Check engine light on

Tiny
JSTAIB1
  • MEMBER
  • 1995 JEEP CHEROKEE
  • 4.0L
  • 6 CYL
  • 4WD
  • AUTOMATIC
  • 88,000 MILES
My check engine light comes on when I am driving and my rpm's get around 2000. Then it starts to kick or buck (only way I can describe, misfire maybe) almost like it is running out of gas. My local mechanic said he did not see anything significant and would have to charge me couple hundred to check it out. Any suggestions on what this could be? I am fairly handy and would love to try and fix myself.
Do you
have the same problem?
Yes
No
Wednesday, December 28th, 2016 AT 12:02 PM

9 Replies

Tiny
HMAC300
  • EXPERT
Scan for codes to see what is causing it, auto parts does for free. Also, check fuel pressure with a gauge auto parts rent it.

https://www.2carpros.com/articles/how-to-check-fuel-system-pressure-and-regulator
Was this
answer
helpful?
Yes
No
+1
Wednesday, December 28th, 2016 AT 12:04 PM
Tiny
JSTAIB1
  • MEMBER
The mechanic says that the only way to scan for code on the check engine light on a 1985 keep is when the light is on. The light only comes on when I am driving. And suggestions on how to scan when light is not on?
Was this
answer
helpful?
Yes
No
Wednesday, December 28th, 2016 AT 3:28 PM
Tiny
JSTAIB1
  • MEMBER
1995 not 1985 keep Cherokee.
Was this
answer
helpful?
Yes
No
Wednesday, December 28th, 2016 AT 4:01 PM
Tiny
HMAC300
  • EXPERT
If the light is on it is stored in the computer unless it is an intermittent. So do like I said in reply.
Was this
answer
helpful?
Yes
No
Wednesday, December 28th, 2016 AT 4:50 PM
Tiny
CARADIODOC
  • EXPERT
Sorry to butt in on brother HMAC300 again, but your mechanic is wrong. Not all diagnostic fault codes turn the check engine light on, so what would be the point of having them if they could not be read?

There may be one problem with having the people at an auto parts store read the codes. Often they can only read them on 1996 and newer vehicles. That is when the OBD2, (on-board diagnostics, version two) emissions system was mandated to be used for all car brands sold in the U.S. Every manufacturer had their own proprietary system up to that point, and many auto parts stores no longer have the equipment to read codes on those cars.

You have it easier though. Chrysler made reading the fault codes yourself much easier than any other manufacturer. Cycle the ignition switch from "off" to "run" three times within five seconds without cranking the engine. Leave it in "run", then count the flashes of the check engine light. There will be from one to five flashes for the first digit, a short pause, then a series of flashes for the second digit, followed by a longer pause before the next code flashes the same way. If you think you miscounted, turn the ignition switch off, then back on once. Cranking the engine takes the computer out of test mode. The last code will be "55" which just means "end of message". If you get code 12, that can be ignored. It just means the battery was recently disconnected.

There are dozens of potential fault codes, but only about half of them refer to things that could adversely affect emissions. Those are the codes that turn on the check engine light. The whole purpose of these fault codes is for the engine computer to store that information when it detects a problem, so it can be read and recorded later. Without that capability, there is no way your mechanic would have any idea which circuit or system to diagnose further for the cause of an intermittent problem.

What your mechanic might have meant, is if you are experiencing an intermittent problem, the check engine light might turn on during that event, and that is when the problem is occurring. Even though the problem stops acting up, and the light goes off, the fault code is still in memory and can be read. If the problem does not act up for a long period of time, the codes will self-erase after fifty engine starts. They can also be erased by disconnecting the negative battery cable for a few seconds, but that is not the goal until the problem has been solved. Erasing codes never makes a problem go away. Instead, the goal is to avoid disconnecting the battery or allowing it to run dead because you would lose that valuable information.

You can get a general idea of the severity of the problem by how the check engine light acts. As I mentioned, the light will never turn on for about half of the problems that can be detected. The best example of that is the code, "Engine running cold too long". The coolant must reach a specific temperature within six minutes of starting the engine. When it takes too long, that code is set. Problem is every Chrysler product in the northern winter climate will have that code in memory if the owner allows the engine to idle too long to warm up. That code never turns on the check engine light.

