Question-------------------------
Followup To
Question -
-------------------------
Followup To
Question -
I HAVE A 94 CARAVAN IT STARTS RIGHT AWAY WHEN ITS COLD AND RIGHT AWAY WHEN ITS HOT. BUT WHEN THE ENGINES WARM YOU HAVE TO CRANK IT FOR 15 SEC.BEFORE IT WILL START.I PUT NEW PLUGS,WIRES,AIR FILTER,FUEL FITER. I WAS THINKING FUEL PUMP GOING BAD? WHEN I TURN THE KEY TO ON I HEAR THE PUMP RUN FOR 5 SEC. SO I DID THIS 4 TO 5 TIMES TO BUILD PRESSURE THEN TRIED TO START START IT.BUT HAD THE SAME 15 SEC BEFORE IT STARTED.
Answer -
Hi Chuck,
I am less familiar with the caravans that with the cars (as my official 'disclaimer' says upfront) but tell me which engine you have if we don't get to the bottom of this the first time. I suspect that there is only one temperature sensor for the engine, the coolant temp sensor. If may be that this sensor is inaccurate and so is not telling the computer how warm the engine is so that an adjustment can be made for warm starts. If you know where that sensor is located on your engine you could measure the resistance across its two terminals; typically it should read 7,000 to 13,000 ohms when the engine is cold (70F) and 700 to 1,000 ohms when the engine is hot (200F) and somewhere in between when warm. So see if you can find it, verify that first. The sensor is unusally mounted near the thermostat housing or on it, and Chrysler tends to use wires that are colored black/light blue and tan/black for the plug into that sensor.
The other possibility involves the intake air handling at the air intake box (where the air filter is located): the heated air temp sensor or the air diversion door that it controls may be disconnected/binding up so that you are drawing in air off the intake manifold when you should be drawing in cool air or vice versa, or a different mixture of the two air sources. So examine the intake box to see if the vacuum line that comes to the underside of the box is attached to the little temp sensor mounted inside the box. That little vacuum hose often will be overlooked when reassembling the air box to the intake of the throttle body.
Those are my first thoughts, generically, without knowing which engine you have.
Hey Roland,
The engine in the caravan is the 3.3 V6. I located the temperature coolant sensor and checked the ohms. The engine was warm and I came up with the reading of 1700.I was guessing this was ok. Now today I tried starting it cold it had to crank for about 15 seconds as well. if I shut it off for about 20 minutes it won't start again for about 15 seconds. After the van has been running, and I turn it off and try starting the van right away it will start just fine.
Answer -
Hi Chuck,
That is an interesting modification of the problem (hard cold and hard warm start). I looked at the 3.3L design and it doesn't have the warmed air intake/vacuum hose control I mentioned in my first reply, so scratch that one. I'm thinking we may have to treat this as a "no start" and begin to look at the quality of the spark, but first why not check the computer to see if it has recorded any faults? You merely move the ignition key: from off to on to off to on to off to on, and leave it at on (do this switching within a time of 5 seconds, on means the 'run' position not the 'starter' position). Then watch the check engine light: it should come on for a couple of seconds, then begin to flash then pause the flash then pause, etc. Count the number of flashes before each pause. You will end with a string of single digit numbers. For example the final two sets of flashes will be 5 in each. Now take the numbers and group them in pairs to form two-digit numbers which are the codes for any faults in the memory. The last two digit number is always 55 because that is the code for "end of readout".
Let me know what the other code numbers are. You can also check out a code list at www.allpar.com but the interpretation of what to do is not always clear. So feel free to let me know. If we don't get a lead that way, we'll go at it in the classical way.
One other idea I had is that you may have a termperature variable vacuum leak that is diluting the mixture except when the engine is hot. For example if one of the bolts on the intake manifold were loose it might self-seal a leak at the gasket as the head heats up, but then open as it cools. So you might want to check the torque on those bolts. Also look for any unattached/cracked vacuum hoses.
But let us see what the fault coded readout shows. It might suggest a problem with automatic idle speed motor for example. There are several actuators and sensors involved in a successful start up: the autoshutdown relay has to close; injector pulse width is determined via coolant temp sensor, MAP sensor, and number of crank revolutions since cranking was initiated; and timing has to be adjusted based on crank/cam position sensor, detonation sensor, and throttle position sensor. Many of those are monitiored by the computer so a code for any one of them may be relevant to the problem.
Your measurement of the coolant sensor suggests that it is o.k. The fuel pump operating for about 5 seconds when you tun the key to run is normal.
Hi Roland
I rechecked the codes and I still have nothing 55.
I had a code about a month ago for a lean oxygen sensor. I replaced the oxygen sensor yesterday. I figured it couldn't hurt since it has 160,000 miles. Oxygen sensor didn't change anything. It seems like it is either getting too much fuel or not enough. Because after it starts it seems like it is sputtering. You smell gas at that time. Waiting for some ideas talk to you soon.
Chuck
AnswerHi Chuck,
So no codes helps to eliminate some of the electical sensors like the TPS as the cause. I would do two things:
Just to be sure, why not take a look at the spark quality when the engine is hot and when the engine is cold. Pull the center wire from the distributor cap and holds its tip 1/4" from the cylinder head and observe for the quality of the spark while an accomplice cranks the engine over for 5 seconds. If you want to also check the rotor, cap, and wires replace the central wire then pull one of the spark plug insulated connectors and using a phillips head screwdriver with an insulated handle insert the screwdriver into the connector so the tip of driver touches the metal connector inside. Then hold the screwdrive handle and position the assembly so the blade is 1/4" from the cylinder head, and again observe the spark for 5 seconds of cranking. That way we'll know the spark is not at issue.
Now that you describe some poor running and apparent odor of gas associated with the start up, I believe the issue may be the fuel/air mixture. Because you have multiple injectors it is less likely that one injector alone going bad would cause such a problem (though you could look at the plugs and see if any look wetter than the others after a failed start attempt). So I 'lean' (pun) toward the cause being the MAP sensor getting out of calibration, but not so far out as to set a code. This indeed happened to me, with our 2.5 tbi engine. About the only way to access that possibility is determine what baro reading it is showing at ambient conditions (what mm of Hg it shows when you turn the ignition switch to run, but before you start cranking). That reading, if it were off true, would set the mixture improperly during the cranking. The sophisticated way to find out is with a DRB (diagnostic readout box) but those are several hundred dollars (a Chrysler dealer tune-up mechanic can do at a price). If you don't have one, what I have done is to measure the voltage on the dark blue/red wire (the middle one on the plug at the MAP) with the igniton 'on' but not cranking. Here at sea level mine read 4.67 volts (=30.3 mm Hg). So depending on how high above sea level you live and to some degree the weather you will read a slightly different value. But I suspect if it is quite a few 0.1's of a volt different from that (unless you are in Denver), that the MAP is going bad. It is a $70 part from the dealer but it a chance you may have to take. To get at the reading, I took a very fine tipped pin and pushed the tip thru the insulation on the wire, so as to make electrical contact without leaving a significant hole in the insulation (which could ultimately lead to the failure of the wire some years hence). Then read it with a digital voltmeter compared to ground.
Let me know what you observe.
Roland
