Electrical Fallout

Many years ago, my dad pulled my sister's '77 Caprice into his shop and replaced the front brake pads. When she picked up the car, the radio wouldn't work. She had enough savvy to find the fuse and replace it, but it blew right away. Her savvy ended there; she plugged in three more fuses before the guy at the parts house advised her to have the car checked for a short. She and my brother both figured my dad had done something to cause the problem. After all, the radio was working when he pulled the car into his shop. To make peace before things turned nasty, I decided to take a look at the problem.

I pulled the radio out and replaced the fuse; it didn't blow with the harness disconnected. Reconnecting the wires to the radio and touching the metal part of the dash with the unit, I heard the fuse snap. I found a shorted noise capacitor in the radio, replaced it and restored the music.

It should be obvious, even to a reasonably intelligent but technically challenged person, that my dad didn't short that big capacitor out while he was replacing the front brakes, but simple customer logic dictates that he must have, because it was working before he did the job.

Having encountered many such situations in my professional career as a spark-chasing wrench man, I was careful to approach the multiple failure on this Taurus with discretion. I didn't want to sling mud at anybody; I just wanted to surgically repair the concern. The first thing I did was search for common links between the failed components and systems.

Powers and grounds

The Taurus didn't have any running problems, just a list of inoperative systems - namely, wipers, windows, Variable Assist Power Steering and fuel gauge. Drawing on this list, I opened the wiring schematic and began my search. Incidentally, some late '90s Tauruses have a tendency to backfeed a renegade 12-volt signal from the Transmission Range (TR) sensor and fool the General Electronic Module (GEM) into thinking the ignition switch was in the start position. This produces a weird set of circumstances that causes the power windows and radio to shut down when the gear selector is moved from one gear to another. If disconnecting the TR sensor restores radio and window function, simply replace the TR sensor. But this car had no radio problem.

Looking for common links, I found that on the 1998 Taurus sedan, the wiper motor ground (G104) is on the left front inner fender near the air cleaner. The power window switch ground (G201) is behind the left kick panel. G101 grounds the fuel gauge and the tank sending unit, and it is near the battery on the left fender. And while G104 and G101 are only a few inches from each other, they checked out just fine.

One thing I noticed as I let my fingers do the walking through the wire map is that G101 and G201 do have one important thing in common: Both feed the GEM, which is responsible for wiper operation (inoperative); Variable Assist Power Steering (also inoperative, defaulted to high effort); and the operation of the Accessory Delay relay that feeds the power window circuit, which was the first problem the owner noticed after leaving the filling station.

Searching the schematic for common power feeds produced no further results in that direction, but the fact that the inoperative systems - except the fuel gauge - were tied to the GEM intrigued me. Finding the Accessory Delay Relay (top right corner of the interior fuse panel), I snatched it and used a jumper wire to bypass it. The power windows came alive, but I still didn't have driver window "One Touch Down" function, and the wipers were still as dead as a hammer. The GEM was rapidly approaching the top of my suspect list.

Scanning for data

My scan tool of choice for most late-model Fords is the New Generation Star (NGS) tester. This old T-bone has been out there since the late '80s, and Ford had plans to scrap it when the laptop-based Worldwide Diagnostic System (WDS) came online. But there was such a hue and cry from the field that Ford decided to continue updating the NGS and even provided a card-based interface with the WDS for reprogramming functions. A guy who is accustomed to the NGS will always grab it first when it's time to scan a Ford.

These cost between $2,000 and $3000, depending on whether or not you buy the reflash kit, but it's well worth the money. The NGS has nice generic OBDII capabilities and even contains a banana plug-compatible signal generator output for test purposes.

Plugging the NGS into the Data Link Connector (DLC), I attempted to talk to the GEM module, but to no avail. Running the car through the NGS data link diagnostics test - it sweeps all the busses leading to the DLC - I found that the Powertrain Control Module (PCM) would communicate just fine, but the GEM communications link (a different wire) was reading "No response/Not equipped."

