Idle FAQ
by Sam Chien-shin Lin.

Originally posted at

Last Update: 29 August 1995

-added fuel pressure regulator, cold-start valve, and 1-wire O2 sensor


I decided to put this FAQ together in the hope that others won't have to suffer the many hours of frustration and hopelessness I went through just to get the car to idle! The info contained herein was obtained from many sources including what I learned from my toils, Bentley, and snippets sent to me by other Bimmernetters (sorry I can't cite sources - there were too many). If you have any additions or corrections, please send them to me. DISCLAIMER: my own experiences were based on my '85 325e and '84 528e, and a friend of mine's '86 325es, so any references to other models are based on info I couldn't verify myself. Many thanks to all who provided information. Hope y'all find this helpful...

Sorry I couldn't cover other years/models, but as it is, writing this up took several hours as it is. The info contained herein applies to models w/ Motronic units prior to 1.1 (which eliminates the ICM and controls the ICV directly). Models covered include 1982-87 5 and 3 series.

After writing this, I found out that a similar FAQ had been added the WWW server. I've appended it after my FAQ.


TechFAQ: The Idle Stabilization System

1. Idle System Overview

In a conventional idle system, idle speed is controlled setting a baseline throttle opening. The idle speed is increased by increasing the throttle rest opening, and decreased by decreasing the throttle rest opening, usually via a throttle-positioning set screw.

With the idle stabilization system, the throttle is completely closed at idle; instead, air gets into the engine via an electronically controlled bypass system. Air for the idle system is obtained via a hose tap in the intake boot upstream from the throttle, and fed into the engine through a manifold tap next to the cold start valve. The idle speed is controlled by modulating the amount of air which bypasses the throttle via an electronically controlled closed loop stabilization circuit.

air filter --> airflow meter -----> throttle ------ manifold --> engine
                               |                        ^          |
                               |----------> ICV --------|          |
                                             ^                     |
                                             |                     |
                                            ICM <- input signals --|

Engine input signals and feedback allow more accurate control of idle speed over changing ambient air pressure, temperature, etc. (and hopefully, reduced emissions).

---> Note that the idle system gets its air _downstream_ from the air flow meter! The air is _metered_ so the DME knows about it!

The heart of the system is a small brain box, called the idle control module (ICM) (or idle control unit (ICU)), which takes engine signals as inputs. Although the ICM works in concert with, and shares some inputs w/ the Motronic (DME), there is no direct electrical communication between the two systems. The sole output of the ICM is a control signal which modulates a solenoid operated valve (the Idle Control Valve (ICV)). The ICV in turn adjusts the flow of bypass air through the idle system. The DME responds to the idle system via the air flow meter and its effects on input signals, such as engine RPM.

From the above description, a common fallacy about the idle system is instantly dispelled:

Fallacy #1: An idle system malfunction can mess up the mixture and destroy my oxygen sensor or catalytic converter, costing me mucho dinero.

Repudiation: Fallacy #1 arises from the notion that somehow an idle system malfunction can result in richening of the mixture, thus resulting in destruction of the O2 sensor and cat. However, _all_ air through the idle system is _metered_ by the air flow meter. Also, there are _no_ electrical outputs passed from the ICM to the DME. Therefore, the idle control system can only affect the idle mixture to the extent that changing the idle speed can cause the DME to vary the mixture.
Salient Points: Your mechanic is full of sh--. Experimentation with the idle system won't destroy your car in some unknown way. The only way you can directly change the idle mixture via the idle system is to introduce a vacuum leak (which results in _leaning_ of the mixture). The only thing that can cause a rich mixture is a malfunctioning DME.

The ICV control signal appears to be PWM (pulse-width modulation), meaning that the ICM varies the duty cycle instead of the voltage to change the valve opening.

