My S5 N/A is still suffering from low idle, at least when the car is hot.

Here are the symptoms,

- Idle fines at cold start.

- Low idle after driving a while. Car doesn't seem to stall as long as I am not stopping for long--e.g. stop sign is fine.

- Car stalls when I stop while going very slow--e.g. in a tight parking lot.

- Car seems to idle better under shade than under the sun.

All these lead me to suspect it has something to do with either with temperature sensing or air flow. Any help? Thanks!

 

Download the FSM for your S5 car. The link is in the FAQ section. It has an excellent flow chart on diagnosing problems with idle when hot.

Here are some quick suggestions:
1. Check for trouble codes
2. Timing
3. Advance
4. Set idle speed per FSM procedure
5. Set idle mixture per FSM procedure
6. Check TPS sensor for dropouts
7. Adjust TPS sensor
8. Check for vacuum leaks
9. Check BAC valve operation (pull the electrical connector off, idle should decrease)
10. Check EGR valve operation per FSM

 

I have heard that such a flow chart exist, but I am not sure where. The fuel and emission control system section (F1) has flow charts for each error code but not for idle when hot. Or are you referring to the table(s) that rank which component to check for "engine does not run smoothly (only when engine is hot)"?

I have already done most of quick suggestions you made,
1. No trouble codes
3. Did that
6 & 7. TPS in spec
9. Checked on both ends

Now I am particularly interested in 8. and 10.
8. If vacuum leak is the cause, shouldn't the car idle worse while running, since more air gets in through the leak?
10. Does S5 even have EGR?

 

Interesting: my car is an S4, and yours an S5. The diagram in your manual on pg. F1-3 shows an EGR, but there is no other mention of it in the manual. Apparently you do not have one and that diagram must be a misprint.

Your troubleshooting guide does indeed assume that all problems are accompanied by trouble codes. If you want, try using the troubleshooting guide in the S4 manual (also available in the FAQ); it looks to me like procedures 2, 4 and 6 might be applicable.

Vacuum leaks cause the most difficulty at low power and idle, because in those conditions the manifold vacuum is high, and the volume of air coming through the normal intake path is low. The high bacuum draws in more unmetered air through the leak, and it is a relatively high percentage of the total flow, and tends to either increase the idle, or if severe, run the mixture lean and cause rough running. At higher power settings, the vacuum is low, and less air is drawn in, and more air is flowing in the normal path, thus the air drawn in through the leak is a smaller percentage of the total, and has less effect.

I am not confident that your problem is a vac leak; it seems to me it ought to cause more problem during the cold condition (leaner mixture) than it does in the hot. Good to check anyway.

the talble of components to check for the engine does not run smoothly only when hot is a very good place to start.

But, before I did anything else I would check the voltage coming from the O2 sensor. At idle it should run rich (more than 0.45v; mine runs about 0.68 - 0.75 volts at idle at operating temp). The voltage might tell you if the problem is due to a too rich or too lean mixture. If the problem is low fuel delivery or a vacuum leak, you should see a low voltage (uner 0.45v); if it is going too rich, you would see a high voltage (more than 0.8v).

I had a very similar problem in my car, and it turned out to be an open in the wiring between the O2 sensor and the ECU. The ECU did not always correctly identify that the sensor was missing, and when it did not it tried to use the open (0v) 'signal' from the missing sensor. It saw that as lean, and ran the mixture rich until the car ran very poorly. It apparently fouled the plugs so badly that it ran poorly even at idle, when the O2 sensor in theory should make no difference.

 

My 02 sensor's reading is definitely lean during idle. I tested a while ago and I remember its below 0.5. It is unclear whether it is the O2's sensor malfunctioning or something else is causing the A/F ratio to be lean.

In your case, the ECU is getting a wrong reading of 0v, which is to say the reading is leaner than it really is. I wonder if a faulty O2 sensor can also cause this.

That said, a smog pre-test I did a few weeks ago gave a high HC reading, a very low CO reading and a slightly high but legal NO reading. I believe the later two would suggest my A/F ratio is running lean?

 

Probably vacuum related. Does your BAC valve operate properly? Often times the idle is set so low that the BAC valve simply can't overcome. You'd be able to tell just by disconnecting it at idle to see if the throttle drops/dies. Almost all idle issues will be one of the following: compression, vacuum leaks, TPS, lack of/overworked BAC.

 

As mentioned above, BAC is working on both ends--the valve itself is working and there is also signal from ECU. Idle has been adjusted, and I agree with calpatriot's assessment that vacuum leak is unlikely to be the cause, since a vacuum leak should show a stronger effect when the car is cold.

The crucial thing is, the car only idles poorly when it is hot.

I have just tested the intake air thermosensors and they seem fine. The next thing to check is the coolant thermosensor, but it seems I would need to take off the alternator to get to it.

 

Heat often has the effect of making vacuum leaks worse, as expansion generally can have negative effects. What is your idle when warm with the BAC unplugged with moderate load? Generally the car will run richer during warmup than when hot, which depending upon the size of a leak could play a role as well.

You could try to check the thermosensor @ the ECU, as if it's in range down there it'll save you some time.

 

As I understand it, expansion has negative effect in general because expanded air has lower oxygen content for the same volume. This would be a good thing when there is a vacuum leak, since this enriches A/F ratio, which is too lean due to the leaks.

Good idea on checking the thermosensor on the ECU side, it will certainly make my job easier. I can probably approximate the temperature of the coolant by measuring the temperature of the metal refill cap instead.

