By Richard McCuistian

Tackling a screwball electronic ignition concern on an older vehicle can be downright enjoyable.

1984 Jeep CJ

Who knows how many miles

4.2L

Manual Trans

4 Wheel Drive

Customer washed engine.  Now engine skips and sputters above 2500 rpm.

Something Old, Something New

  From its humble beginnings as a military vehicle designed by the Bantam car company to Tomb Raider Lara Croft’s shiny new Rubicon, Jeeps have always enjoyed an intrinsic machismo, along with almost universal recognition and a longevity very few other vehicles on the planet can match, either in the real world or on the silver screen. 

I “horse traded” my cousin Billy Joe out of a rusted out ‘47 model Willys Jeep in 1983.  It hadn’t run in years and I bought some parts to rebuild the engine.  The mill was a little tight when I finally got it bolted back together, so I hooked the old bomb to the trailer ball on my dad’s 1966 Chevy pickup, found a forward gear, and dragged it across a pasture to get it started.  It was in 4 wheel drive when it fired up and it pushed the Chevy all over the place before I could get it stopped.  After that it would crank with the key and I had a lot of fun with it until my money got tight and I had to let it go.

Computer Controls, American [Motors] Style

With ever-tightening emissions standards nipping at their heels in the early eighties, American Motors pieced together a feedback fuel system for the 4.2L platform that included a microprocessor control unit (MCU) and an oxygen sensor.  This engine is outfitted with a Carter carburetor equipped with a special stepper motor for feedback response and a two-stage throttle kicker for limited idle speed control. Air management is handled via a vacuum solenoid used by the MCU to operate the innards of plastic diverter valves that are piped through a pair of hand grenade-size pulse air units to provide fresh air for upstream and downstream exhaust processing.

For me, one of the most interesting wrinkles on this system was way the AMC system handles computer-controlled ignition timing.  More about that later. 

Problem Washing

A computer programming contractor named Chris walked into my department one day last week to tell me how his 84 Jeep CJ had been running just fine until he washed the engine.  Immediately after the under hood cleaning, he found that he couldn’t raise the engine rpm above 2500 rpm without it sputtering and popping.  He had already done everything he could to dry everything out, but the problem remained.

We don’t usually work on vehicles as old as this Jeep in my department. I get requests for cheap engine overhauls on ancient pickup trucks and muscle cars, but I’m not running a hobby shop, so I turn those away.  But since we were just finishing up a course on ignition systems, and the old Jeep was blessed with an interesting ignition system problem, I told Chris to bring it in and the students would have a look at it. 

I started the Jeep when he brought it in and raised the rpm up to 2500.  He had an aftermarket tachometer, and when the engine started misfiring, the tach began bouncing toward zero, then back up to 2500.  A downward bouncing tach generally indicates a primary ignition fault, and an upward bouncing tach points to secondary ignition faults, particularly high secondary resistance concerns.  This problem almost sounded like the rev limiter strategy on a fuel injected vehicle, only not quite so rhythmic.

Knowing how easy it is to drift off track during the diag process, I kept reminding myself that the problem had manifested itself after the engine compartment was doused with high pressure spray.

Checking for Moisture

Incidentally, teaching auto mechanics in the shop/lab is a whole lot different than fixing cars, and planting problems for students to troubleshoot isn’t always as easy as it would seem.

For instance, I tried to ‘bug’ a ,91 Chevy S10 trainer vehicle for my students by spraying water into the distributor cap, but after reinstalling the wet part, the 4.3L fired up ran just fine.  I was flabbergasted until I realized the water has to condense in droplet form on the underside of the cap in order for a moisture-related crossfire to occur. That’s why an engine might run just fine right after washing but stall out on the way back home from the car wash.

If the problem had been an intermittent concern, we would have connected the oscilloscope first, but since the engine never failed to dump about half its horsepower and torque at 2.5 grand, the students opted to take some exploratory measures before connecting the scope. 

I generally teach my trainees to check the easy stuff first.  What that means is that even though the ignition switch might be the first component in the problem circuit according to the schematic, it doesn’t make sense to drop the knee bolster to check the ignition switch connections only to find later on that the problem was in the easily accessible 42 pin inline connector on top of the engine. 

With that in mind, the student I had assigned to the Jeep removed the easily accessible distributor cap right away just to make sure there were no beads of condensation hanging around in there.  The cap was dry, and before connecting the o-scope he snatched the plugs out for a quick look.  They were in fair shape, but since spark plugs are cheap, the student screwed in a new set, but the sputtering and popping remained unchanged, as we all figured it would.  It was time to move to the next level.

Scope Doping

            The o-scope showed us an interesting pattern; to begin with, with the Jeep idling, the scope kept getting confused as to how many cylinders the Jeep had, and there was nothing we could do to change that.  The inductive pickup was properly clamped around number one plug wire, and even moving it to a different wire made no difference.  This particular scope doesn’t seem to like some vehicles very much, but it works just fine on most units. 

But when the rpm was increased, the scope gave us our first relevant piece of data.  Above idle, it knew full well that there were only six sparks popping, and at 2500 rpm, we literally lost every other cylinder in the firing order, starting with number five.  The Jeep uses the ancient 153624 spark pattern common to straight six engines, but above the 2500 mark it turned into 1—3— 2 —.  

