Mechanics

ENGINE IDLES ROUGH, POOR GAS MILEAGE. ADVICE?

2002 Jeep Grand Cherokee • 143,000 miles

2002 Jeep Grand Cherokee Laredo 4.0. Brand new reman engine. Gas mileage went from 17.8 to 7.4mpg. I am assuming this has to do with the reason the engine idles rough and worse when there is a load put on it. (D or R and/or when the AC is turned on max). Tested and checked IAC, TPS, and the coil rail. Has brand new plugs (Champion RC12ECC). I also cleaned the throttle body and IAC real good. Also tested all the injectors today.
AD
Avatar
Jfazio5188
January 12, 2012.




You're really barking up the wrong tree. First you have to determine whether there's a misfire and fix the cause or there's way too much fuel going in causing the engine to run rough in a flooded condition. That can be done by viewing fuel trim numbers.

The Engine Computer uses the automatic idle speed (AIS) motor, also know as the idle air control, (IAC) to control idle speed when your foot is off the gas pedal, that's all. It has nothing to do with other running problems. The throttle position sensor just tells the computer which way and how fast the gas pedal is moving. They cause very little trouble and definitely not what you're describing. Injectors can be tested electrically on the vehicle but that won't show if one is sticking open and spraying fuel constantly instead of in short pulses. Chrysler has extremely little trouble with injectors so they should be the last thing to look at. You can check for that by monitoring fuel pressure when the engine is off. It should hold steady for weeks. If an injector is leaking badly, that pressure will drop off within a couple of seconds.

One misfiring spark plug can cause a lot of fuel consumption but not nearly enough to cut the fuel mileage in half. With a misfire, unburned fuel and air go into the exhaust where the oxygen is detected by the oxygen sensors. They do not detect unburned fuel. All the computer knows is that side of the engine is too lean. It doesn't know why or if it's being caused by just one cylinder. It responds by requesting more fuel on that side from all of the injectors. No matter how much extra fuel goes in, there will still always be that unburned oxygen from the misfiring cylinder. It's a never-ending cycle, but the computer can only add or subtract fuel by about ten percent.

That's where long and short-term fuel trim numbers come in. You need a scanner that can display live data to see what the computer is doing. High positive numbers means the computer is requesting more fuel in response to an incorrect sensor value, typically the MAP sensor. He has the biggest say in how much fuel is needed. If the numbers are high negative, the computer is trying to cut back on fuel metering, obviously without success. That's where you have to look at things beyond its control. That includes an injector sticking open, (very rare), or high fuel pressure. Look for a leaking or disconnected vacuum hose going to the fuel pressure regulator, or a crushed fuel return line to the tank. Also look in that hose for signs of wetness. Your problem is real common on GM trucks due to a leaking regulator. I've only heard of that once on a Chrysler product.

A lesser-known problem that causes poor fuel mileage is an exhaust leak ahead of the first oxygen sensor(s). Between the pulses of exhaust flow, the momentum creates little pulses of vacuum that can draw in outside air. The oxygen will be detected as a lean condition that the computer will try to correct by adding more fuel. You will often smell the unburned fuel at the tail pipe but the computer will be showing a lean condition.

Cleaning the throttle body can solve some hesitation or stumbling complaints, again, more common on GM vehicles, but that will slightly restrict air flow if severe enough. You have way too much fuel going in, and that has nothing to do with the throttle body.

If none of these things seems to point to the cause, you're going to need a scanner to see what the Engine Computer sees. If you unplug the MAP sensor's electrical connector and the engine runs better, (not well), that's a clue that it's providing incorrect information. The main fuel delivery calculation starts with load on the engine which is evidenced by a drop in intake manifold vacuum and is measured by the MAP sensor. All the computer cares about when self-testing that sensor is its signal voltage must be between 0.5 and 4.5 volts. The readings can be way wrong, but as long as they remain within that range, no fault code will be set in memory.

Have you even checked for diagnostic fault codes? If the Check Engine light is on, and it should be, there will definitely be at least one code. The light is only required to be turned on when the problem detected could adversely affect emissions. There can be other codes that don't turn the light on.

Caradiodoc
Jan 13, 2012.
Thank you so much for your information. I have done some test and I will list the results. There is still no change in my problem. Cleared the pcm and rechecked for codes a couple days later. I had no codes stored. I replaced the fuel filter just to rule it out because it was cheap enough. I re-torqued the exhaust and intake manifolds to ensure that there is no leak. I also replaced both upstream O2 sensors because they were the originals. I clean and cleared the vacuum on the MAP sensor. I also tested the map sensor and the connector. At the connector with the key in the "run" position it measured 5.0v which checks out. I then probed the MAP sensor, with the key in the "run" position it was reading 4.7V. With the engine on it was bouncing back and forth between 2.3V and 2.0V.
AD

Tiny
Jfazio5188
Jan 18, 2012.
Now we have some more variables to confuse the computer. All sensors are different. The computer learns their characteristics as you drive by comparing their signals to other sensors and known conditions. I doubt replacing the O2 sensors is going cause problems, but randomly replacing parts is the least effective and most costly way to diagnose this type of problem. You'll be money ahead to just have it professionally diagnosed.

