Hunting engine speed at no load
 

I have a high miles '88 3.0L 5-speed (one owner) that still has excellent oil pressure and gas mileage, but it has recently lost its smooth idle and easy starting character.

The problem started gradually with multiple tries at cold start. It has a fresh battery, but I've been letting it sit a lot, sometimes for a month at a time. I replaced the passenger side valve cover gasket in early summer to fix a bad oil leak. I didn't start it for a week afterward and it was surprisingly difficult when I did, but it ran OK afterward.

Now that the weather has cooled off (I'm in snow country) it "idles" like it has a huge cam. I took the air bypass valve off and thoroughly cleaned it with carb cleaner and a Q-tip. I tested the valve solenoid with the battery and reinstalled it--no change. The rpm hunts even at higher rpm, sometimes rhythmically going up and down. It doesn't do this under load. I wedged a vacuum tee (it was the right size and close by) in the throttle linkage so that I could keep a steady throttle while I pulled the air valve electrical connection (to defeat and close the air valve). The rpm decreased, but continued to hunt.

I replaced the throttle position sensor with an old used one that I had lying around--no change. (swapped it back)

I looked for vacuum leaks, greased a couple of them, and put an O-ring tourniquet (like a hose clamp) around the cab heat control line at the manifold tree. Squeezing and wiggling various vacuum line connections: brake booster, pressure sensor, fuel regulator, etc. has no effect on the engine rpm.

Pulling the PCV has no effect and has good suction.

The fuel pressure is a rock steady 42 psi no matter how I jerk the throttle. It is less than 10 at start, but rises above 42 at hot start. I have a turbine style pump, not the positive displacement type it originally had. (Those seem to beat themselves to death.)

At speed the van runs quite smooth, but maybe once or twice per trip on the freeway I will feel a brief "headwind". I don't fight it and anyone behind me wouldn't notice any slowing down. It comes back right away, but no backfire.

While checking the vacuum tree I got blasted by the distributor and replaced a bad plug wire. Now I can put my bare hands all over the distributor with no sensation. (I've had a coil go bad on this vehicle years ago, but that would cause a shut down while cruising and not start again until the coil cooled down.) I checked the plug connected to the bad wire and it was clean, gap at .047.

Removing the air cleaner does nothing.

I disconnected the injectors on the passenger side, one at a time and it would very slightly stumble, but recover. Reconnecting each injector seemed to have little or no effect. They must be upstream far enough to share? The driver side are inaccessible with the engine running. I dumped a pint of Gunk injector cleaner in an almost empty tank and added a 1/2 gallon of fresh gas, took it on the freeway, then let it sit overnight. I started it again in the morning: ran like crap. Added 1.5 gallons of fresh gas--no change.

We are in the 40's and 50's now. I wish I could retest it in 70 degree weather.

Check the vacuum line going to the MAP sensor.
BTW how difficult was that valve cover job? Mines been leaking for quite some time. I keep putting it off. It looks like a daunting task.

When I said, "pressure sensor", I was referring to the MAP. I even beat on it with a screwdriver handle. The gasket job isn't too bad. The driver side is easier. You might need penetrating oil to remove the oil filler bracket.

My memory might be amiss about the coil. It could have been the ignition module that was causing the shutdown more than a decade ago.

I want to clarify the way the rpm varies. At about 1500 rpm the speed variation is more swirling and sinusoidal than pulsating. It's not an on/off pulse train. I'm guessing about one cycle per second. When it tries to do this at normal idle speed it dives so low that it stumbles so it behaves more randomly.

Looking at the cross section of the bypass valve, I'm wondering if a weak return spring could be overcome by manifold vacuum. The valve is opened by the solenoid, closed by the spring. Since the valve lifts into the vacuum it should open more easily than close. A cycle could develop when high vacuum opens the valve which kills the vacuum, allowing the spring to close, repeat. Seems like poor design. I would have made the solenoid pull the valve open instead of push.

My next trick will be to remove the valve again and block the holes with a plate or rubber corks. It won't start or run this way without something to crack the throttle.

I think it is important diagnostically that the problem goes away under load. What happens under load? Higher intake pressure (weaker vacuum). This seems to rule out things not affected by the intake like a bad PIP sensor or a bad TPS,

Another intriguing possibility is the steering pressure switch. I've been wrestling with air in the steering. It collects enough bubbles to sometimes belch out the cap vent. Another poor design feature is the long pressure line that arcs up well above the pump, making it difficult to bleed. Anyway, I noticed this morning that the engine responded to my opening the steering pump cap. I thought it was coincidence, but maybe the pressure switch is acting goofy. The pressure switch is supposed to control idle (or add its 2 cents anyway!). That line was loose so I snugged it up with zip ties not long ago. Maybe I worked the electrical connector loose?

