Driving isn't necessarily the fix. The Engine Computer has to see high manifold vacuum for at least seven seconds to relearn minimum throttle. That may not happen from normal driving unless you coast off an off-ramp.
The computer is looking for the lowest voltage reading it sees from the throttle position sensor. Once it knows, or "learns" that reading, any time after that it sees that same voltage, it will know your foot is off the accelerator pedal and it has to be in control of idle speed. Until that is relearned, idle speed can appear to be okay, and it could drop too low at times, especially when shifted into gear.
The high manifold vacuum is what tells the computer the accelerator pedal is released, but you can achieve that by snapping the throttle in the shop. The only way to see that for seven seconds is by prolonged coasting, and the only way to have that is seeing road speed. The things that will abort the relearn include even the slightest change in TPS reading during the seven seconds, a momentary loss of vacuum, typically due to a failing MAP sensor, which was real common many years ago or tapping the brake pedal during the seven second coast. A tight throttle cable and a misadjusted brake light switch could interfere with that relearn.
A definite way to know if the relearn has taken place is to view the "idle steps" on a scanner. GM and Chrysler used the same idle speed motor. It has a slowly rotating armature inside four electromagnetic coils that are pulsed with varying voltages and polarities by the computer. As that armature rotates, it extends or retracts a pintle valve on the end of its threaded shaft. That valve opens the air bypass passage around the throttle blade. At the same time the computer varies the length of time the injectors are pulsed open. Together that's how it controls idle speed.
The computer can set that armature to one of 256 positions or "steps". For a properly running engine, step 32 is typical. With a single-cylinder misfire on a V-6 engine, you can expect to find it at around step 50. The secret here is when minimum throttle hasn't been relearned yet, you'll find it at step 0.
The first thing I'd do if you do find minimum throttle won't relearn on a test-drive is to hold the brake pedal up with your foot while you're coasting. I've solved more than one by doing that.
If you see a step number shown on the scanner, as I mentioned, typically around step 32, give or take, minimum throttle has been learned, and we have to look somewhere else. A lot of our experts talk about cleaning the throttle blades of carbon. I've never had to do that on any car, but that doesn't mean this isn't a valid solution. The next step would be to use the scanner in "record" mode to take a snapshot of what's happening when the stalling occurs. I suspect you're familiar with that, but for the benefit of others who might be researching this topic, when the "record" button is pressed, the scanner will record about five seconds worth of sensor data that can be replayed slowly, later, to see what changed when the problem occurred. Because this data passes through the scanner's memory, the recording actually begins a couple of seconds before you pressed the switch, so it should catch the data starting before the stalling occurred. Some sensors are inputs to the Engine Computer, and the computer bases its fuel-metering and ignition timing decisions based on those readings. Other sensors, mainly oxygen sensors, report to the computer what happened as a result of those decisions. Often it takes an experienced engine performance specialist to interpret those readings. A typical example might be a loss of engine speed when the idle step and fuel injector pulse width do not change. You might look for an EGR valve that is being held open by a chip of carbon.
The more expensive scanners also have graphing capabilities. This also takes some experience to interpret. You have to know which sensors or engine operating characteristics, (idle speed, for example), you want to include, then the point of interest is observing when things change when the problem occurs, which one occurred first? Did the engine stall because idle speed dropped too low, or did idle speed drop too low because a cylinder started misfiring? The sequence the events occurred in will show which one was the cause and which ones were the result.
Also, of interest is whether you can keep the engine running when the stalling can be expected to occur. For example, if you know it's likely to stall when approaching a stop sign, holding the accelerator pedal down 1/4" will prevent that if minimum throttle hasn't been learned.
After rereading your initial comments, this could also have nothing to do with idle speed. Based on extended idling time, the camshaft position sensor and crankshaft position sensor become good suspects. Those often fail on any brand of vehicle by becoming heat sensitive. A real common failure is for them to work fine as long as the vehicle is being driven. Natural air flow keeps them cool. While standing still with the engine running, or more commonly, after stopping a hot engine for a short time, as in when stopping for gas, engine heat migrates up to those sensors causing one to fail. When this failure starts to show up, it is very often intermittent at first and the sensor will start to work again after cooling down for about an hour.
One of the frustrating parts to finding an intermittent sensor is it takes the computer some time to detect the missing signal and set a diagnostic fault code related to it. For that reason, it's rare to find that code is set just from cranking the engine after any codes have been erased. Even stalling while idling might take place too quickly for a fault code to set. Typically, the computer needs more time, as in when a stalled engine is coasting to a stop, to set the appropriate fault code. This is another place the graphing scanner is priceless. Even viewing live data isn't going to catch this. I have a Chrysler DRB3 for all of my vehicles. They list the cam and crank sensors with a "No", or "Present" to show if those signals are showing up at the computer. They both must switch to "Present" during cranking. The problem is takes too much time for the scanner to update its readings. While expensive scanners are real fast, you're going to find a signal could drop out, the engine stalls, and then both sensors switch to "No" because the engine isn't rotating. You won't be able to see that the signal dropped out first, but you can on a graphing scanner.
Tell me if it sounds like these are valid things to pursue or if we have to consider something else.
Tuesday, October 12th, 2021 AT 12:44 PM