I would have told you also to not waste your money on a new fuel filter. Except for diesel trucks, they last the life of the vehicle on Chrysler products unless they start to leak like mine did at 180,000 miles. The pickup screen in the tank will collapse or become plugged before the filter will, and even that is not very uncommon. Replacing the filter will not have any effect on the computer. He won't notice any change from that. Also remember we're dealing with way too much fuel. A plugged filter or screen will reduce volume and pressure. The engine would run poorly but the fuel mileage would not get worse.

Two volts on the MAP sensor at idle seems a little high from the experimenting I did when building "bugged" cars for my students to troubleshoot, but it is within the acceptable range to not set a code. The higher the voltage, the higher the load and the more fuel is needed. Since that voltage could be incorrect, the fastest and easiest way to know if it is causing the problem is to pop a new one or known-good used one on and try it. If I had to guess, I would have suggested around one volt at idle would be normal but if I found two volts on a good-running engine, I'd accept it.

How did you "reset" the computer? If you disconnected the battery, that erased all of the stored fuel trim data it had learned during previous drive cycles. Knowing that is what helps it figure out when there is a problem, and why. Once the battery is reconnected, it has to relearn all of the sensors while there's still a problem present. Any fault codes that were in memory are lost too. That can provide real valuable clues. Part of the strategy the computer uses is to constantly compare sensor readings with each other. As an example, it knows that when the engine has been off for more than six hours, the coolant temperature sensor and ambient air temperature sensor, (or battery temperature sensor), had better be reading the same temperature. If there is a fault code stored for one of them, the computer knows it can't rely on its reading to compare to the other one, so it will not set a fault code for it except for extreme conditions such as being unplugged. Every fault code has a long list of conditions that must be met for that code to be set in memory. Every list usually starts with a whole bunch of "xxx code must not be set". THAT is where mechanics got the notion to disconnect the battery to erase existing fault codes but that is counter-productive. Those original codes that set will be related to the current problem. Erasing them allows the conditions to exist that allows the other codes to set. The computer could have been watching the MAP sensor and trying to figure out which related code to set. Now that all the stored data is wiped out, all the computer might be able to figure out is the exhaust gas is staying too rich too long. That does not indicate a problem with the oxygen sensors; they're just the messengers. I used to read a lot about mechanics resetting GM computers by disconnecting the battery, but we never did that at the Chrysler dealership I worked at. Those '90s GM computers had a lot of issues, but on other brands nothing is going to be magically fixed by erasing the memory and starting over. The fuel trim numbers the computer learns over time are what it bases initial fuel metering on. Those numbers will tell you whether it's trying to reduce fuel without success or whether it's adding fuel in response to incorrect sensor information. As soon as the problem is repaired and the sensor information is correct, the short-term fuel trim numbers will begin to update immediately. You don't have to do anything except drive the vehicle.

As another point of interest, MAP sensors are so sensitive, they could be used to measure engine speed by counting the pulses of intake manifold vacuum each time a piston takes a gulp of air. If you watch the signal voltage on an oscilloscope, you would see four little voltage ripples for each crankshaft revolution on a V-8 engine. Those ripples are a part of how the computer figures out which fault code to set. Electrical codes are easier to understand. If the 5.0 volt supply is good and the ground wire has 0.2 volts, but the signal voltage is wrong, the sensor is suspect. But the computer can also detect pneumatic codes. On older engines that meant a leak in the vacuum hose or a defective sensor. On newer engines the sensor's vacuum connection plugs right into the intake manifold. No vacuum hose is used. While experimenting on prepared cars in the shop, we used to disconnect that vacuum hose and run the sensor with a hand vacuum pump to see how the engine responded. It was very common to set a "MAP Pneumatic" code even though we kept the signal voltage at the proper level. The computer set that code because it saw there were no ripples in the voltage and it knew that wasn't proper operation.

The reason your meter reading was bouncing around was due to those ripples. The meter takes a reading, thinks about it, displays it, then leaves it there while it goes back for another reading to think about. That one might have been taken, (like a snapshot), just when the voltage was at a high spot of that ripple. If you were to read that voltage on a scanner you would see it holding steady.

What you need to do is either drive it long enough to get a code to set or you'll need a scanner to view live data to see what's going on. If the MAP sensor is indeed reporting a load higher than actual, the computer is going to request more fuel. THAT one can easily double fuel consumption.

No matter what the oxygen sensors see, the computer can only modify that initial calculation plus or minus about ten percent based on their readings. If it subtracts fuel, you'll have a noticeable stumble or hesitation on acceleration. If it's adding fuel, you likely won't even notice unless you carefully keep track of your fuel mileage. Another clue when oxygen sensors or their readings are related to the problem is they are not even in the fuel metering calculation until they get to 600 degrees. For many years the O2 sensors have had electric heaters that turn on only when the engine is running to get them up to that temperature faster. Until that happens, the engine will run differently for that minute or two.

Caradiodoc
Jan 18, 2012.

AD