 

Hot air = less dense, less fuel. So as a vehicle warms up, and the ECU begins to remove even more fuel from the tables, it can make a vacuum leak even more noticable (depending upon the size of the leak) as the unmetered air will cause a lean scenario. If your hot idle seems to be rough or slightly uneven, something you can try to do is removing your airbox/filter and gently pushing the AFM flapper open. It often times gives you a good indication if everthing else has checked out well. I know how frustrating it can be to look over everything and spend forever searching for an idle issue, especially when you've checked and double checked for vacuum leaks but i'd wager almost 90% of the time it always ends up being vacuum related in the end. Sometimes though, the mechanical idle can be set a hair too low. I almost always set the idle to around 730rpm with the BAC disabled, just so I know the BAC has full operation to maintain the idle under load/varying weather/etc.

 

due to the air pump injecting air into the exhaust ports, the o2 sensor SHOULD read lean at idle. if you put a wideband on it mixture will read in the 16-18:1 area.

furthermore the TPS and idle should be set HOT.

 

With respect, no. Once the car is warmed up, the O2 sensor should read well into the rich range (more than 0.45v) at idle. FSM pg F1-80 lists the O2 sensor reading at idle should be 'less than 1.0v'. 1v is the max (full rich) output.

Mine runs, once warmed up, at between 0.65 and 0.78 v at idle.

This is almost definitely related to your problem.

The lean mixture at idle can be caused by a variety of factors, such as unmetered air (vacuum leak); improper metering (faulty AFM); clogged injectors; low fuel pressure; faulty wiring to an injector; bad ECU (not likely, but possible); other sensors not functioning correctly (boost, air pressure, throttle position).

Pull codes. Check for vacuum leaks.

 

I spent the evening doing more checking, and here is what I know,

First, it seems a lean mixture is indeed the problem.
Second, vacuum leak is probably not the cause.

I did two things--first I sprayed water around the engine, which affected absolutely nothing. I then hooked up a new oxygen sensor and connect it to a voltmeter. Here is when things get interesting. The mixture was rich (0.6-0.7) initially, until it all a sudden drops to below 0.3. The drop was very discreet, taking less than a second.

What could have caused this? Does S5 has some sort of fuel cut or air dump feature?

One more piece of information is, the drop seems to occur soon after the stock electric fan comes on--just realize my S5 is a manual-swap, but that's another story. Could it be some sort of heat protection being triggered by a faulty thermosensor?

 

that must be a misprint, if you were to say, put a wideband in, they idle between 16-18:1, due to the extra air.

 

they DO have a heat protection thing, but it would have turned on the exhaust overheat light. it also goes the other way, it dumps the air pump air to atmosphere and goes way way rich to try and cool itself off.

normally the air pump air is going to the exhaust ports, and the air fuel is "normal"

 

Given the discreet nature if the drop in rpm I am starting to consider the possibility that the ECU is malfunctioning. Anyone has experience with ECU not working?

 

They usually fail in such a way that it's catastrophic. I do find it amusing that you say a lean mixture is likely the cause, yet you dismiss the absolute leading cause of lean mixtures @ idle. Start the car, let it get warm, and push the AFM door in. It'll fix your mixture and see if the car idles any better.

 

By the absolute leading cause I guess you mean vacuum leak. I am not completely dismissing that possibility, but at the very least I wasn't able to locate any. Furthermore, lean mixture is only likely to be the cause if the idle mixture shouldn't be lean, and in this thread alone I see opinions ranging from lean to rich.

Now for some new findings--I just tried your suggestion of pushing the AFM door. the car stalled on the slightest force, and reading from the oxygen sensor jumped up to 0.9v. I believe this means my engine stalled because the A/F ratio is too rich, which could mean my idle mixture is actually too rich to start with? If that is the case, it would explain why my FC seems to do fine as long as I am moving along, as more air enters the engine when the car is moving.

On a separately thread earlier on, I mentioned that the resistance of my AFM is non-monotonic. At both ends it meets FSM specifications--which isn't saying much, since the spec. accept a really large range of readings--but it actually peaks somewhere in the middle. Could that be the cause?

 

The AFM's aren't linear. It's interesting that it want's to stall with the AFM, were you careful to not choke the air supply off at all while doing so? The problem with the stock O2 sensor is it's range is so pathetically narrow that even trying to diagnose something as simple as the idle is probably asking to much. at .9v the AFR could be from about 12.8-13.5, and it can vary in response depending on the EGT as well. They tend to idle best around 12:1 to 13:1 (varying on car), but most of them seem to run best near the low 12's. I checked 3 car's that idle happily earlier today just to kind of compare, and every one of them was over .9 at idle. Most narrow bands are well into the 14's ~.6v, which would be way lean and give you troubles.

 

I have been informed a few times that the AFM's reading is non-linear, but I not on its non-monotonicity. Maybe by non-linear people also means non-monotonic, which is wrong mathematically.

Just tried the pushing the AFM door again with a screwdriver instead of my hand. Still stalls almost immediately.

 

To confirm, it's definitely not monotonic.

The AFM door thing opens up quite a few idea's as it definitely leads me to believe otherwise with regard to being lean. Do you happen to have a vacuum gauge or something similar? If so check the FPR line, is it pulling vacuum @ the FPR? If you don't have a gauge, just pull it off and see if it's pulling vacuum at idle. I had a car that I spent quite some time fiddling with just the other week that had some hot start/hot idle issues and we eventually tracked it down to being plugged into the wrong vacuum nipple (it saw no vacuum, ever). Switched them around and it took care of most of the problems all at once.


Edit: What does the car idle at when hot anyway? What does it tend to drop to when you're coming to a stop/etc?