Your Chrysler vehicle runs perfectly, but the engine stalls when slowing to a stop.

The Symptom

Intermittently, the 2.2 liter injected engine in my 1986 Shelby GLHS died during coast-down when the clutch is disengaged. The starting, idle, acceleration and cruise are unaffected. Eventually, the condition was constant, and I developed a habit of roll-starting the car as I slowed to every traffic stop.

The Cause

The suspect item is called the Distance/Speed Sensor. It's an electrical item, mechanically driven by the speedometer drive mechanism in the tailshaft of the transaxle. On our two L-bodied (Omni) Shelbys, it's mounted where the mechanical speedometer cable enters the transaxle. It caused such a spectacularly mysterious set of symptoms that I documented this repair in March of 1995.

The DSS apparently contains (I'm guessing from external experiments, I've never sawed open the faulty device) 15 or 16 reed switches wired in parallel. These reed switches are arranged radially around a shaft with an eccentric magnet. As the shaft turns, the DSS produces 15 (or 16) open-close cycles per rotation.

At least one of the many reed switches sticks closed intermittently, shorting the entire sensor closed; hence, the engine control computer thinks the vehicle is at rest (at least when Distance Sensor parameters are considered by the system.)

The Distance/Speed Sensor provides road distance telemetry to the engine management computer - this data is used for cruise control and engine control. My guess is that one of the engine control applications related to road speed is maintaining high idle during coast to provide additional vacuum for power brake boost. When the vehicle groundspeed drops below a certain speed (7mph on my car,) idle drops to normal. In the journal text below, I suggest that it's emissions-related, but now that I think about it, that really could be accomplished with just manifold vacuum, oxygen sensor and throttle-position data - groundspeed wouldn't matter.

A Simple Test

My experience was with my 1986 Shelby GLHS. This is a modified Dodge GLH Turbo, and the engine is Chrysler's 2.2 liter inline four, but I suspect Chrysler uses the same part in many injected vehicles. The Shelby does NOT have a factory cruise control, but if your vehicles has one, you can perform a simple test with your factory cruise. First, make sure your cruise control works at all. If the cruise works: next time the vehicle stalls, try to use the cruise control (remember that you must exceed a certain minimum speed to engage the cruise, usually over 30-35mph.) Keep in mind that any mechanical shock or vibration could free up the stuck reed switch, so if the cruise WORKS, that's not conclusive. But if it DOESN'T work, and then DOES work later, I'd say the DSS is a good $50 gamble.

