Read post #5.

Spent some time re-doing some basic checks to no avail:

• Checked CHT resistance (within range)
• Checked Idle switch for continuity
• Checked ICV is humming at power on and that it snaps open and closed (9v test)
• Checked Idle switch harness by disconnecting from idle switch (RPM's increased)
• Checked for vacuum leaks, removed several hoses and re-fitted
• Started car, removed oil cap and RPM changed, refitted oil cap and RPM dropped back (I understand this can be a good test for a pre-existing) vac leak
• Tried turning idle screw (with cable to short port), no change in RPM just surged right through my adjustments

When I turn the ignition off the car shakes slightly (presumably due to rough running?)

Time to try and find a reputable garage to check mix and reset idle (and probably look for any leaks I've missed).

Booked the car into a garage but cant get an appointment til mid-Jan. In the meantime I thought I'd do a few more tests:

* Issue is not temperature sensitive
* Changing the idle speed (winding screw in and out a long way has no effect)
* In the absence of a smoke machine, I took a rubber glove with a straw attached, attached it to the bottom hose on the ICV and tried blowing the glove up- interestingly this is not possible, the glove deflates immediately

On the basis of the checks above and given I only replaced the ICV I am wondering if I have introduced a vac leak after the ICV (hence there is unmetered air being introduced post ICV and thus changing the idle speed has no effect). Does this sound plausible? If so I may fabricate a smoke machine.

They have cheap, home made smoke machines for sale on fleabay. Work pretty well using mineral oil. Before I bought one, I blew smoke from a cigar thru a hose into the system. It located a bad intake manifold gasket for me on my BMW. Really sounds like a vacuum leak to me... easy to rule out using a smoke machine so that's where I'd go

Did you ever run the engine with the air valve hose squeezed-off, i.e. no air valve control?

Also, did you ever read this as posted in reply #5;

If it had an air leak, the air leak would have to directly bypass the throttle body butterfly, idle screw, or the idle valve,
i.e. the air leak must draw air directly thru the AFM into the intake plenum.

Sorry, missed this one - good test. Will try it and report back. Thanks

At this point the thing I see that hasn't been checked on is the intake gaskets. A couple easy checks for issues with them. (This is not an exhaustive check.)

1) - Check the allen head intake bolts to see if any are loose. They are 18ft/lbs and if any are loose the gaskets underneath are almost certainly cracked and leaking.

2) - With the engine dead cold (after sitting overnight) get an IR thermometer gun, put the car up on a lift or jack stands. Fire up the car and quickly check if the 6 exhaust ports heat up more or less equally. If some of the ports are cold there are probably leaks in the intake. (When there is a vacuum leak in one port the airflow from the leak can pull the air/fuel mixture out of one port into the others.)

If the intake gaskets are bad you will need 12 gaskets and expect to need to replace the 6 plastic spacers too. (Not sure how the plastic can erode but they can degrade heavily.)

BTW - There has been some discussion of the AFM as a possible source of this issue but when the idle micro switch is engaged the AFM's output is ignored by the DME.

you probably mean "when the WOT switch is engaged (closed) the AFM's output is ignored by the DME.

Did you overlook reading post #26, or did you not understand it?
An intake air leak at the intake gaskets will cause a lean condition and a low idle.

No.
911-606-013-00 is the idle micro switch. When you move the throttle by hand you can hear it clicking on and off on the left side of the throttle body. (Finger is pointing at the shadow where the switch is hiding.)
http://forums.pelicanparts.com/uploa...1638332792.jpg

The connector to the idle micro switch is the upper, 2 wire connector on the right side of the throttle body. You can test its correct operation there with an Ohm meter or continuity tester. (Finger is pointing at the wire that runs from the connector, back around the throttle body, and connects to the switch.)
http://forums.pelicanparts.com/uploa...1638332792.jpg

The idle micro switch tells the system that it is at idle and then the DME will use the Idle Stabilizer (commonly called the ICV) to maintain the idle at 880 RPM or 800 RPM depending on which DME chip you have.
If the idle micro switch doesn't correctly engage the DME won't know it is supposed to be at idle and won't go into its idle control loop.

__________________________________________________
An air leak at the intake, if it is small enough that the idle stabilizer can control the idle speed, will cause some cylinders to go lean and some to go rich depending which cylinder(s) has the leak and where its valve timing is lining up to the injector pulse. (BTW - I have never heard this being discussed anywhere. I discovered it during my "replace all the rubber and intake gaskets" incident.)
If the leaking gasket is on a port where the intake closes right before the injector pulse the airflow from leak will pull some of the air/fuel charge out of that runner into the plenum where it will feed other ports resulting in them receiving an overly rich charge.
If the leaking gasket is on a port where the injector pulse is immediately before the opening of the intake valve it won't really change a thing because all the intended fuel will still end up in the correct cylinder. (If it was a ridiculously large leak it would obviously start screwing things up.) This can easily be demonstrated by loosening up some intake ports and using the "dead cold engine and IR test the exhaust port temps" method that I mentioned above. Completely and consistently repeatable.
BTW - If the 3.2 had sequential injection it wouldn't act inconsistent like this.

Adding too much air will obviously create a lean condition but it will not reduce the RPM unless it is so lean that it won't light off. And the major thing you will see from this kind of vacuum leak chaos is that it will not idle smoothly at all. (But will run fine at full throttle.)

This is not the situation being described by the OP, i.e. the RPM surging from above idle to about 1700 RPM (injector shut-off occurs with the idle switch closed).

You are correct. He needs to read what he posted;

Please verify you checked the vacuum line to/at the brake booster. It’s the largest vac line in the back left side near the firewall. It’s not fun to service.
My ‘87’s hose was toast.

The idle speed adjuster on my '84 was the original slotted style for use with a flat blade screwdriver, which someone hacked up into nonadjustment.

I purchased a replacement with the 7mm hex head and swapped them out, counting the number of turns out it took to remove the original.

Since they were the same length I counted the same number of turns in, then checked to see how many more were needed to bottom out (closed). That number was 3.

I jumpered the ICV as noted above, and eventually turned down the adjustment to between 1.5 and 1.75 turns and my cold start warm-up oscillation described in the first post of this thread was no more.


--------------------------------


I would suspect that the quickest way to check for an idle air leak is to take the oil cap off. No change in idle, you have a leak; idle drops and stumbles a bit, no leak... at least that's what others have noted.

Cheers

The OP's 1st post indicated this;

That's a different mode than what you describe for your problem!

Will do some of the checks described over the weekend. I noticed today that if I take the oil cap off the RPM drops, there was no stumbling and the car didn't stall but the frequency of surging also slowed substantially.

Good, typical of an air leak that doesn't draw air thru the AFM, i.e. the idle drops.
You're making progress!

Excellent test, we are on to something.

When I took the ICV off the car I noted it was fully open.

I then fully closed the ICV, left it disconnected (car wouldn't start), I then connected power and the car started with valve fully open.

I then fully opened valve, left power off and started the car; it started and began surging.

I then manually moved ICV to a central position and started the car with power disconnected, perfect smooth idle, as the car warmed the surging gradually started to return (presumably the mix started to adjust as it warmed).

I've just completed the 9v battery test and the valve clacks open and closed, the resistance is 23ohms from inner pin to each outer and 41 ohms between the two outer pins (which I believe is broadly inline with the 20/40 Bentley suggests).

Given the valve opens fully when powering on it looks as though the valve is getting a signal to open but it seems to only move to fully open (my understanding is that it should be able to balance in the middle).

Ideas on any tests from here?