Random no-code failures can be the most frustrating problems to diagnose because the condition usually cant be duplicated in the shop and the “fix” often cannot be verified by test-driving.
In this instance, I use the word “random” to describe a condition that might occur only once in two weeks. Sometimes I’ll find something suspicious in the form of an obviously worn or defective part, but not in most cases. Without being able to diagnose the problem or verify the repair, I find that the best approach to solving a random no-code complaint is to devise both a failure scenario and a diagnostic scenario.
To illustrate the point, I present two case studies. The first is a manual transmission 1999 Jeep Cherokee Sport, 4.0-L, that had a random no-code, no-cranking complaint, and the second is a 2005 Ford Crown Victoria with a random no-code, loss of power complaint. The Cherokee Sport is a standard model with few options while the Crown Victoria is fully accessorized.
1999 Jeep Cherokee Sport
The Jeep owner reported that his Cherokee often starts if the passenger door is slammed or after waiting a few minutes or perhaps a full day. The failure can happen at any time, but is most disconcerting when it happens on a deserted mountain trail. Among the moving parts in the starting system, the starter and the ignition switch have been replaced, leaving only the starter relay and the clutch switch. I’m speculating that the previous tech might have swapped the starter relay with another relay to test its reliability, but I have no way of knowing that.
If any of the above is true, solving this particular Diagnostic Dilemma involves failure probabilities. For example, could two starters, two ignition switches or two relays produce the identical random failure? I think not. Could it be a random open circuit in the wiring harness? Possibly, with the exception of the clutch switch wiring harness, but most of the open wire failures are the result of manufacturing errors early in the vehicle’s life.
At this point, I replaced the starter relay based on my previous experience with relays. I could replace the clutch switch as a preventive measure, but I want more evidence. Beyond replacing those parts, I might need to inspect and perhaps remove the wiring harness for repairs, which can be a very expensive and time-consuming operation that’s recommended as a last resort only.
Before diagnosing, a diagnostic tech must acquire an intimate knowledge of the system and how it works. In my case, I always begin by thoroughly examining a wiring schematic. On this Cherokee, the ignition switch powers fuse No. 19 in the instrument panel fuse box located in the passenger-side kick panel. Fuse No. 19 powers the clutch safety or “interlock” switch. The clutch switch provides primary circuit power to the starter relay located in the underhood power distribution center (PDC). The starter relay closes when the ignition switch is turned to the “start” position and the clutch pedal (switch) is depressed. When the starter relay closes, current is supplied through the No. 10 PDC fuse, which is a 20-amp maxi fuse that powers the starter solenoid. The starter solenoid simultaneously applies battery voltage to the starter motor and engages the starter drive with the flywheel. Apparently, harness connectors rarely fail in this particular system.
The Failure Scenario
Because this ‘99 Cherokee Sport is unreliable, the owner is willing to sell at a loss if the problem can’t be solved. The Jeep is in otherwise above-average condition. Probabilities rule in this instance because I can’t duplicate the problem, much less verify a repair.
So, do I gamble on a new starter relay? The answer is a resounding “yes” because one new relay is cheap compared to one diagnostic hour. Should I also gamble on replacing the clutch switch? Well, parts availability and expense might be issues for a 1999 vehicle, and according to my favorite diagnostic database, the clutch switch is usually reliable.
The Diagnostic Scenario
Let’s go back to the owner’s statement about when the no-cranking problem occurs: Sometimes the Jeep will crank after waiting a few minutes or an entire day; the failure can happen at any time; and sometimes the vehicle starts if the passenger door is slammed. From my experience, these symptoms sound like a failing starter relay.
The starter relay is activated each time the engine is started and should therefore develop the most wear at the contact points that close the circuit to the starter solenoid. Since the relay feeds a circuit protected by a 20-amp maxi fuse, the relay obviously carries more amperage than its companion relays.
The symptoms also indicate a classic problem with worn contacts inside the relay, which means that the roughened contacts occasionally might not make full contact. Consequently, the contacts might realign when a vibration — such as a door closing — is introduced into the chassis. And, since the PDC is located on the passenger side in the engine compartment, slamming the passenger door could definitely jostle the relay contacts enough to re-establish contact.
Another issue is the fit or “drag” between the blades or pins on the fuses and relays and the connectors inside the fuse and relay boxes. The thickness of both the fuse and relay blades is about 0.034”. Fortunately, a common universal blade connector is about 0.034” thick as well, so I fabricated a “drag” tester from a universal blade by narrowing it to fit in a fuse connector. Testing indicated that the drag was adequate at fuse No. 19 in the I/P fuse box, at the starter relay and maxi fuse in the PDC in the engine compartment. I also verified the relay connection by observing the scratches or “register” marks left on the contact blades of the relay.
Diagnosing The Clutch Switch
I’ve verified to the best of my ability the correct performance of every component in the starter circuit with the exception of the clutch switch. How do I diagnose the clutch switch? Let’s begin at the starter relay connector located in the underhood PDC.
My service information had a number of discrepancies on both wiring color codes and fuse location in the PDC, but according to my information, a yellow wire connects the clutch switch to the starter relay primary circuit.
