Typical failure symptoms of individual components:

Throttle Min Max Switch either out of adjustment, intermittent failure, or hard failure:
1. High and/or unstable/surging idle speed

Idle Stabilizer: (should be accompanied by Check Engine light, but not necessarily in practice)
1. Engine dies at idle or if does idle is rough/surging
2. Engine runs at off-idle throttle settings
3. Can be intermittent or a hard-fail

False Air Leak:
1. Higher idle speed
2. Poor idle quality
3. Can have one with little notice since the system will try and compensate

Fuel Pressure Regulator:(leaking raw fuel into intake through the vacuum hose)
1. Engine stalls or runs very poorly at idle or returning to idle
2. Raw fuel smell

Motronic/ECU Coolant Temp Probe:
1. Loss of signal to the ECU will result in a momentary spike of engine RPM until the ECU recognises that the temp value is bad and uses a default value. Idle can be slightly rough/unstable.
2. Can be intermittent due to a broken wire

________________________

Detailed description of direct compenents of the Idle System.

Idle Stabilizer (IS) (a.k.a. the Idle Air Control Valve, IAV, IACV)

The Idle Stabilizer is mounted on the rear of the plenum and regulate the amount of air allowed into the engine at idle. It is common to several Bosch Motronic systems. It is controlled by the ECU and only powered at idle. It is basically a stepper motor that operate a rotary valve. It is what is known as a duty-cycle motor. The ECU sends variable length pulses (5-95% range) to the IS which commands the valve how much the valve is held open against the return spring. It is spring loaded to the minimum position. Supposedly, when the idle air stabilizer valve is functioning correctly, it should vibrate and hum slightly whenever the engine is running at idle. However, from my experience, if it is humming slightly, it's too slight to hear or feel. Off-idle, all air goes through the normal Throttle Body. Hence you can have a failed IS and not know it until you close the throttle and command an idle. With a failed IS, it's likely the engine will stumble badly and probably die when the throttle is let off and it approaches idle. You can usually prevent it dying initially with a little throttle, but if it does idle, the idle will be low, rough, and surging at first, then settling down some. But certainly not normal. If the IS fails while at idle, the engine will likely die, but will restart with some open throttle. Although monitored by the Check Engine light, I have disconnected mine with the engine running and not received a CE light

 

Throttle Min/Max Switch

A further component is the Throttle Min/Max Switch (small black box attached to the front side of the main throttle body).

The Throttle Min/Max Switch is simply a three position on-off-on switch. It’s a ground switch with the middle terminal (18) being grounded. Its job is to tell the ECU when the throttle is fully closed (idle) and when it’s about 75% open. Hence, there are two separate circuits/contacts inside the Mix/Max Switch. With the throttle fully closed, the Min side of the switch is actuated and a signal sent to the ECU. The ECU then leans the mixture slightly for idle.

In decel mode, like with the engine at speed (the ECU knows via the Crank Sensor) and you close the throttle, the Min/Max Switch closes and fuel delivery is cutoff completely if the engine is above 1100 rpm or so. Actual fuel cut/restore points are determined by a complex mix of engine speed, engine temp and engine load, but to keep it simple we'll say about 1100 rpm.

Once the engine decels below this rpm, the ECU restores fuel delivery and allows the engine to recover to idle. At full throttle operation, the MAX side of the switch is closed, the ECU commands a slightly richer mixture to prevent detonation at full power demand. Between idle and full throttle, the Min/Max switch is not actuated (off) and inputs nothing to the ECU for normal running.

While at idle, if the connector is removed or the switch fails, the idle speed will immediately increase to about 1500 rpm. Again, the Min/Max Switch is monitored by the CE light, but I have disconnected mine and not received a CE light.

____________________________

Diagnosis

First thing to do is run the Self-Diagnosis (Self-Test) of the ECU. The Electrical Manual Vol, page 602 has the procedure for both Test A (extracting check engine stored codes) and Test B (actual test of the injectors, the Idle Stabilizer, and the Evaporation Solenoid Valve). It's possible that you can get no codes, or random/intermittent codes, yet your idle problem remains. So, crosscheck the CEL codes, but they are not necessarily reliable for final diagnosis. For that, you'll need to look at the individual components and their failure modes.

I. Idles too high (1500-2500 rpm) or idle surges between approx 1200 to 2500 rpm.

Suspects:
a. Throttle cable adjusted too tight not allowing throttle valve to completely close (idles too high only)
b. Maladjusted Min/Max Switch (too high and/or surging)
c. False air (vacuum) leak (too high, slightly rough, or surging)
d. Dirty/broken electrical connector to Min/Max Switch (too high and/or surging)
e. Failed Min/Max Switch (too high and/or surging)

II. Engine dies or almost dies when returning to idle, either intermittently or consistently. Engine will generally restart (perhaps roughly) with accelerator pedal depressed, but will idle roughly or not at all. Engine will run if held above idle.

