Engine warm idle issue

[NigelStu]

Recently bought my 'new' E30, an '87 es and I've come across my first head-scratcher problem. Searching has not been my friend on this. Seems like most people who had hunting / fluctuating idle issues were generally all the time, and lots of good info/answers for those (vac leak, TPS, ICV, etc.), but it isn't helping me.

I'm having an issue specific to when the engine is up to operating temp. Start-up seems to be OK, and initial idle is OK, maybe a bit high, but its even within +/-50RPM. Once the temp gauge gets up to near the half mark, the high idle and fluctuation starts. Idles about 1000 and fluctuates to 1500-2500.


This all started after I did some maintenance / clean-up (ran OK before). Cleaned out ICV, throttle body, injectors and AFM. Adjusted valves. New gasket for valve cover (proper tightening sequence per bentley) and throttle body.


-I've checked vacuum leaks, and replaced a few hoses/vac lines based on what I found. Blocked off the evap canister at the throttle body. (although based on issue only at operating temp, I'm guessing not a vacuum leak issue)

-I ran through the electrical tests.

• ICV hums at idle, ICV closes with 12V applied.
• ICM numbers all looked OK.
• TPS is in spec at idle and full throttle with correct throttle openings.
• Temp sensor is in spec at operating temp.
• Thermo time switch seems to be in spec at operating temp.
• Cold start solenoid doesn't have impact
• No change if O2 sensor is unplugged (update - O2 voltage and heater passed bentley test)

I'm at a loss who's got some ideas of what to check next? Or re-check?

And if anyone in the Detroit area has a known-good ICV/ICM that I can try, lets meet up this weekend.


TIA

Ben

[jlevie]

Unless you've had a smoke test run there may still be intake leaks. I'd do that before going any further. And if the O2 sensor is at or past it's useful life (100k), I'd replace it.

[NigelStu]

Can you explain how a vacuum leak would cause an idle issue ONLY once the engine is up to operating temperature? This is the part that I'm not understanding; if there is a leak, I would think it would impact the system at all times, independent of coolant temp.

[jlevie]

A significant intake leak will casue noticable problems on a cold start. But a small intake leak may only be noticable once the engine switches to closed loop operation.

[NigelStu]

Can you explain this a bit - add some detail?

What all is changing between open loop and closed loop? I'm assuming that the switch from open to closed is based on the thermo time switch, Is there anything else that the computer uses to determine open/closed loop condition?

How does the system work that in open loop a tiny leak has no effect but then in closed loop it shows up? What is the system reading (or ignoring) in open loop that would basically be ignoring (or compensating for) the extra air that would be coming into the system via a small leak?

I'm trying to learn so I can pin-point the issue and fix it. Whether it be a vac leak or some electrical gremlin.


I'd much prefer ideas and tests I can do before spending any more $...

I did conduct a home 'smoke test' to test last weekend for vac leaks, I pressurized the intake manifold via FPR line and sprayed glass cleaner on all accessible joints, concentrating on all parts that I changed/unbolted (vavle cover, ICV hoses, valve cover breather hose, intake boot, EVAP line, throttle body gaskets, brake booster). Bubbles were found and leaks fixed with new hoses/clamps.



BTW - add in o2 sensor elec test over luinch today. Both O2 voltage and heater circuit passed per Bentley, so it also appears to be OK.

[jlevie]

When in open loop mode the DME is only using data from the AFM, ECT, and CPS. Open loop maps are intentionally richer than stoich because a cold engine needs a richer mixture. The cold start valve found on an ETA engine provides even more enrichment when starting the engine. The time-temperature switch controls whether the cold start valve will open and when it closes.

The DME shifts to closed loop mode when data from the ECT idicates that the engine is warm enough (around 120F i think). Once in closed loop the DME uses data from the AFM, ECT, O2 sensor, and CPS. The maps change to those designed to produce stoich operation and the DME adjusts fuel trim based on data from the O2 sensor in an attempt to run at stoich conditions.

An ETA has a separate ICM not integrated with the DME. While a similar affect can occur on an M20B25 (or SuperETA), a small intake leak that is leaning out the mixture can cause hunting as the DME is trying to achieve stoic idle conditions. And a small leak will raise idle speed causing the control system to attempt to bring it back down, which changes the mixture because the leak has a greater affect with a reduction of the air flowing through the ICV.

Other causes of what the engine is doing are possible, but until it can be proved that there are no intake leaks you are chasing your tail. Once I knew for certain that no intake leaks are present I'd next want to make sure that the cold start valve isn't leaking, verify correct adjustment of the throttle & TPS, and then swap in a known good ICM and ICV.

The O2 sensor is a scheduled maintenace item that needs to be replaced every 100k. The tests in the Bentley can't tell you if the sensor is good (for that you need a mixed gas flow bench), only if it has failed.

[NigelStu]

Thanks Jim - that description makes a lot of sense out of this.

I am looking at potential vac leaks. Ruled out EVAP canister and brake booster by plugging the inlets to throttle body/manifold. Also unplugged the cold start valve and blocked the fuel line to it to take out that aspect. Using various methods, I'm still not finding any leaks (carb cleaner and propane at any joints with engine running and pressurizing intake manifold and spraying windex and looking for bubbles - anyone have a suggestion for smoke testing near Novi MI?)

Did some more poking around with the car. The temp switch is apparently bad, with continuity at low temps (should be none) and high temps. But that leads to a questions. What tells the car to switch between open loop and closed loop? I thought it would be the temp switch, but unplugging that switch and then jumpering the wires does not seem to do anything (forcing 'hot' temp). What else feeds in to switch from open to closed loop?

[jlevie]

The DME shifts from open to closed loop based on data from the ECT (Engine Coolant Temperature) sensor. That sensor is (normally) in the top of the thermostat housing, just under the sensor for the temperature gage. The time/temperature switch is only used to control the cold start valve.

You really need to have a smoke test run. That is the only sure way of locating intake leaks. Call around to shops in your area to see if someone has a smoke machine.

[NigelStu]

Did some more digging on this, assuming there are no actual intake leaks (I'm still not finding any), does this point to a potentially fubared ICV?

I tried adjusting the ICV with the engine surging. Turning the screw on the ICV would change the range, RPMs and speed/frequency of the surge. With the adjuster screwed in almost all the way, surge took about 2 seconds to go from 1000 to 1300. With it screwed most of the way out, it took about 1 second to go from 1100 to 1800. Going from all the way out to all the way in, the idle would drop and stabilize around 900 until any throttle is applied; then it went back to surging idle.



Secondary question - could an intake or exhaust valve that is not seating properly create a vac leak somehow?