My Experience

3/7/95 10:16p

• About a week ago, I started experiencing a very specific and annoying symptom in my 1986 Shelby GLHS. At nearly every traffic stop, the engine would stall. Every other operational state was apparently unaffected. Idle, cruise, acceleration, deceleration all seemed fine. Only in the transition from closed-throttle coast-down to clutch disengagement did the symptom appear. The smallest application of throttle would keep the engine alive. If the engine did die, a brief re-engagement of the clutch was sufficient to roll-start the car, and became a habit over the past few days which I'm now having to unlearn. Once roll-started, the car idled flawlessly.
• I should mention that the problem was intermittent - a potentially frustrating diagnostic factor. Both our '86 and '87 GLHS' have multi-point fuel-injected engines, controlled by a Chrysler microcomputer engine management system. This "logic module" controls ignition timing, fuel injector delivery rate, turbocharger wastegate, electric radiator fan, and electrical charging systems. Engine management is based upon information collected from a number of sensors, including:
  • brake switch
  • distance/speed sensor
  • neutral safety switch
  • oxygen sensor
  • temperature (water) sensor
  • charge (manifold air) sensor
  • detonation sensor
  • throttle position sensor
  • distributor fuel sync pickup
  • distributor ignition reference pickup
  • MAP (Manifold Absolute Pressure) sensor (also samples barometric)
• Upon assessing these sensor states, the logic module picks the appropriate ignition timing, injection timing and turbo boost settings from a "map" of factory programmed data. If some of these sensors fail, the logic module lights a dashboard "Power Loss" light - in some of those cases, the module goes into "limp home" mode, allowing the vehicle to operate, albeit less efficiently, to get the driver to his/her destination.
• Or so the theory goes. In reality, there are many electrical connections which the logic module is not designed to monitor, or can't monitor. Should any of these fail, the engine could falter or fail to run. Moreover, there are probably some twenty vacuum hose circuits, all prone to leakage. Some of these vacuum systems affect engine performance, such as the turbocharger wastegate solenoid and the EGR valve. An electrically controlled vacuum valve is used to switch the MAP sensor from measuring pressure in the manifold to reading ambient barometric pressure - but there's no way for the vehicle to know whether the solenoid opened or not - and it may assume that a near-atmospheric reading from the manifold is a barometric reading.
• The factory Chrysler manual describes a troubleshooting process to diagnose engine control problems with the multi-point EFI system. The first step of that process is to check all the electrical connections and vacuum connections. A daunting task, at best. At worst, a nightmare of uncertainties. Especially with an intermittent problem. Often, the process of inspecting electrical connections changes their state. So the culprit connection may be overlooked in the first phase of diagnostics - and can mislead future diagnostics terminally.
• So, all of the above is running through my head, and I'm trying to figure out what should be happening during the de-clutching period, but isn't. I knew that there was some system for letting the car control throttle-closing conditions. When you lift the throttle to decelerate, there is an immediate rich-mixture condition. The fuel just leaving the jets or injectors is provided in anticipation of the air denied it by slamming the throttle butterfly shut. This unburned fuel is an emissions no-no and tends to get catalytic converters a bit frisky for their own good, unnecessarily depleting the catalyst and heating up the converter more that necessary. Most modern vehicles have a servomotor which meters additional air around the closed throttle plate for a sufficient period to burn the unused fuel. Chrysler calls this an Automatic Idle Speed motor (AIS.) Just for fun, I visited a couple of auto parts stores looking for the AIS on Saturday. I wanted to know how expensive a gamble it was, and whether it was available in the aftermarket. My first stop was a NAPA store, because they tend to carry more obscure (read: "dealer") parts, although at a premium. I was still surprised that the counterman actually found a listing: $69 and he'd have to call "the warehouse" on Monday. If they didn't have one, 8 to 10 working days. Good intel, but I'd keep going.
• Now, as I'm visiting auto parts stores, I'm also looking for this "thing." The Chrysler manuals always list "special factory tools" and their part numbers. Which is pretty aggravating, since you can't get those tools, unless you're a dealer. I've tried. The heck of it is, those factory tools exist because it's otherwise a Cast Iron Bitch (pardon the language - it's an automotive expression) to do whatever it is that those tools facilitate. So the "thing" I'm looking to find is a little box with two 8-segment LEDs on it and a slide switch and a push button. Chrysler calls it a C-4805 Diagnostic Read-Out Tool. Much of what it does, many Chryslers do already. If you turn the ignitions on our Shelbys on-off-on-off-on within 5 seconds (without starting the engine,) they'll "report" any fault codes the logic module has stored. The aforementioned "Power Loss" indicator lamp on the instrument cluster displays these codes as sequences of flashes. For instance, the sequence 2-3-5-5 indicates that the Air Charge Temperature Sensor has as some point, while the engine was running, sent a voltage to the logic module which was out of range; the 23 is the fault code, 55 is "end of message." (I have posted fault code charts for our 1986-7 Shelby GLHS' and our 1988 Caravan 3.0 liter V-6.)
• So even without a C-4805 we can benefit from the logic module's built-in fault detection. And we have. Including this problem. More on that later. But the really cool thing that the 4805 does is ATM. Not a cash machine (although that'd be a cool tool as well,) but Actuator Test Mode. Oh, this is neat. Push that one button on the C-4805 and watch the display increment from "01" to "11". Depending on which of the eleven codes you choose, the system will now:
  • 01 - Spark Activation - Once every 2 seconds
  • 02 - Injector Activation - Once every 2 seconds
  • 03 - AIS Activation - On pulse open/one pulse closed every 4 seconds
  • 04 - Radiator Fan Activation - One pulse every 2 seconds
  • 05 - A/C Cutout Relay Activation - One pulse every 2 seconds
  • 06 - ASD (Automatic Shut Down) Relay Activation - One toggle every 2 seconds
  • 07 - Purge Solenoid Activation - One toggle every 2 seconds
  • 08 - EGR Solenoid Activation - One toggle every 2 seconds
  • 09 - Wastegate Solenoid Activation - One toggle every 2 seconds
  • 10 - Barometric Read Solenoid - One toggle every 2 seconds
  • 11 - Alternator Full Field Activation - One toggle every 2 seconds
• This is great stuff. You can work on a cold, non-running engine, and (attempt to) operate devices that normally work only when the engine is running. And it'd eliminate several possible component failures from the diagnostic process. The thing is, the Diagnostic Connector on our cars is just six pins. One's ground, one' s 12VDC. The rest go to the logic module or the power module (electronic power supply.) The data's probably moving serially. I'd bet the ATM button is just jumping one of those four pins to hot or ground. Unfortunately, I might blow up a $300 logic module if I was wrong.