Suspects:
a. Fuel Pressure Regulator leaking raw fuel into plenum
b. Idle Stabilizer dirty
c. Idle Stabilizer electrical connector dirty/broken
d. Failed Idle Stabilizer

III. Rough idle/misfire.

Suspects:
a. Bad wire to Motronic coolant temp sender or bad sender itself
a. Bad distributor cap or burned out resistor in rotor
b. Bad plug wires/spark plugs.
c. False air leak


Troubleshooting

Do a quick assessment of the possible problem areas. Check the quick and easy stuff first.

1. Check for any signs of air leaks by looking at the large rubber connector between the AFM and the throttle valve. Check hose clamps are snug. Check the hoses to the crankcase vent system and the Idle Stabilizer. With the engine running spray carb cleaner around any suspicious areas and listen for change of tone or tempo in engine that would indicate a leak. One at a time, remove and cap the three vacuum lines to the plenum, looking for improvement after each cap-off in order to isolate a possible leak.

2. Pull the small hose from the Fuel Pressure Regulator to the intake plenum. Check for evidence of raw fuel.

3. Check that Throttle Valve is reaching the full idle stop position.

FIXES:

1. Throttle cable too tight. Remove the plastic cover on the throttle valve relay crank and adjust the throttle cable so there is a very slight slack in it at idle. Ensure the relay crank is fully against the stop at idle.

2. False air (vacuum) leak. Broken or cracked intake hoses should be replaced. As a temporary measure, JB Weld or tube silicon can be used to plug the leak in the AFM to Throttle Valve connector.

3. Fuel Pressure Regulator leak. Replace FPR

 

4. Idle Stabilizer. The most common fault with the Idle Stabilizer is that it gets dirty and the modulation valve sticks. Or, it can hard-fail, or fail intermittently. The intermittent failures can be hard to diagnose, especially if the CEL throws no code at all . . . . or, even more confusing, an incorrect Check Engine Light (CEL) code.

If the engine is above idle when the failure occurs, you probably won't notice it until the accelerator is allowed to come back to idle (closed). Then the engine dies or recovers very roughly to idle. Unless the accelerator is depressed, the engine will not start. Pressing the accelerator and keeping it above idle will usually result in a restart.

Remove the Idle Stabilizer from the rear of the plenum. To remove, take out the one machine screw and capture clamp holding the IS onto the plenum. Disconnect the electrical connector. Then simply pull the unit out from the plenum. This may take a bit or force due to the slip-fit rubber grommet that joins the IS to the plenum. Once off, spray brake or throttle body cleaner in the end and swish it around. Repeat and allow it to dry when clean. Then spray a tiny amount of WD-40 (or equivalent) in the intake hole of the IS and swish it around to lubricate the mechanism.

When reinstalling, the rubber grommet joining the IS to the plenum will likely not make a tight fit anymore due to the age and wear. Wrap some electrical tape around it to make it fit snugly and without air leaks in the plenum. Reconnect the electrical connector and start the engine. Check for normal operation.

5. Min/Max Switch Adjustment.

You can adjust the Min/Max switch using your feel/hearing, or using an ohm meter. I prefer using an ohm meter because it tells you exactly when the switch trips and verifies that the contacts within the switch are functioning normally.

a. Disconnect the electrical connector and set it aside.

b. Use a 9/32nd ignition wrench to loosen the two machine screws holding the Min/Max Switch onto the throttle body. Unfortunately, there is insufficient space around the machine screws to use a socket. The throttle valve shaft goes all the way through the throttle body and into the Min/Max Switch. The switch rotates within an adjustment range around that shaft.

c. Connect an ohm meter to the top two terminals (2 & 18) ("Min" side of switch)

d. With the engine and ignition off, rotate the switch CCW until it hits the stop. Slowly rotate the switch CW until you see the switch “just” trips (ohms go to zero, showing a closed switch). Snug down the top machine screw. You can also listen and feel the “click” when the switch trips, which is what you have to do if you don’t have an ohm meter.

e. Open and close the throttle using a thin blade screwdriver to "just" crack the throttle slightly. Verify when you crack the throttle, the switch opens (off) and when the throttle valve goes back to the idle stop, the switch closes (on).

f. If satisfied, snug down the bottom machine screw and reconnect the connector. Start engine and confirm normal idle.


6. Motronic/ECU Coolant Temp Sender.

The coolant temp senders are know to be pretty reliable, however, the wires to the sender connector were cut slightly too short from the factory and it was the subject of a technical change order. This can cause a wire break at or near the Bosch connector. The connector is blue. Don't confuse it with the dashboard gauge sending unit that is next to it. Check for the integrity of the wires as well as that the connector. Ensure the sensor connections are clean and corrosion free. If the wires are good, then check the resistance of the sensor according to the table shown.

If the coolant sensor is bad (out of range or implausible for the conditions), then the ECU will use a default value for cold weather starts, then switch to a default value of about 180F for normal running. This is the "limp home" mode.


End

 

Nice article, John.