2. ICV Overview

The ICV is a dark plastic or silver-colored metal cylinder 3.5" long and 1.5" in diameter mounted on top of a two-legged support bolted to the valve cover near the firewall. An intake hose gets air from a tap on the side of the intake boot between the air filter and the throttle (the intake fitting is directly opposite the electrical connector). The output is at a right angle to the intake, and feeds to a short 2" hose, which in turn feeds into the intake manifold next to the cold start valve. A black plastic two-pin connector feeds control current to the solenoid-operated valve flap with the ICV.

When there is no current, the valve is completely open. Increasing the current decreases the opening. The solenoid valve in the ICV does not entirely control the air flow though it; an adjustable bypass system within the ICV allows air to flow through it even when the solenoid valve is completely closed (hmm...a bypass system within a bypass system!). When the adjustment screw on the side of the ICV is turned all the way clockwise, no air bypasses the solenoid valve; turning it counterclockwise increases the bypass air flow.

----> The adjustment screw analogous to throttle position screw in a conventional idle system. Opening it up is like cracking open the throttle.

Bentley describes adjustment of the ICV screw as "adjusting the ICV current." In reality, the adjustment screw only indirectly affects the control current to the ICV because the ICM reacts to the RPM fluctuation caused by the change in idle air. When you open the screw, it lets more air bypass the valve, increasing the idle speed; the ICM responds by increasing the duty cycle (current) to decrease the ICV opening.

Bentley also says that adjusting the screw isn't supposed to affect the idle speed. I haven't found this to be the case (even when borrowing my friend's working ICM and ICV), so don't be too alarmed by it.

ICM Overview

The ICM is made by VDO. It is a 2"x2" box located above the glove box, to the left of the Motronic unit. To access the ICM, open the glove box and remove the black plastic upper cover (two phillips screws facing you at the junction of the dash and the cover; two black plastic retainers, in the back - rotate and remove). The big box w/ the large connector is the DME. There are different colors of ICM: solid black, black with a green stripe, and solid green. Black is the oldest. The solid green one it is the latest update (the one you want). The ICM is held in by a single bolt, and a 2x6 12-pin connector is connected to it. The pin numbers are clearly marked on both the ICM and the connector:

1) ICV (output)
2) Power supply
3) RPM sensor
4) Ground
5) ICV (output)
6) Coolant temperature switch
7) Automatic transmission range switch
8) N.C.
9) A/C switch
10) Air temperature switch
11) Coolant temperature sensor
12) Throttle rest position switch

All of the signals are inputs except the ICV outputs.

Inside the ICM is an analog circuit mounted on two circuit boards w/ a flexible connection. The circuit consists of an assortment of resistors, capacitors, op-amps, etc. Its job is to decrease the duty cycle of the ICV signal when RPM's dip and increase it when RPM's rise, contingent upon its various other inputs. Sounds like a pretty easy task, doesn't it? Why it doesn't do a better job is a puzzle to me. It seems like any idiot could design a better circuit. For one thing, the ICM can't seem to compensate very well for changes in ICV friction (that's why it's a no-no to clean or lube the ICV! It'll mess up the calibration and the stupid ICM won't be able to control it right!).

4. Troubleshooting

Of course, the easiest thing to do is buy a new ICM and ICV but that costs a lot of buck$ - about $300 from my sources. Getting used stuff from a yard might seem tempting at first, but bear in mind that the ICM is a delicate electronic circuit that overheats easily (that power transistor doesn't have a heat sink!), and the ICV can appear to function correctly, but still be out of whack. If you want to buy a used ICM, at least get the solid green one, because it's supposed to be the most reliable. Even if an ICV passes electrical tests, it could still be out of calibration. Personally, I think it's playing Russian roulette to buy any parts that you can't verify are good. If you're lucky enough to have a good friend who will let you borrow his/her *working* ICM/ICV, that's the best bet for testing, of course.

It's probably a good idea to make sure the car is tuned up, the air filter is clean, and the fuel filters and injectors are clean, etc. before proceeding w/ the diagnosis. If your motor is too out ofwhack, the stabilization system probably won't be able to compensate.

