cold start, idles 5 seconds, sputters, dies

[Aurică]

Hi guys,
I have a problem with my Fiat Bravo ’08, 1.4 turbo T-jet 120hp( similar engine as in the Chevrolet Cruze, Dodge Dart, and Fiat 500).
The problem:
Every time i do a cold start ( the car has been sitting for more that 3-5 hours) the car starts up just fine, idles for 3-5 seconds and then it sputters and dies. It does this 2 or 3 times and then it won’t die anymore, but for the first 1-2 minutes the car is very jerky/ lacks power.

After it’s warmed up there’s no sign of any problems: idles ok, runs ok. The problem started about 1 year ago and it got increasingly worse, at the beginning it would just sputter for a bit and then it was fine.

I had a scanner on it:
-it has no stored codes.
-coolant and air intake temperature sensors report ok.
-did not find a vacuum leak (tested as Eric has shown)
-the MAP sensor reports reasonably close to the ECU expected amounts.(and since it’s a MAP i would not expect a vacuum leak to cause problems)
-tested the fuel pressure in the rail at cold start: 3.5 bar with engine off or without the vacuum hose connected, pretty stable, it drops to 2.5-2.8 at idle .. so that seems to be ok.
-fuel pressure bleed off seems a bit more than expected.. drops to around 2 bar in 10 minutes.. but it goes right back to 3.5 in 1-2 seconds with the ignition on… so not that bad.
-no blockages in the intake system.
-replaced the spark plugs.
-replaced the fuel filter.
-checked the EVAP solenoid- works fine.
-had the injectors out, tested the with an ohmmeter and they are all really close in readings( don’t remember the exact figures. but i remember looking up the official figures and they were close)
-tested the injectors (using the two-way obd2+laptop combo I’m using) spray pattern looked fine and uniform across all four.
-cleaned the throttle body that is ok.

-It has a coil on plug system. don’t know how i could test them.
-With starting fluid added to the intake it no longer sputters.
-LTFT was only about -7% at one time, i then reset the ECU and the LTFT to ZERO to see if that would make any difference- it did not.
– What makes a difference is the level of fuel in the tank, when full it only sputters; as it drops it sputters and dies, even tho the pump is still completely submerged in fuel( and altho fue pressure is still ok)

Any suggestions are welcomed!

Topic

[Tom]

Mate, I know you say pump pressure is fine, but if fuel level makes a difference, AND starting spray keeps it from sputtering, then I almost have to say it is a fuel related issue.

[Aurică]

Forgot to mention: at one time i replaced my pump with one from another bravo ( same engine, same pump) … issues did not go away.
( replaced it in the evening, drove it, and in the morning it sputtered again – no change)

I know it’s a strange one… that’s why i posted.
If i can get that guy with the Bravo to lend me his car for a day.. i would replace all the fuel delivery system ( minus the lines) ..his car is working fine.

[Tom]

Well, I am at a loss, but fear not, there are more experience people here than I, and I am sure one (ore more) of them will chime in with other ideas.

[EricTheCarGuy]

You’ve done a good job covering the basics so far. Well done. That said you can have issues with things like coolant temperature sensors or O2 sensors that can cause symptoms like that and not set codes. This is because they may be giving inaccurate readings, but according to the computer they’re OK. If you can hook up a scan tool and observe the coolant temp sensor readings you may be able to find the problem that way. Say, if the vehicle is cold and the temp reading says it’s at 100º already, you know something is up. Also, if you look at both the air temp sensor and coolant temp sensor with the engine off and cold, the readings should be about the same.

The O2 is a little tricker. What you’re looking for is a rapid movement between rich and lean once the engine warms up. If the O2 reading stays steady for a while before moving, then perhaps it’s not giving correct readings and causing the performance issues you’re experiencing.

More info on issues like this here.

http://www.ericthecarguy.com/faq/solving-automotive-performance-issues

Good luck and keep us posted on your progress.

[Aurică]

Thanks for the reply Eric!
I’ve already looked at the coolant and air temperature sensors.
They generally have similar values (maybe 2-3 degrees Celsius difference)
I would think they look alright.

The O2 sensor; I’ve checked them as per one of your videos ( this one: P0420 How To Diagnose A Bad Catalytic Converter -EricTheCarGuy ), looking for fast changing of the firs O2 sensor.(i don’t have a graph of that at the moment -but i will tomorrow morning) to me it seems to be ok. wen i press the accelerator pedal the sensor’s output increases quickly… and then when the ecu works out there is to much fuel it decreases as a response.

Anyhow I don’t think the O2 readings would influence the starting, because the ECU does not take it into account until the engine is up to temp.