• I'd noticed in the past year that Sunpro (automotive diagnostic devices) was selling diagnostic boxes. I wasn't sure if they made a Chrysler box. Turns out they don't. GM and Ford, around $40 each - cheap. One of their products tests ignition modules only, including Chrysler, but it's no help. I'm still looking for this device. Or a few minutes with a C-4805 and a VOM. If you know anything about C-4805s or the pinouts of the Diagnostic Connector, please drop me a line at .
• An absolutely invaluable resource that I do have at my disposal is an identical engine and control system, parked just shy of 2 feet from my car: my wife Joni's 1987 Shelby GLHS. As long as it doesn't involve serious surgery, or risk to the healthy system, I can swap parts. One caveat: when the problem is intermittent, you might end up with a very long waiting game.
• So I swapped AIS motors, rather than gamble 70 bucks. Good thing. It wasn't the problem. In fact, I haven't swapped them back. As a matter of fact, I think the Air Charge Temp Sensors have been switched in our Shelbys since a diagnostic swap in '87. That turned out to be a faulty Shelby Automobiles splice in the harness to said Charge Sensor, which had been relocated into a custom intake manifold.
• First time out with the exchanged AIS motor, my car died a the first stop sign, a half block from home. Before this point, I'd ruled out any temperature sensors, reasoning that if the sensor failed only enough to mislead the logic module, but not enough to generate an out-of-range fault code, then at some point in its temperature range, the engine would run better. But when it had the problem, it was at all temps.
• I was still convinced that the AIS motor, although functional, was not being controlled appropriately. If it never opened, then the car might tend to be over-rich as the engine fell toward idle speed, flooding the engine at its weakest point. I'd read in the Chrysler books that AIS operation depended on the state of the Throttle Position Sensor (TPS,) Distance Sensor, Coolant Temperature Sensor and Neutral and Brake switches. I'd ruled out temp sensors (and the temp gauge seemed OK, which gets its data from the logic module.) I'd wondered about the TPS, but for some reason decided it was an unlikely culprit. Ditto the Clutch and Brake switches.
• Monday we were in my car, and Joni suggested I check the fault codes. I'd done it a couple of days before with no result, but tried it anyway. There was a new code, indicating a fault in the "vehicle speed/distance sensor circuit." A chart Joni'd gotten from the Shelby Dodge Automobile Club (we've been SDAC members since purchasing the cars) said that this fault code was "put into memory . . . If the speed sensor signal indicates less than 2mph when the vehicle is moving." I've wondered for some time how the system knows the car is moving when the Speed/Distance sensor doesn't. The chart also said that this circuit is monitored "Over a 7 second period during deceleration from highway speeds when the throttle is closed." So it apparently looks for closed throttle and higher than idle vacuum and rpm for a 7 second period, and assumes you must be moving.
• Late Monday, we drove Joni's car to a nearby Dodge dealer to buy a $50 Distance Sensor, and I paid close attention to what her healthy 1987 GLHS did as I approached a traffic stop. Experimentally, I disengaged the clutch early, at around 35mph, and coasted to a stop. The RPM would hold at 1400 until the speedo dropped below 7mph, and then dropped to a normal 1000rpm idle. Subsequent experiments with longer and shorter coasts suggested that the rpm change was speed sensitive, and not time-from-throttle.
• Even with this evidence, I chickened out of buying the sensor at the Parts Counter, deciding that I'd swap between the cars again to be sure. Upon returning home, I prepared for underbelly surgery. It turned out to be a pretty tough operation. The sensor is actually an adapter in-line between the speedo cable and the transaxle. Poor accessibility necessitated removal of turbo air plumbing (a familiar operation, at least) just to unplug the sensor from the harness. It was enough to discourage me from doing the swap, and I decided I'd go ahead and buy the part the next morning. After an hour and a half or more, I had the old sensor in hand. Optimistically, I stuck a VOM across the two pins. The Dodge parts man had suggested that the sensor was a sort of "generator," which seemed reasonable. I set the meter to check continuity, and it beeped, suggesting that there was, at least, some circuit within the mysterious shot-glass sized cylinder. With the meter set for AC volts, I spun the speedometer pinion (attached to the sensor.) Zero. DC volts. Nada. Hmm. Figuring that I might not be making enough revs, I chucked a piece of vacuum hose into the old Makita and made myself a flexible shaft drive. 1800rpm later. Zilch. Zip. 0VDC. 0VAC. 0Ohms dead short. Huh. In the past, I'd fiddled with some other spinning thing that I never figured out, so I was resigned to playing $50 auto parts roulette.
• Joni came down to the garage to see how I was doing, and I showed her the part. I'd been digging around for an old magnetic compass that's been floating around the workbench for a couple of years. I found the compass, and a quick experiment suggested that something magnetic was spinning around inside the case.
• I still wanted some concrete evidence of malfunction. But I didn't know what I was looking for. In desperation, I switched the VOM back to continuity check and picked up the part while the meter screamed away. I twisted and bent the connecting cable, hoping to find an open short. Ah! A break in continuity. Which didn't return. Fiddling with the part, I discovered that when I spun the shaft, the continuity tone was evenly interrupted, what turned out to be 15 or 16 times per revolution. Eureka! Solid evidence. Suddenly, I understood what had happened. Holding the part up to my ear (without the VOM,) I spun the shaft, and heard regular little metallic tinkling sounds. Reed switches. Like the burglar alarm sensors on doors and windows, a magnetically-operated switch. Inside this sensor, 15 or 16 of them, arranged radially about a shaft with a magnet mounted eccentrically. All of the switches are wired in parallel, so every switch makes and breaks the circuit through the sensor. However, if one of those reed switches should stick, it would dead short the entire device, regardless of the state of its peer switches. Reed switches rely on spring properties to pull against the magnet and return to their "rest" state. A quick calculation suggests that my wheels have made some 7 million revolutions to date. Depending on the gear ratio, these switches might've cycled from half a million to many millions of times.
• I bought the new part at a Dodge dealer. The Chrysler part number is 4740088. In March of 1995 it was $46.20 (tax not included) in Van Nuys, California. Interestingly, the replacement part is radically different looking from the original. Included documentation explaining how to "convert" to the new part is dated 1988, so it might have been a fix for a poorly designed original. Indeed, around 6 or 7 parts failures that have occurred on the 1986 GLHS have appeared, with nearly identical modes of failure, and usually around 6 months later (the difference in production dates) on Joni's '87 GLHS. The good thing about that (or at least, a consolation) is that I'll know how to approach the 2nd repair. If parts are inexpensive, I'll typically buy a spare, knowing it'll probably go in the Charger.
• So I put the new Distance Sensor in this afternoon, and no problems. I'm not absolutely convinced, but I'd say I'm 95% confident. Dunno why the doubt, really. Guess I just considered so many other possibilities, and can't forget them.
• (NOTE 4/29/97: To date, the problem has NEVER reappeared.)