Incidentally, I've had a couple of idle stabilizers fail intermittently on the Spider, and when they start to go you can get very random CEL errors. When my last one started acting funny I was getting codes for the purge solenoid!

 

Addition made.

 

I've only had a quick skim (there's a *lot* of information there to digest!) but a couple of points:

The fuel mixture is not static at closed throttle - it's varies in relation to RPM and engine temp, as well as an initial enrichment factor after starting.
If the revs drop too far at idle, the fuel mixture will go quite rich as a form of last-ditch stall saver.

Coolant sensor - if the coolant sensor signal is out of range or implausible for the conditions then the ECU will use the MAT value for cold weather starts, then switch to a default value of ~80*C.
This does however trigger a form of "limp home" operation, wheres your suggestion of substituting a resistor wouldn't.
You might find cold weather starts a bit tricky though, because the cold start enrichment won't be activated.

Decel fuel cut/restore points are determined by engine speed, engine temp and engine load. It's actually a fairly complex process for what you'd think was a simple function

Another option for checking the TPS adjustment is to hook a laptop up to the ECU and it can tell you when the contacts are open/closed as you move the throttle, as well as show you all the sensor readings, engine parameters etc. You can also read and clear fault codes this way.
In the pic below, the ECU could see that the TPS idle contacts were currently closed.

 

Thanks. Good info. I incorporated it into the verbiage.

 

This should be a sticky, or added to Steve's Maintenance Tips.

 

I can copy this thread link and add it to my sticky.

 

Also keep in mind that there are at least two different coolant temperature sensors, with slightly different curves. I discovered that my 91S has one which is slightly "leaner" than that called for in the S parts listing, but since the car runs great, I haven't ever changed it. It might be for the Milano, I don't remember what I found out.

 

Both the parts book and IAP list only one type of temperature sender for all 164 models. According to IAP, the sensor is different for the Milano (both 2.5 and 3.0L engines).

 

That's what I figured as well. I don't know why a Milano sensor was installed by the original owner (an Alfa rep in Florida as far as I could tell). but, since it has always started well, run well, with decent fuel mileage, whatever, I have just left it alone.

Tells me you could use either without problems if you had to.

 

You will find both the cream/creme colored 023 and the blue/bleu 026 sensors in the 164.

They are very nearly the same spec.

 

"They are very nearly the same spec"

Yes, I plotted out the real measured curves for mine, the (Milano) one in the S and the 164 one in the LS, and they were a little different, matching the published curves.

 

Roadtrip

Thank you so much for this post, I was having trouble with my idle stab and with your help on this post I was able to fix(I hope, so far so good)the problem. I have had my 1991 164L for 3 years now. I will refer back to this form now that I have found it. Where I now live, no one has even heard of an Alfa before. Thanks again!

 

High or 'dancing' throttle isn't always as you expect.

I once had this problem at idle that my 164 Q4 revved to high rpm's without me actually touching the gaspedal,
also when driving sometimes I could let go of the gaspedal and the car kept driving, and I have no cruisecontrol installed!

I searched for false air leakage, checked the temp sensor and the throttle sensor... everyting was OK.
Than I checked the mechanical gascable and construction itself and I saw a piece of the inner insulation, some transparant
insulation that covered the steel cable inside, has broken off and I immediately suspected that it
sometimes became stuck in the construction (that half moon thingie) and caused the throttle to hang on that point.

After cutting away that piece of insulation I never had this problem again!

 

This is such a valuable thread.

John's (Roadtrip) write up on the whole idle system has brought me back from the brink of Alfa suicide on several occasions.
It really is just a matter of stepping through the various potential causes of the bad idle.
My cars purr thanks in no small part to this thread.
For the most part the remedy has cost me nothing....just a case of bad seals and bad connections.

Ta,

Neville.

 

Just found the source of my 164 surging/unstable idle issues. Not an air leak, MAF meter or idle control valve failure.
It is actually a broken lambda sensor.

Had the cat changed a couple of years back and they said they needed to torch the sensor out of the old one - seems to have damaged the body. It is screwed in tight enough but grabbing the protruding body, while under the car, was able to wobble it. Not good. Temporarily patched it with self-hardening exhaust paste, which seems to help. Time to source a new sensor but at least the problem has been found. Progress...

 

Thank you Roadtrip for posting these detailed instructions for repairing the idle stabalizer!!! After following them precisely, I was able to get my 164 running after being disabled for over sixth months!! The rubber grommet was to blame all along.

Sometime after pressure washing my engine bay, my 164L with 71k began acting up. It would stall several minutes after start and would not restart until it cooled down. The idle was fine when cold, but began to worsen as the temperature rose to about 150 degrees. The car could be kept running with my foot on the accelerator, but once removed, the car would immediately stall. After establishing the rubber grommet from the idle stabalizer was too loose, I proceeded to apply gasket maker around the circumference. It worked! Thank you so much, I would have never guessed a leak so minor could disable a car.