For those codes that do turn the light on, when the problem is intermittent and it stops acting up, the light will turn off while you are driving if the cause is relatively minor. If the problem is more severe, and it stops acting up, the light will remain "latched" on until you turn the ignition switch off and restart the engine. Then it will remain off until the problem acts up again. If it is still more serious, even if the problem is not currently acting up, the light will be on all the time. That indicates the most serious of conditions as far as increased emissions, not necessarily seriousness of repair cost. Where it can cause real problems though is when people ignore the light, or put tape over it.

There is one more condition that can occur on the 1996 and newer cars. That is the check engine light is flashing. That means you are supposed to stop the engine right away because too much unburned gas is going into the exhaust system where it will burn in the catalytic converter and overheat it. That can melt the catalyst and plug the exhaust system. That can turn a minor repair, like a worn spark plug, into a very expensive repair.

Two things can happen when the check engine light is ignored. First, the problem could be relatively minor and inexpensive to repair, but what if a really serious problem develops that will turn into an expensive repair? You will never know it because the check engine light is already on and you will continue to ignore it.

The other potential problem is the computer is constantly testing all of its sensors and other circuitry, and it looks for discrepancies between various operating conditions and sensor readings. There is always a long list of conditions that must be met for any fault code to be set, and one of those conditions is certain other, related codes cannot already be set. For example, the computer knows when the engine has been off for at least six hours, the coolant temperature sensor and the intake air temperature sensor had better be reporting the same temperature. When they disagree, it has ways of determining which one is wrong. However, if you unplug the intake air temperature sensor, a fault code will immediately be set related to that. Even though you reconnect it, the computer may assume that sensor's readings cannot be trusted, so it will suspend any other tests that rely on that sensor for comparisons. If an actual problem develops with the coolant temperature sensor circuit, and the tests are suspended, the defect may go undetected.

Where we run into a problem is we read the codes to be able to give you a rough estimate for the cost of repairs. At least we know which circuit we need to look at. You give approval, and the mechanic starts the diagnosis. Most of the time the repairs are fairly straightforward. Half of the time there ss a bad sensor. Half of the time the wiring and connectors for that sensor are the cause of the fault code. Once the repairs are authorized by you, and completed, the mechanic erases the fault codes and goes on a test-drive. That is when the suspended tests resume, and any additional, related problems are first detected. The check engine light turns right back on again, and we have to tell you there are more problems to diagnose and more parts will be needed. Having to tell you that is extremely frustrating for us, and uninformed car owners assume we are trying to defraud them or sell things that aren't needed. They also often assume we did not diagnose the original problem correctly or repair it correctly. This is an especially big problem on GM vehicles related to defective anti-lock brake wheel speed sensors. Multiple failures are very common because they happen so frequently, but only the first one will be detected. It is not until that one is replaced that the tests resume, and the second defective sensor is detected. This is mostly a problem when the first problem is ignored for a long time, as in "when I get the money". When the first problem is repaired right away, it becomes evident it was repaired correctly when the warning light no longer turns on.

HMAC300 is a busy boy and does not have as much time as I do to dispense all this wondrous information. Hopefully this will help you help him solve your problem. If I had to guess, I would suspect a problem with the crankshaft position sensor, distributor pickup assembly, or MAP sensor, but I am only saying that so I can say, "I told you so", when you post the solution.
Was this
answer
helpful?
Yes
No
Wednesday, December 28th, 2016 AT 4:59 PM
Tiny
JSTAIB1
  • MEMBER
Very helpful. We think the code we got was 12-24-55. I am assuming the 12 is from what you said above and the 55 was the end of code. I googled the 24 and it says throttle position sensor input above maximum acceptable voltage or below. Does this sound like it would be what I described above? When the rpm's are above 2000 it feels like jerky? I added our video of the light flashes
Was this
answer
helpful?
Yes
No
Wednesday, December 28th, 2016 AT 5:31 PM
Tiny
CARADIODOC
  • EXPERT
To describe further, the TPS is fed with 5.0 volts and ground, (0.2 volts). The movable contact picks a voltage along its range of travel between those two extremes, but it has mechanical stops that prevent the signal voltage from going below roughly 0.5 volt or above 4.5 volts. Those numbers are approximate and are different for every sensor, but they are used for explaining theory.