DON'T CLEAN THE ICV W/ GUMOUT OR WD-40! It just might change its friction enough that the ICM won't be able to control it anymore.

Vacuum Leaks

First, look for vacuum leaks! Vacuum leaks introduce unmetered air, causing a lean mixture and rough running.

One good trick is to spray a little carb cleaner where you suspect a leak. If the RPM's change, then you know you've found a leak.

Throttle Cable

If your idle is too high, make sure your throttle is really closed at the rest position. On my 325e, the previous owner had adjusted the throttle cable so that the throttle was cracked open when my foot was off the gas pedal.

To adjust the throttle, take off the intake boot. Adjust the cable until you can just barely slip a .0015" feeler gauge between the throttle plate and the throttle housing. The purpose of this tiny clearance is just to prevent the throttle plate from gouging a groove into the housing. After adjusting the throttle plate, don't forget to recheck the throttle rest position switch; you may have to readjust it. See below for a description of how to do that.

Fuel Pressure

A bad fuel pressure regulator can cause the following symptoms:

1) rough idle

2) running rich (black smoke) - this can cause black soot on your spark plugs

3) buzzing noises from the fuel pump which may vary w/ engine speed

4) general lack of power

My 528e manifested the problems only when warm. It ran fine until the temp gauge got in the the mid-range, and then would not idle the next time I got to a red light. Then it would lose power and wouldn't idle. Large clouds of black smoke and power loss ensued. Interesting thing, too, was that the car wouldn't stall as long as I left the A/C running - I guess the idle circuit's compensation for the extra load of the A/C did the job (TIP: if the car won't idle, try turning on the A/C!). At first, I thought it was a bad transfer pump, because the fuel pump would buzz loudly once the car was running badly, varying w/ the RPM's.

Lacking a fuel pressure gauge, and being the buffoon I am, I swapped out the transfer pump, main fuel pump, and cold start valve from my 325e, as well as replacing the oxygen sensor before I realized the fuel pressure regulator was the culprit. Then, reading the old digests from the list, I found a posting which mentioned the fuel pressure regulator as a possible cause (READ THOSE OLD DIGESTS!). Voila, w/ the new regulator, it runs perfectly now.

For the '85 325e and '84 528e, the regulators (last 3 digits of part number are 225) are rated at 2.5 bar; most regulators have the rating stamped on the side. I won't go into detail here about checking it, but suffice to say that if it's way off spec, your regulator is bad - you can e-mail me or consult Bentley for more details on test procedures.

The fuel pressure regulator is located at the front of the engine, and is attached to the front of the fuel rail. It's easy to identify as a brass colored metal cylinder about 2" in diameter it has a fuel hose going to it on one end, and a vacuum line on the other end which goes to the manifold. If you see other metal cans in the fuel line w/ fuel hoses coming out of both ends, those are vibration dampers - they cut down on fuel pressure variations caused by the injectors opening and closing.

The vacuum line is attached to a diaphragm in the regulator which allows it to adjust itself according to manifold pressure. To check the diaphragm, unplug the vacuum line from the regulator and plug the hose w/ your thumb. You should see a change in the pressure (or if you don't have a pressure gauge, you should at least see a change in idle speed). Alternatively, you can unplug the hose from the manifold and suck on the end of the hose. If you can't build up a vacuum, then the diaphragm is leaking and the regulator needs to be replaced.

Cold Start Valve

The symptoms described above (except for the fuel pump buzzing) could also be caused by a bad cold start valve. The cold start valve is an extra fuel injector which is mounted on the intake manifold, usually above the valve cover. There is a two-pin electrical connector going to it, as well as a fuel feed hose. The valve is supposed to inject a little extra fuel into the engine to help cold starting. Controlled by the thermo-time switch, it's supposed to shut off after a few seconds. A worn valve could constantly drip fuel, causing a rich mixture.