I think something is messing with your fuel/air mixture (or a sensor is giving bad input) and the "computer" keeps trying to compensate for it. A vacuum leak can do this. By any chance do you have a vacuum gauge?

I have a vacuum guage (at least one) and I see the unused port on the tree.

How about the manifold temp sensor?

update
 

I just checked the steering pressure sensor and it is not in the flexible part of the pressure line. I disconnected it anyway--no effect.

I made a nifty little aluminum plate with the same footprint as the bypass valve. With the valve removed and the plate in its place and my foot on the gas pedal, I was able to fire it up. The rpm was more stable and I was able to get the speed way down without it killing, but the vehicle was still shaking--running rough.

I've read that the 4.0 liter intake bolts come loose, causing rough idle and coolant leak. I might check those even though this is a 3.0.

The same post also put clogged fuel filter at the top of the list, but would a steady fuel pressure reading rule that out?

Do you mean the one just behind the idle air valve? O'Reilly has that for 14 dollars. The book says you can't test that without "specialized equipment", however an old one I have reads 33.7 ohms at 70 degrees.
There's also a coolant temp sensor that's separate from the one for the gauge, but I can't remember where it's located. It's not very expensive either.

update
 

Even with the valve removed I got the high rpm cycle this morning. I got it to idle real low (and rocky) so I could check again with ether for vacuum leaks. I sprayed under the throttle body for loose manifold leak. Sprayed the tree, the map, the brake booster, fuel regulator, cab control lines, etc. I suppose the booster could leak internally, but the brakes work fine.
This thing does not have an EGR system, but it has an O2 sensor. I had a GM 3.8 of the same model year lose an O2 sensor. All it did was throw a code. No effect on drivability, mpg or anything. (maybe emmisions, I don't know) I trust the OBD would tell me if the O2 was bad. This thing isn't throwing any codes.
Bad fuel can cause rough idle, but can it cause rhythmic cycling at high no load rpm?

I should study if the cycling is limited to a narrow rpm range.

Does you A/C work? When you turn on the A/C (or vent if you don't have A/C) does the air blow out the dash vents or the defrost vents?

Cab controls work fine.

I'm letting the thing cool right now so I can check a plug wire that looks like it's lying on the exhaust. Man, that would be an easy fix, but I think an intake manifold leak is more likely.

I put the van up in the air so I could look underneath for exhaust leaks when I saw moisture around the drivers side head gasket at the rear. I thought, "an exhaust leak under the head?" Then I followed the wetness back up to the distributor, intake manifold area. I haven't been able to keep the radiator full for many years, but I always thought it was because of the cracked overflow reservoir. I've removed the throttle body and injectors years ago to clean the injectors, but I didn't take the lower intake off. Not looking forward to it!

(BTW: The exhaust was tight--no leaks.)

I would replace the Coolant Temp sensor, not the Temp Sender, they are two different parts.

Also hunting idle can be caused by an air leak, not vacuum, AIR LEAK. check the air intake tube from the air filter box to the intake.

Thanks for the most thoughtful reply yet. The problem seems to be both vacuum and temperature related. If the sensor is not working the computer can't know when the engine is cold. (Meanwhile I have already tested the intake temp sensor. The resistance varies smoothly as I exhale over it.)

The air leak you suggest does not apply. My favorite feature of this injection control system is the absence of a direct air mass sensor--no flow meters, hot wires, etc. that seem to cause headaches with other systems. Air mass is derived from throttle position, pressure, and temp. Simple and rugged. I love it!

I mentioned a possible coolant leak at the rear. I was wrong. There is no water jacket up that high. It looks more like oil splashed up from the valley is weeping out.

was the idle hunting present before you replaced the valve cover gasket?

not neccesarily related but this idle hunting also happens in the Jaguar 4.2 liter fuel injected engine and it is always caused by the Coolant Temp SENSOR, (not the Temp Sender), when it fails.

another cause of idle hunting in fuel injected Jaguars is the grounding of the O2 Sensor. As the exhaust system ages and rusts, Ground to the O2 Sensor is lost. The solution from Jaguar techs is to add a separate ground strap to the O2 sensor. In my car it stabilized the rough idling completely. I was shocked by the change in englne idling behavior when I installed a ground strap between the base of the O2 sensor and the manifold.

Gotta' love the old Jags. Two gas tanks means double the troubles.

Any locksmiths out there? I need to convert a Jag key code to the dimensions. I even have a dimensional drawing of the key, but no way to convert the code. The code was on a sticker inside the lock/switch.