With a broken ground wire, 5.0 volts will show up on the signal wire, regardless of throttle position. That is an unacceptable voltage and will trigger a fault code, "TPS voltage too high". If the 5.0 volt feed wire is broken, 0.2 volts will always show up on the signal wire. That is also an unacceptable voltage and will also trigger a fault code, "TPS voltage too low". Those are fairly easy to explain.

A break in the signal wire is more involved. If the wire is cut, you'll still see normal signal voltages at the sensor, but not at the computer. Because the wire at the computer is interconnected to all the other circuitry inside the computer, the signal voltage the computer might see can "float" to some unknown random value. As long as that random value remains between 0.5 and 4.5 volts, no fault code will be set, and the computer will try to make fuel metering calculations based on those incorrect readings. To prevent that, all sensor signal wires for all computers use a "pull-up" or "pull-down" resistor inside the computer. Those resistors are so extremely high in value that for all practical purposes, for a properly-working circuit, it's like they aren't even there. It's only when there's a break in the signal circuit that the resistor places 5.0 volts or 0.0 volts on that wire to force it to go to a defect condition that will be detected. You only know the circuit that needs to be diagnosed. Voltage measurements will lead you to the cause. Most Chrysler computers use the pull-up resistor, so 5.0 volts will appear in the computer. The fault code will again be, "TPS voltage too high", but the cause is different than with a broken ground wire. This is why you need a few voltage measurements to verify the cause before just replacing the sensor.

When you have an intermittent problem, you will not always find it with voltage readings because most digital meters respond too slowly. The engine computer will detect tiny glitches that occur too quickly to be seen with a voltmeter. Your mechanic would start by connecting a scanner to view live data, then he'd watch the TPS reading as he worked the accelerator pedal. If the signal voltage is always high, taking individual voltage readings is the next step. If the signal voltage appears to be normal until around half or three quarter throttle is reached, that is almost certain to be caused by a dirty or worn contact inside the sensor. That is the time to replace the sensor.

The TPS will not cause a crank/no-start condition but it can cause running problems. Besides throttle position, it tells the computer when you are at idle, (and the computer must be in control of idle speed), when you're at wide-open-throttle, (and it should turn off the AC compressor and alternator to leave more power for passing that freight train), the direction of throttle change, (to reduce fuel momentarily while slowing down, to increase fuel mileage), and rate of change to improve engine response.

When the computer sees an unacceptable TPS signal voltage, it will disregard it and "inject" an approximate value to run on, based on other sensor readings and operating conditions. Those injected values are never perfect so you will usually observe some type of symptom. For basic fuel metering calculations, the TPS is probably the least important sensor. The most influential sensor is the MAP sensor. Chrysler is the only manufacturer that uses only the MAP sensor to measure engine load, (vacuum), as the main component in fuel metering calculations. All other manufacturers need a mass air flow sensor to measure the weight of the incoming air so it can be matched with the corresponding amount of fuel. Most of those cars also use a MAP sensor, but only for barometric air pressure readings before the engine starts, then as a backup strategy when a problem with the mass air flow sensor is detected.
Was this
answer
helpful?
Yes
No
Wednesday, December 28th, 2016 AT 7:11 PM
Tiny
JSTAIB1
  • MEMBER
Changed out the TPS this morning for $37.00 and it fixed the problem!
Thank you so much!
Was this
answer
helpful?
Yes
No
Thursday, December 29th, 2016 AT 6:42 AM
Tiny
CARADIODOC
  • EXPERT
Dandy. Happy to hear it is solved.
Was this
answer
helpful?
Yes
No
Thursday, December 29th, 2016 AT 11:55 AM

Please login or register to post a reply.

Recommended Guides