new transmission pan with an invisible hairline crack that leaked, quite common (JB Weld saved the day). Smoke, propane & Map cleaner reveal nothing. Opinions please. I don't know how to test for this. Is there anything other than a vacuum leak it could be? I could just smash my scanner and enjoy the car.




























stft
sf

 

I have never checked, but it is possible that is normal! When you stop your car LTFT is for the hot engine. STFT will have to get pegged at +25, for the cold engine, until LTFT catches up. Less additional fuel required for a higher RPM than at idle to keep the engine running. So STFT will drop some when rev'd.

 

Wait, do you mean "open loop". I runs fine when warm, closed loop. It's on a cold start that it has a problem, open loop? It's important to me since I have the same problem.

 

These cars have heated O2 sensors so Open Loop only lasts for about 30 seconds after start. After that time it starts to believe the O2's output and swings STFT to give proper combustion. It will start offsetting the remembered LTFT until it starts to correct that value. So STFT may stay high for a while during warm up until LTFT is corrected.
I have never watched it do this so have no real idea how long the process would take.