To test it, unbolt the two allen bolts which affix it to the manifold, plug the hole in the manifold, and have someone else start the motor while you observe the cold-start valve (leave the electrical connector attached). If the motor is cold, it should spray for a few seconds and stop. If it's hot, no fuel at all should emanate from it.

If it keeps spraying forever or drips, either the valve is stuck on or the thermo-time switch is stuck on. Disconnect the electrical connector. If it keeps spraying or dripping, the valve is bad. If it stops, then the thermo-time switch needs to be replaced.

NOTE: check the fuel feed hose to the cold-start valve for cracks while you're at it. On both of my cars, it started to drip fuel on my valve cover. Lucky I discovered it before the engine blew up!

ICV Diagnosis

NOTE: The diagnoses below don't all make sense unless you follow the steps in the sequence shown.

1) Turn the ignition key to run position, but don't start the car. You should hear quiet buzzing sound from the ICV, and if you touch it with your fingers, a vibration. If not, either the ICV is bad or there is no control current.

2) Start the car. Run the system "open loop" by pulling the electrical connector from the ICV. The RPM's should climb to about 1500-2000, and then oscillate back and forth between about 600-1500rpm. If reconnecting the electrical connector has no effect on RPM's, your ICM is probably at fault. (For the curious, your RPM's fluctuate because when the ICV is disconnected, the valve is stuck wide open, and the DME is the only thing controlling your idle. The RPM's rise until it cuts the fuel flow, which causes RPM's to dip. Then it restores fuel flow, and the cycle begins again.)

3) Cut the motor. Pull the electrical connector from the ICV and connect an ohmmeter across the terminals. The reading should be about 9-10 ohms at temp 73+-9F(23+-5C). If you get an open circuit, it's time for a new ICV. If the resistance is much lower, you've got a short, and your ICM may be roached too, from the resultant excessive current draw.

4) Disconnect the ICV hoses, and look into the outlet. Obtain jumpers and connect 12V across the ICV terminals. The valve should close tightly when voltage is applied, and open strongly when the voltage is removed. (Yes, it might look grungy and black in there, but resist the temptation to clean it w/ solvent for now - it could throw it out of whack!). If there is no movement or the movement is sluggish, your ICV is bad.

5) Plug in the ICV electrical connector and turn on the ignition (engine not running!), all accessories turned off. Looking into the outlet again, the valve should be partly closed. If the valve is wide open and there is no vibration, you aren't getting any control current. To verify, unplug the ICV connector, and verify that you're getting voltage across it. If there's no voltage, your ICM is at fault.

6) Reconnect the ICV hoses and electrical connector. Hook up an ammeter in series the ICV. W/ the engine fully warmed up and idling w/ all accessories turned off, the current should be between 400-500 mA. If the current is wrong, adjust the ICV current. Turn the adjusting screw until you get 460+-10 mA at 700+-50 rpm.

KLUDGE: If you can't get the current in the proper range, just try to adjust the screw until your idle stabilizes at 700 RPM and ignore the current reading.

If you can't adjust the control current properly, proceed to ICM Diagnosis. If the ICM checks out ok, then the ICV is probably out of whack. Maybe an ICV Kludge can help you peg the diagnosis (or fix the problem well enough for you to live with it).

ICM Diagnosis

First, check to make sure the ICM is getting the proper input signals.

Checking the ICM Inputs

Disconnect the 28-pin connector from the ICM, and perform the following measurements on the connector with the ignition on.

1,5) ICV

This is actually an output (the only one). These two pins connect directly to the ICV. Hook up an ohmmeter between pins 1 and 5. You should get 9-10 ohms, the ICM DC resistance. See ICV testing section for more details.

2) Power supply

A voltmeter hooked up between pins 2 and 4 should read battery voltage.

3) RPM sensor

Hook up an LED test light between pins 3 and 4. While cranking the starter, the light should flicker.