After removing the PDC and removing the bottom cover, I located the yellow wire connected to the starter relay. For testing purposes, I stripped a 0.25” of insulation from the clutch switch wire and soldered a 12” length of similar yellow wire that could extend out of the fuse box and to the front of the vehicle. After insulating the connection with electrical tape, I attached a set of long meter leads so I could observe the clutch switch voltage when I turned on the ignition switch and pumped the clutch pedal. The voltage graph in Fig. 3 indicated that the clutch switch is working perfectly and consistently.
Is the no-cranking problem fixed or not? Sometimes it’s hard to find a smoking gun, but I did disassemble the old starter relay. As Fig. 4 illustrates, the surface of the relay contacts had a large pit burned into it, which is caused by the tungsten coating on the contact point burning away into the steel base metal. While it’s impossible to say that the random no-cranking complaint is fixed, I’d have to say that a worn starter relay is the most likely cause of the no-cranking complaint.
I also found that touching my test lead to the battery positive post will activate the starter relay and starter. The end can be safely tucked into the split loom on the battery ground cable. So, I’m going to leave the test lead in place and instruct the owner on how to safely use the test lead to crank the engine.
Why wouldn’t I mount a by-pass switch for the clutch safety switch below the instrument panel? In my opinion, because the clutch switch wiring flexes slightly as the clutch pedal is depressed, all I’m doing is building a stress riser into the clutch switch wiring harness. I’m also not fond of modifying any safety device like a clutch switch. All I can do is let the customer drive his Jeep and report back if he experiences another starter failure. If using the test lead cranks the engine, at least I’ll know that the failure is located somewhere between the ignition switch and the starter relay, narrowing the scope of my Diagnostic Dilemma considerably.
2005 Ford Crown Victoria
Last spring, an old retired friend by the name of Maury asked me to investigate an intermittent loss of power complaint on his 2005 Ford Crown Victoria. The Ford would lose power only when passing other vehicles on a long, open stretch of highway. The problem couldn’t be duplicated locally on trips of 20 to 30 miles. As with many higher-end Ford products of that timeframe, Maury’s Crown Victoria was equipped with an electronic throttle and pulse-modulated fuel pump. Either of these systems should store a diagnostic trouble code (DTC) if they malfunction, so I scanned those and all other on-board modules for DTCs, but found nothing.
Only thinking inside the box, I was sure that the loss of power complaint was caused by a bad MAF sensor, clogged fuel filter or worn fuel pump. Of course, most pulse-modulated fuel delivery systems are monitored closely by the PCM and, when the pump fails to produce the commanded fuel pressure, a trouble code should be stored. At the very least, the P0171 and P0174 lean codes should have stored. And, as with any electronic throttle system, double and triple redundancy safety features are built into the system that cause the PCM to store a DTC when even the slightest glitch occurs.
So, thinking within the box, I was deeply troubled by the lack of DTCs and by the predictable occurrence of the loss of power complaint.
The Failure Scenario
Fortunately, the human brain works at the subconscious level to solve problems. In this case, I began speculating on what type of condition would cause the electronic throttle to not respond to a wide-open throttle command from the driver. One possibility: most rear-wheel-drive vehicles equipped with electronic throttles include a traction control feature that closes the throttle if the PCM senses that the drive axle wheels are spinning faster than the front wheels.
Since the Crown Victoria is one of the last rear-wheel-drive vehicles produced, I called Maury to ask if he had his snow tires on. He did, but he would be replacing them with his normal street tires shortly. I asked him to report back on his loss of power complaint. A few weeks later, he called to report that his car ran perfectly with the street-tread tires mounted on the rear drive axle.
The Diagnostic Scenario
Thinking outside the box, I realized that Maury would be driving at least 70 miles per hour passing vehicles on long stretches of highway. Due to the difference in the rolling circumference between the snow tires on the rear and the regular tires on the front, it might appear to the PCM that, at 70 mph, the rear wheels were indeed spinning faster than the front wheels.
Keep in mind that rolling circumference on a radial tire is affected by tire casing size, tread wear, inflation pressure and vehicle load, all of which affect the distance between the center of the axle and the road surface. To measure rolling circumference, align a straight edge through the center of the wheel and then mark the tire and the contact point on the floor. A short carpenter’s level works especially well because exact vertical can be established. After the mark on the wheel rotates exactly one turn, mark another spot on the floor. The distance between the two marks on the floor is the rolling circumference of the tire.
A typical rolling circumference would be in the 80” to 90” range for most passenger and light truck tires. As a rule of thumb, rolling circumference shouldn’t vary more than 0.5” among all four tires. While this minor difference in wheel speeds wouldn’t be apparent at lower speed limits, the difference at 75 mph might be enough to cause a false traction control signal.
As for diagnostic strategy, we’ll have to wait until winter arrives before I can verify the solution to the above scenario.
Maury first called me in March and now, as of mid-September, he has had no random loss of power complaints. Of course, this particular problem could be a software programming issue in the PCM. Ford might have published a technical service bulletin about this problem, but I didn’t see any such bulletin during my preliminary research. Come winter, I’m going to measure the rolling circumference of both the regular-tread front tires and the winter-tread rear tires to see if I can verify the solution to this month’s Diagnostic Dilemma.