Resisting the urge to throw a black box (the GEM in this case) at the problem, I decided to gather more information. I've been burned too many times by assuming a solid-state black box was bad only to find that there was a loose connection somewhere else.

A short course in GEM-olgy

The GEM is an interesting little creature Ford introduced on 1995 Windstars, Explorers and Rangers. The base model Rangers came equipped with a watered down version of the GEM called a "Central Timer Module" (CTM). On these three vehicles, the GEM is stand-alone. The early Windstar module is above the driver's right shin, and on the Ranger/Explorer, if you remove the radio and its trim panel, you'll see it in there on the left. The GEM/CTM has its wiry little fingers in a lot of pies, and it handles anything that needs to be timed, as well as many things that don't. The wipers get all their marching orders from the GEM; the stalk switch on the steering column merely interfaces with the GEM, which produces wiper operation by way of a pair or a trio of relays in the Power Distribution Center under the hood, depending on the application. The GEM also handles all the chime functions, the One Touch Down driver-side power window, and some of the instrument cluster warning lamps.

Another GEM-driven responsibility is to time-out the Battery Saver relay, which shuts down battery-killing interior lights 45 minutes after the vehicle is parked and, on the '98-and-later platforms, when vehicle speed surpasses 5 mph.

The GEM also handles the Accessory Delay function, which allows the driver to listen to the radio and operate the power windows for a given amount of time after shut-down or until a door is opened, whichever comes first. The GEM on some vehicles will even sound a warning chime if the vehicle is driven a mile with the turn signal engaged.

The GEM is not really a body computer per se, but it's close. Taurus received its own GEM in 1996; the F-Series pickup and the Expedition got theirs in 1997, both riding piggy back on the interior fuse panel/junction box.

Crown Vics and Lincolns had a "Lighting Control Module" (LCM), which is a different animal.

Isolating the problem

Printing the connector pinouts and the schematic for the three accessible GEM module connectors, I found that all the fuses feeding the GEM were good and that powers and grounds were present everywhere they were supposed to be. Furthermore, as far as the communication problem goes, the blue/white DLC bus wire had good continuity all the way from the GEM connector to the DLC unit. The remaining inaccessible GEM module connector is a direct plug-in to the rear of the interior fuse panel/ junction box. I have seen numerous Expeditions and F-series pickups with junction box problems, but never a Taurus. As I considered the big picture, it became clearer: The power windows wouldn't work because the GEM couldn't energize the Accessory Delay Relay, which is always energized when the key is on. The wipers wouldn't work because they depended on the GEM to fire up their relays.

The Data Link communication problem appeared to be yet another symptom of a GEM module that had either died of natural causes or had been murdered. However, there was yet another mystery: The fuel gauge was totally separate and apart from the GEM module, and it was on the blink as well.

Concerns pinpointed

The fuel gauge was reading past full - the way these gauges do when the sender is open or unplugged. I always like to go to the tank connector (front right side of the tank) and ground the yellow/white wire to see if the gauge will drop to empty. If it does, the sender is at fault.

In this case, grounding the sender wire did nothing. Reconnecting the tank sender and moving to the point where the yellow/white sender wire passes through connector C214 en route to the instrument cluster, I checked the wire for continuity through the sender. A high impedance test light works well for this. I found a gentle path to ground through the tank sender.

Removing the cluster revealed a scorched transistor on the slosh module circuit board. This little signal filter keeps the fuel gauge from reacting to sloshing gasoline, but in this case, it had opened, causing the gauge to read as if the sender were unplugged. The slosh module for this car is available only with a new gauge at a cost of $40. A new GEM brought the wipers and windows to life, as well as normalizing the power steering, and the DLC had good communication.

In retrospect, it appears that somebody connected a pair of jumper cables or the new battery backwards and fried these two solid-state parts.

Because we don't charge labor at the college, the bill for the repairs was slightly less than $200. It could have been a lot worse.