The resistor is connected in series with the LED, and alligator clips are connected the resistor and LED leads.

alligator clip >>--------------|>----------<< alligator
                  1-2K ohm          LED        clip

4) Ground

Use a continuity tester between pin 4 and any unpainted part of the chassis. There should be almost zero resistance.

6) Coolant temperature switch

Measure continuity between pins 6 and 4. It should be open below 86F(30C) and closed above 118F(48C). If it doesn't close, check the connection at the switch. The temperature switch is mounted on thte cylinder head coolant outlet, to the front of the thermo-time switch. It is the only sensor in that area which has two separate push-on spade terminals. The brown wire goes to ground and the white wire goes to ICM pin 6.

7) Automatic transmission range switch

Hook up a voltmeter between pins 7 and 4. With manual transmission, should get battery voltage. With auto transmission, should get battery voltage w/ gear selector on Neutral and Park positions, 0V in other positions.

8) N.C.

9) A/C switch

A voltmeter between pins 9 and 4 should read battery voltage when the A/C is turned on, zero when the A/C is off.

10) Air temperature switch

Voltmeter between pins 10 and 4 should read battery voltage below 18F(-8C) and 0V above 39F(4C).

11) Coolant temperature sensor

Hook up an ohmmeter between pins 11 and 4. Verify the resistance at the following coolant temperatures.

	|Model          |  325,325e,325es|    325i,325is|
	|connector color|     white      |      blue    |
	|temp           | resistance  (ohms)            |
	| 14 +-2F       |   7000-11600   |  8200-10500  |
	|(-10 +-1C)     |                |              |
	| 68 +-2F       |   2100-2900    |  2200-2700   |
	|(20 +-1C)      |                |              |
	| 176 +-2F      |    270-400     |   300-360    |
	|(80+-1C)       |                |              |

If you just get an open circuit, check the connection. The sensor is located on the cylinder head coolant outlet, behind the thermo-time switch. You can also measure resistance directly across its terminals.

12) Throttle rest position switch

Ignition doesn't have to be on for this one. Hook up a continuity tester between pins 12 and 4. You should get continuity when the accelerator is in the rest position, and open circuit otherwise. If not, check the switch.

At the bottom of the throttle housing, there is a 3-pin connector. First make sure it receives voltage. Pull the harness connector. With the ignition on, you should get 12V between the center and either of the outer terminals of the harness connector. Next, test the switch. The left and center terminals are for rest position. Open the throttle part way by hand. Slowly let it return to its stop. The switch should close when the throttle lever is approximately .2-.6mm from its stop.

While you're at it, even though it doesn't affect idle, you should test the full throttle switch. Move the ohmmeter probes to the center and right terminals. Open the throttle slowly. When the throttle is within 10+-2degrees of full-open, the switch should close.

If the switch is out of whack, unbolt the throttle body via the 4 retaining nuts. There are two screws on the switch body. Loosen the screws and rotate the switch body until it works as specified. If it's broke, replace it.

Testing the ICM

Although the inputs the ICM are well understood, the ICM is still somewhat of an enigma. I still haven't completely figured out the input/output relationship. If any one on the net has a good understanding, please e-mail me so I can add it in. In the meantime, here are some basic tests you can perform to see if it's working at all.

Plug in the ICM and turn on the engine.

Hook up a voltmeter across the ICV terminals while it is still connected to the ICM. Introduce a lean condition by pulling up on the oil dipstick. You should see the voltage dip slightly, then go back up as the idle speed stabilizes. If the ICM is working properly, detects the RPM drop and attempts to bring it back up by reducing the ICV current (opening it up more). If you have a scope, I guess you could check to see if the duty cycle changes instead. If you have an AC voltmeter, check AC voltage, too. Since the signal is pulsating DC, you should get an AC voltage reading. If it's just 0V, then your ICM isn't modulating and is roached.