Epilogue

This may all sound like a condemnation of computer-controlled cars. On the contrary: this car's computer actually told me what its problem was - although I was hesitant to accept it. Thanks to digital controls, my car makes 200bhp from a 2.2 liter four, gets 19-28mpg and is perpetually tuned. It's only had three sets of spark plugs (two, really - I just changed back to hotter plugs to improve driveability.) And it starts every time I turn the key. Remarkable, and for the average auto owner, a turn for the good. On the other hand, take a look under the hood of my friend's 1965 Plymouth Valiant. It's a backyard mechanic's dream. It's got maybe six things you could call a hose. Maybe less wires. My car's like a bucket of black linguine. And any of those noodles could ruin the whole meal.

All in all, both our Shelbys (we have '86 and '87 GLHS') have been pretty good cars. Mine's coming up on its 11th birthday, and just passed 90,000 miles. Considering how much time it's spent at double the design horsepower of the motor, boiling the rubber off the front tires, I'd say it's been worth it.

Ellsworth's Other Pages

• Ellsworth's other Hopefully Useful Pages
• Ellsworth's Ellsworth's Hopelessly Useless Pages

---

Future Pages

• Problems common to both our 1986 and 1987 Shelby GLHS'.
• Macintosh troubleshooting.

If you're interested in any of the above proposed pages and can't wait, drop me a line at . I'll email you the files or try to answer your questions.

Contact

Email comments to (Please: no general automotive questions - I'm not an expert, this was just a solution I deduced to one problem.)

[Back to the Top]