Attachments:

• 1temps.PNG

[KZ 259]

This problem kind of reminds me of a faulty camshaft sensor. From my experience, a faulty camshaft senor will act up in the way that is described, and sometimes, it will throw a code or not. But if this same issue has been happening for over a year, and with no codes, then my suggestion is very unlikely. Have you tried resetting the ECU? Is there any subtle fuel-like smell around the vehicle?

[Aurică]

@ KZ 259
The problem is ongoing for over 1 year, slowly getting worse.
I was thinking of a possible crankshaft/camshaft sensor issue… but that would mean that the fuel level would be a genuine coincidence… still it’s possible.
There is no gas/fuel smell around the vehicle.

[Bryan Umberger]

I had a similar issue with a truck I used to have. It would act strangely after a fill up or on cold starts. It ended up being a bad fuel pump module. Nothing wrong with the pressure it had to do with the evap system. Somehow raw fuel was getting drawn up into the system and when the solenoid would open to dump what was supposed to be fumes it was dumping fuel into the intake system causing it to sputter and die. Once the engine got hot it would vaporize the fuel prior to it entering the cylinders.

Check your lines coming from the evap solenoid to see if it is wet. If they are you may have the same issue as I did.

[Aurică]

[quote=”umbergbr” post=104819]I had a similar issue with a truck I used to have. It would act strangely after a fill up or on cold starts. It ended up being a bad fuel pump module. Nothing wrong with the pressure it had to do with the evap system. Somehow raw fuel was getting drawn up into the system and when the solenoid would open to dump what was supposed to be fumes it was dumping fuel into the intake system causing it to sputter and die. Once the engine got hot it would vaporize the fuel prior to it entering the cylinders.
[/quote]

I already thought about that.
The evaporative system is separate with a charcoal canister next to the tank and line to the valve, it’s not connected to the fuel pump at least in my car.
So one evening I unplugged and caped the pipe coming from the canister to the EVAP valve (it’s situated on the engine) an also blocked the pipe to the valve (so there would be no vacuum leak).
Next morning when i started the car I also logged the rpm vs canister purge valve opening. It did not open at least for as long as i logged it (until the car died)

I’ve also tested the valve (bidirectional testing with the laptop), blowing air in through one side, and it would only pass through when the valve would be opened.

What i was thinking now:
Knock sensor… maybe (but still would not explain the difference in running with more/less fuel in tank)
Reasoning: when the engine is cold it’s dry (until the oil comes up) ..so it’s much noisier, so maybe the ecu is retarding the engine to much.. so it’s not able to keep running.
With starting fluid it’s running richer .. so less chance of pinging… so it’s not going to die.

Now.. i should unplug the sensor and see how it goes.. just that the sensor is only accessible from underneath – it’s somewhere above the starter motor.

[EricTheCarGuy]

I think the knock sensor is a stretch. Without any codes stick to the basics like fuel delivery, vacuum leaks, and basic tune up stuff.

Looking at your chart though, was the engine cold when you took those initial readings? If it was those sensors are reading very different things. That might be a place you might want to start.

[Aurică]

I know, you’re probably right.

About those reading there is a difference of 8 degrees Celsius, to me that seemed ok because the car was sitting in the sun when that was taken. Made sense that the air would be hotter.

The basics, I’ve checked them 3-4 times each.
I don’t think it’s the fuel delivery because:
-pressure is good;
-LTFT is ok;

I’ll post back as soon as i have some more info, maybe a video so you guys can see exactly what is going on (suggestions on video content would be welcomed)

[Aurică]

Hi again,
I’ve managed to upload a video of the cold start Bravo cold start, I also have one of a regular start about 1-2 hours after I’ve been driving the car Bravo warm start

[James]

Do you have access to an oscilloscope or even a fast graphing multimeter? I would look at the cam and crank sensor and see if you’re inputs are good. I would also look at the secondary ignition waveform and see how it looks. It’s either an issue with fuel or spark, so it could be an input problem from the crank/cam, or an issue with ignition or fuel. If you don’t have access to a 5 gas analyzer, there is a little trick you can do to see if it might be a fuel issue. Start it in your garage and see how much it stinks. If it’s an ignition issue, it will probably stink of gas real quick because fuel will be spraying and not igniting. If it’s an input issue, eg. the injectors and spark is not firing, it should smell normal, or if it’s and injector issue, which I don’t think it is, your spark will be firing with no fuel. A secondary ignition waveform will be able to tell us a lot about this.

If you don’t have access to these tools, it might be a worth while investment to get an hour or two of diagnostic time at a shop that has the tools, and also has the knowledge of how to interpret these waveforms.

[Lorrin Barth]

After all you have gone through with this vehicle I feel I should suggest something complicated. However, to me it says vacuum leak.

[EricTheCarGuy]

The coolant temp and air temp should have been closer than that in my opinion. If it were me, I’d be looking closely at the coolant temp sensor operation.

[Author]

Replies