With the engine cold, shorting the coolant temp switch should cause an RPM dip. Also, you should see 8-12V measuring from the pin 6 to ground when the switch is open.

The ICV control signal appears to be PWM waveform. Hooking up a frequency counter, I found the frequency to be fairly stable at about 145 Hz across various RPM's.

With the engine fully warmed up, unplug the throttle switch. Blip the throttle a couple of times. If the idle speed doesn't change, ignore this test. If the idle speed rises to somewhere around 1000-2000 RPM's after the blipping, plugging the throttle switch back in should bring it down again (throttle closed). This shows the ICM is responding to the throttle rest position switch.

If all else fails, open up the ICM and look for burnt parts, shorts, opens, bad solder joints, etc. An assortment of screwdrivers can be used to pry the connector out of the case. I used an X-acto knife to cut off all but two of the little retaining tangs, and then pulled the connector out by the pins. If you have a transistor or diode tester, you can test the transistors. Pay special attention to the output transistor, which connects to the ICV outputs. Electrolytic capacitors often go bad, so you can unsolder them and see if they are shorted or open circuited. If you don't have a capacitance meter, one rough method is short the capacitor to drain the charge, then hook up an ohmmeter. If it isn't roached, you should see the resistance go from very low to very high as it charges from the ohmmeter current.

5. ICV Kludging

Now for the part you've been waiting for...

If the ICM doesn't seem to be the problem, and the ICV seems to function properly except that you can't adjust it right, maybe you can kludge it. ICV Kludging will only work if your idle speed is too high or is fluctuating wildly when the adjustment screw is turned all the way in. If your car just idles too low, you're out of luck, unfortunately - go get a new ICV. First, warm the engine up to operating temperature.

If turning the ICV screw all the way in still doesn't get your idle down to where it should be, or the idle speed is still oscillating all over the place, maybe your ICV is still letting too much air into the engine. First, turn the ICV screw to approximately the middle of its travel. This way, once your kludge is in place, you can still fine-tune with the screw.

The kludge procedure is iterative. The basic idea is to add restriction to the air flow through the ICV until the car idles right. Fashion a plug by cutting a circular piece of metal to fit into the ICV intake port. Plastic is ok, too, as long as it can take the heat. I cut my plug out of a metal bottle cap.

1. Cut a tiny bit off one of the sides or drill a ~1/16" hole into it to let some air through.

2. Insert the plug into the ICV intake port and reconnect the hoses and electrical connector.

3. Start the engine and see if it idles.

4. Repeat Steps 1-3, increasing the opening in the plug until the idle speed is around 700 RPM.

Once you get it to 700 RPM, you can fine tune it with the screw. Mine stalls the first time it's started in cold weather, but otherwise, the idle speed is rock stable now.

NOTE: Using a penny w/ a 1/4" hole drilled in it as described below may get you to the right ballpark w/ less hassle.

NOTE: After kludging, your ICV current may go way off spec.

6. Other Things to Try

Some netters have said that fuel injector cleaner fixed their idling woes.

Maybe your oxygen sensor is bad. I doubt this would be the cause, but some have had success w/ changing it. Testing of three wire O2 sensor:

Locate the oxygen sensor connector. On E30's, it's clipped to the battery pan in the engine compartment. The 3 pins are labeled on the connector.

Heater: Bentley doesn't specify a resistance, but you shouldn't get an open circuit between pins 3 and 2. On the other half of the connector, you should get 12V between pins 3 and 2. Otherwise, your heater relay is probably bad.

Sensor output: The sensor generates a very small voltage (less than 1V). Warm up the car. Connect voltmeter between pin 1 and ground. Let the car idle for about 2 min. Lean the mixture by loosening the oil filler cap or pulling the oil dipstick. The voltage should drop. Try accelerating a few times or running at fast idle for a few minutes. If the signal doesn't fluctuate or there is no voltage, your O2 sensor is bad.

