Theory only here
My car has an idle issue where when I stop at a light, the RPM (ideally around 700 @ idle) will droop down to ~500, then overcomensate and shoot up to ~1000, the re-overcompensate and drop to 200, then shoot back up to 1200, then go so low that it dies. Reminds me of an unstable PID loop. Does a car use a PID loop to idle? I never gave it much thought, but I guess if your foot is not on the gas, it would have to have some kind of control loop or else it would die when you turn on the AC + headlights or just over rev until it blew up.

I'm talking late model car here, electronic fuel injection + it's had a "speed density tune" - basically meaning there's no mass air flow sensor anymore.

 

There are a LOT of sensors in the loop; absolute pressure (MAP) sensors for ambient and manifold pressure, RPM sensor(s), oxygen sensors, temp sensors, throttle position sensors, the list goes on.

If your vehicle has a lot of miles on it, the O2 sensor might be going bad, you might have deposits in your EGR system, a plugged port on your intake manifold, a leak in your vacuum controls, or a number of other items.

Any fault in any of the feedback loops can cause erratic operation. Your fuel economy and engine life will suffer if you don't get it corrected.

I had an odd problem with one of my vehicles last summer; the "check engine" light would come on after driving over 45mph for a period of time. Turned out to be a clogged EGR port in the intake manifold.

Keeping your vehicle in good tune (and serviced according to the manufacturer's specifications) will ensure a good trouble-free service life from the vehicle.

 

Wouldn't hurt to check for a vacuum leak also. A leak would cause a high idle, which in turn would try to be corrected by the ecm..

 

I created a problem like that by replacing a small diameter, hard plastic vacuum tube with an old fashioned piece of rubber vacuum tubing. It apparently changed the timing of whatever the sensor was sensing and the engine would go through backfires and correction if you held the accelerator in just the wrong place. The moral of the story: Put it back the way it was.

 

Yep, the systems are really touchy nowadays.

I mentioned O2 (lambda) sensor(s) earlier. As they age, their response slows down, which can cause the engine speed to vary as the air/fuel mixture is changed too much because the O2 sensor is slow to report the actual change.

The original lambda O2 sensors are still in use. They only report "oxygen present" or "no oxygen present", which translates to "too lean" and "too rich". The ECU constantly adjusts the fuel mixture up and down, always seeking the opposite side of the proverbial fence to get the mix as close to 14.7:1 (stoichiometric) as it can.

Later "wideband" O2 sensors can report the actual air/fuel ratio. I think they first appeared on Volvos. Late model Chevrolet Corvettes have them. I don't know how many manufacturers/models have changed over to the wideband O2 sensors, but they are certainly an improvement over the lambda sensors.

Anyway, you really need to keep your vehicle as close to "as manufactured" as possible, as so many systems are interrelated that changing something somewhere can very well affect something on the other end of the vehicle.

The wife's car occasionally had the "check engine" light coming on. Turned out that she was not tightening down the fuel tank cap after adding gasoline; and the system detected an air leak in the fuel tank.

 

Hi Strantor. The most common reason a vehicle looses control of Idle speed is when the battery has been replaced. Here in Australia Ford Falcons are what i worked on. This fault was cured different ways with each different model. The early ones all you needed to do was drive the car till it was at operating temperature then pullup in the vehicle. Put it in park if auto or in neutral if manual. Then let it idle for 1.5Minutes to re learn its Idle position.(all accesorys must be off &power seering not moved) as these will upset the learning process.

 

Heres a bit of info from a Ford course I did a few years ago on Engine RPM Management. Just to give you a bit of an idea of how it works.

 

Hi Strantor. The most common reason a vehicle looses control of Idle speed is when the battery has been replaced. Here in Australia Ford Falcons are what i worked on. This fault was cured different ways with each different model. The early ones all you needed to do was drive the car till it was at operating temperature then pullup in the vehicle. Put it in park if auto or in neutral if manual. Then let it idle for 1.5Minutes to re learn its Idle position.(all accesorys must be off &power seering not moved) as these will upset the learning process.

The statement that this is the most common "reason" that an unspecified vehicle (old/new? car/truck? gasoline/diesel?) loses idle speed control seems a very sweeping one, and not something I had heard of before. It is not clear to me whether experience with Ford Falcons can be generalised in this way, but here I must admit to knowing little about modern engines.

Does this mean that a specific "running-in" protocol is required on battery replacement for many current vehicles? Should battery replacement therefore only be performed by a dealership for the model concerned, who can reprogram the system to cope with the new conditions?

 

Quite a few ECU controled engine management systems will loose engine Idle memory if the battery has been removed. Ive tried to give a small insite to how the system works just for strantors curiosity. Every manufacturer will have a slightly different aproach. As a mechanic its not easy diagnosing faults in vehicles these days as there isnt a lot of info avalable un less you are working for a dealership. & code readers dont always diagnose faults.

 

what system does your vehicle employ?

an air bypass valve placed to shunt air across the throttle body is a typical strategy. Nested PID loops would be the norm.

I'd look for that valve (or device) and check it for smooth operation (cleaniness)

 

Heres a bit of info from a Ford course I did a few years ago on Engine RPM Management. Just to give you a bit of an idea of how it works.

Thanks! that's a great writeup, I would assume that most manufacturers use the same or close strategy.

what system does your vehicle employ?

an air bypass valve placed to shunt air across the throttle body is a typical strategy. Nested PID loops would be the norm.

I'd look for that valve (or device) and check it for smooth operation (cleaniness)

Not sure, I will have to look

 

This may seem rather obvious but if replacing the battery causes the ECU to loose memory, then why not jump the system with a 12V lantern battery - Diode isolator before you remove the battery?

 

This may seem rather obvious but if replacing the battery causes the ECU to loose memory, then why not jump the system with a 12V lantern battery - Diode isolator before you remove the battery?

They make a device that has a 9V battery, that you plug into the cigarette lighter socket . It is used when changing the battery, to keep ECM and radio memories alive. The 9V is enough since the memories are at 5V.
http://www.amazon.com/CTA-Tools-Computer-Memory-Saver/dp/B000GKD5Q0

 

They make a device that has a 9V battery, that you plug into the cigarette lighter socket .

Can't say that I'm surprised.

 

Well I'm surprised. I'm surprised that ECU's would be designed to lose important information like that during a battery change. Thank you for telling me (and a lot of other people) about this and how to avoid having a problem with it!

 

A friend of mine has a new car with electric door locks. He works on ships for 10 or 15 weeks, and when he got home the battery had dropped so far the door locks would not work and he could not open the doors OR the bonnet (hood) as that is operated from INSIDE the car of course!

So he could not open the car to charge the battery.

I made him a very simple maintenance charger that puts out 13.0v at about 300mA and he now connects that to the car when he goes away.

 

I'm pretty sure that the only thing lost on battery disconnect is the adaptive or learnt data, along with error codes. Loss of such sends the ecu back to mapped data, Something an operator wouldn't notice.

 

I'm pretty sure that the only thing lost on battery disconnect is the adaptive or learnt data, along with error codes. Loss of such sends the ecu back to mapped data, Something an operator wouldn't notice.

Typically correct.
Many modern cars have an IAC motor which bypasses the throttle plate to allow air to enter/exit to maintain idle.
Idle hunting can be related to many issues.. Could be a vacuum leak, MAP/MAF sensor issue, TPS sensor issue, IAC motor dirty/broken,etc...

 

There's a video on this page: http://www.wonderhowto.com/how-to-test-and-replace-your-cars-idle-air-control-motor-267188/
regarding testing/cleaning/replacing IAC motors.