Normal Pattern below 2500 rpm - bottom illustration is above 2500

Clamping the inductive pickup of a timing light around the plug wire leading to each of the six cylinder, we moved from front to rear and found that cylinders 4, 5, and 6 were all going dark and dead above 2500 rpm, an anomaly I either hadn’t seen before or had long since forgotten.  So what on earth could cause the engine to lose half its spark events at a certain speed?  Well, since this Jeep ignition system includes a microprocessor control unit (MCU) as a part of the ignition equation and we didn’t have any spare parts to use for an A – B – A swap, it was time to do a detailed analysis to determine whether the MCU or the Duraspark ignition module was at fault.

Jeep’s Strange Blend

While the Jeep Grand Wagoneer ignition system remained identical to a normal Duraspark setup until the end, the 83-90 4.2L Jeep ignition system engineered by AMC for the CJ/Wrangler series is strange mix, with a familiar old Ford Duraspark module mounted low inside the driver side front fender.  The system even uses a Ford style distributor and pickup coil, but with the distributor’s pickup unit feeding its signal to the MCU instead of the ignition module. 

   This MCU basically looks like a 60 pin Ford processor and is mounted between the passenger and the bulkhead, with the dashboard in between.  The MCU sorts out the pickup coil signal, then sends a pulsed ground in on the orange ignition module wire to trigger the Duraspark module, which in turn fires the ignition coil. This setup is in place to give the MCU control over ignition timing, but the system still uses a plain old vacuum advance.  There were a couple of vacuum switches and a temperature sensor that provided additional input to the MCU, but those had been disconnected and removed on this unit (Chris had recently replaced the engine), and they never seemed to make much difference anyway.  The vacuum switches and the vacuum advance have to be disconnected when checking or setting the ignition timing. 

Unless the MCU and its hardware has been done away with, the engine rpm should be raised to 1600 when adjusting the ignition timing on one of these, and the sticker specs under the hood call for 9 degrees plus or minus 2. 

To strengthen the ignition system’s foundation and on general principle, we adjusted the ignition timing to Jeep specs.  I’ve seen misadjusted ignition timing cause odd problems I hadn’t expected, and I wasn’t about to be made a fool of this time around.  The timing wasn’t set right, and we had to retard it several degrees and bring it into spec, but to no avail.  The sputtering and misfiring remained.

Taking the MCU out of the Loop

            Whenever I have built a firm foundation in regard to the integrity of the coil, wires, and spark plugs, I go to work on the ignition system’s superstructure to eliminate different components and portions of the circuit. 

I worked on a 98 Explorer a few years back that was sputtering and popping rather famously at 4000 rpm, and I determined that the problem was in the primary ignition signals to the coil pack. Acquiring some connector terminals, I built a set of primary wires to bypass the existing harness circuits between the PCM and the coil pack so as to find the root of the problem. It was a grand and glorious troubleshooting attempt, but the problem was still there.  The Ford field service engineer suggested that I replace the 12A581 engine wire harness, and since I was out of ideas, and to make a long story short, the Explorer cleaned up its act after I replaced the harness.  In spite of the fact that the harness fixed the problem, I get really annoyed by situations like that because they just don’t make sense; the bypass I engineered should have cleared up the concern, but it became obvious that another circuit was at fault, and I never could determine exactly what the problem was with the original harness.  Oh well…

Explorer ignition - the red arrow represents the part of the wire harness I overlaid...

As far as our Jeep was concerned, we decided to remove the ignition module from its mounting and use jumper wires to connect it directly to the distributor and ignition coil.  In this way, the MCU would be out of the loop. 

The ignition module has a two-wire connector with a hot-in-start circuit (white wire) and a hot-in-run (red wire) feed.  The four-wire connector is the business end of the module. The student quickly removed the ignition module and laid it on the passenger side fender so it could be rewired to take its signal directly from the distributor like an old Ford system. He attacked the job with alligator clipped jumper wires from Radio Shack. 

The black wire on the four-wire connector had to be grounded to the engine block; the module uses this ground to fire the coil (the harness anchor screw in the distributor bowl provides this ground).  The four-wire connector’s green wire connects to the tach side of the ignition coil, and the remaining two wires (orange and purple) connect to the pickup coil windings in the distributor.  With the jumper wires connected this way, we managed to fire up the Jeep and observe the scope pattern again using nothing but the Duraspark module itself to operate the ignition system. The symptom and the pattern were still there; the ignition module itself had to be at fault, and when I had a replacement part sent from the local jobber, we connected it to the jumper wires to be rewarded with victory.  Mounting it in the original module’s place and plugging it into the harness, we found all six spark lines.  Chris was standing by, and he was measurably impressed at how quickly the students located the problem.  Why the problem appeared immediately after Chris washed the engine remains a mystery.

Concluding Thoughts

Troubleshooting any system without being able to understand the big picture is a frustrating exercise.  And while it may have been a whole lot easier to just plug in a new module, I’ve been in too many situations where a known good part wasn’t available except across the parts counter, and we all know how unpopular we can get with the jobber if we use his new stuff for troubleshooting!  That situation turns even nastier when the part has to be ordered, particularly if it’s expensive!  And if the problem is an intermittent, somebody will wind up having to eat the cost part if the vehicle comes back with the same concern.  Can I get a witness?

Personally, I’ve found that I learn a whole lot more if I can gather enough hard data to determine (as nearly as possible) exactly what the heck is going wrong.  It’s a bit more time consuming on the front end but gaining insight and understanding adds to our mental library and makes us more accurate troubleshooters.                                    R.W.M.