For 1-wire O2 sensors, just follow the instructions for "Sensor output" above.

On cars which have transfer pumps (later model cars only have 1 fuel pump in the gas tank. earlier ones have a "transfer pump" in the gas tank and the main pump is under the car in front of the left rear wheel), a bad transfer pump could be the culprit. However, bad transfer pumps usually affect off-idle performance more; fuel starvation at large throttle openings causes flat spots in acceleration or lack of power at full throttle. One symptom of a bad transfer pump is that the main fuel pump buzzes all the time. To check the transfer pump on E30, take out your rear seat and unbolt the access panel under it (oval plate on passenger side). Start the engine. While listening to the pump, unplug the electrical connector. If there's no change in tone, it either isn't getting electricity or is dead. No room to elaborate here...e-mail me or others for details.

On 5 series, the access to the transfer pump is thru the trunk. There is a round access cover under the trunk mat on the passenger side of the trunk. I have a good kludge for using a Vega pump if yours is dead - $30 vs $150 for BMW.

Try cleaning the ICV w/ WD-40 or Gumout. What do you have to lose at this point?

If all else fails: Pray to the Bimmer God.



Idle speed control

Information on BMW idle speed problems.

Pin assignments used on BMW idle speed control units.

(Used on 1982-1987 5 series and 3 Series. Not used on Motronic 1.1 cars)

  Pin     Function
-----    ----------
 1&5   Idle air stabilizer valve (9-10 Ohms between the 2 pins at 73 Deg. F)
  2    Battery
  3    Engine speed
  4    Ground
  6    Coolant temp. switch
  7    Auto. transmission range switch (12 Volts when in Neutral)
  8        "                           (12 Volts when in Park)
  9    Air Cond. on signal
 10    Air temp. switch  (12 volts if below 18 deg. F, 0 V above 39 deg F)
 11    Engine temp. (resistance to gnd. varies with temp.)
 12    Throttle switch(Continuity to gnd when throttle closed, open otherwise)


Tale of fixing a "surge" problem on a '84 528e.


Using a solenoid type control (VDO unit), feedback loop, and a control unit to stabilize the rpm to 700 rpm under various conditions.

Input to the control box:

engine speed

engine temperature

idle contact switch

system voltage

Output from the control box:

PWM (pulse width modulated) to control the stabilizer valve.

When the solenoid is disconnected, or failed, it will remain wide open. The idle rpm should go "open loop". A warm engine will rev until 1200 rpm, causing the the fuel injection system to intervene, cutting off fuel flow. Shutting off the injectors causes the engine to decrease in speed... cutting the engine out. RPM drops back, and the fuel starts flowing again. This will give a surge from 700 RPM to 1200 RPM. Important to realize the idle stabilizer is wide open, so the rpm is actually being controlled by the fuel injection system running, then cutting out. (oscillation period for this engine was about 4 seconds).

With the solenoid connected directly to 12 volts and ground, it will fully close the valve, disabling any idle function. Once the car is started, it will stall unless the throttle is depressed to keep it running. Thankfully this is a fairly simple test to check the solenoid.

With a direct connection from the battery to the solenoid, current should be around 1.2 Amps with the valve connected between +12 and ground. Should the valve be shorted out, it will have a much higher current flow, possibly damaging the meter. If this is the case, the control unit may have also been damaged. This also explains the dealerships desire to replace both units at the same time.

Make an adapter to allow the current flowing through the stabilizer valve to be measured. An analog ammeter is preferable, to see any fluctuations, but a digital one will work since the pulse train is fairly fast, and the inductance of the stabilizer valve will dampened the fluctuations.

A neat "emergency" trick is to drill a 1/4" hole into a penny, place the penny next to valve input, replace the feed hose, and leave the stabilizer's electrical connection disconnected. This will result in an idle of approximately 800 rpm, allowing you to temporarily drive without the engine surging.

Once the solenoid was checked, both under cold and hot conditions, the ammeter was left attached and placed into the passenger compartment for a test drive. Normal idle should produce a current reading of 470 ma. This was confirmed.

On the vehicle tested, the idle current would suddenly drop to a value much lower value than expected, and on occasion completely to zero. Since the valve checked out good previously, [it may have been a heat related open] the most likely cause was the control unit itself.


A small "black box" [about 2" x 3" x 1" thick] is located in the upper portion of the glove box. The much larger unit is the Motronic engine management system, the one we are looking for is usually attached to a support bracket.

The control box was marked with a green band of tape, showing it to be of the later recall that affected the '84 model year. The control valve is silver, as the earlier [recalled] units were black.

Once I had the box out, I was able to open it with care using a half dozen paper clips, a small screw drive, and lots of patience.

First glance didn't show anything unusual, only four Integrated circuits (IC's) all of which were LM2902 quad op-amps, a few active components (transistors), and maybe five dozen passive components (resistors, capacitor, and an inductor).

The output transistor, an BD438 [PNP 4Amp] , showed a fair amount of heat being dissipated, blackening the PCB (parts component board), and was a fair suspect for a "heat related failure". The heat, and vibration, of the unmounted device had caused the solder joints to degrade in appearance. Desoldering to component it was obvious the collector lead was barely making contact.

Once I had the power transistor out, I tested it over night on a curve tracer, looking for an intermittent failure. The transistor performed satisfactory, but left me perplexed as to why VDO didn't use a heatsink on it, but left it with only its leads to hold it in place and provide a heat dissipation path.

Further examination of the PCB showed a hairline crack in the trace leading from the transistor to the connector. This would close at cold temperature, then OPEN at higher temperature. Since the contact was marginal it will allow some current to pass, but not as much as expected. A repair was carried out by soldering a wire direct from the transistor to the connector.

So far everything has held together, and the car is back into operation as hoped! While a new control stabilizer valve and control unit would have set me back almost $500, I was able to track down the failure which give a higher sense of satisfaction, al a bit at over 40 hours involved! [It would not have been unusual for a BMW mechanic to have billed over 20 hours labor, and the new parts, to a tune of $1300 to repair this!]

Since a new control box, from the dealer, cost $250; wrecking yard would be approximately half that cost. But given the failure I described, the life expectancy of a used unit would be questionable.

It seems a simple microcontroller with built in diagnostics could be built to replace the original. Advantage of the aftermarket unit would be a proper heatsink, current limiting (so a bad valve can not destroy the control box), and could incorporate an LED to flash a diagnostic code allowing a quick check of the system functions. In order to develop such a box, its full cost would have to be around $20, to insure a resale price less than $120.

How many BMW's use the "green strip" or "green box" control system? Typical life expectancy of the VDO unit?


Tom Walter

Beaverton, Oregon.

P.S. for BMW owners of the similar model: BMW recommends a full tune up, vavle adjustment, fuel pressure check, before proceeding with any electronic diagnoses. Also the 528's were prone to developing carbon deposits on the intake valves, causing driveability problems [they developed a system to "walnut blast" the intake valves, with the intake manifold removed].

Hats off to someone who could FIX one of those boxes

ICU/ICV Color Coding

Heres my knowledge of the coding of said ICUs/ICVs

             Early Unit     Recall              Current
  Engine     (Recalled)   Replacement            Unit

  1.8l       blk/blk    blk(grn stripe)/blk     grn/yel
  2.8l                                          grn/red (looks purple to me!)

  3.5l                                          grn/blue

The color is Body/Connector

Early ICV's were black plastic with INTERNAL current adjust

Newer ICV's are silver with external current adjust

BEWARE .. some parts places and dealers are selling BLACK boxes with yel/red/blue ends these are NOT the current and best option ... the green (olive green) bodied boxes are purported (by CCA tech tip reps) to be the BEST as they are designed to NOT burn out (as much)