Intermittent Insanity

ave you ever had a vehicle with an intermittent problem in your shop? Of course you have, and you know all too well how difficult and time consuming it can be to diagnose correctly.

Recently, I had a call on a 2006 Dodge Durango that had an elusive intermittent problem. The gauges would quit working at times, yet the odometer would light up but wouldn’t indicate new miles while driving. If the problem was present at key on, the radio and power windows would not work for about 20 seconds after the key was turned ‘on.’ The gauges could work correctly for a period of time, and then some days, only intermittently. The dealer had already installed a new instrument cluster and an ignition switch in an attempt to repair this with no change in the symptoms.

The instrument cluster replaces the Body Control Module (BCM) on this model and directly or indirectly controls many electrical functions. The instrument cluster is referred to as the Cabin Compartment Node (CCN) when connecting a scan tool or when referencing its function as a BCM.

I asked the technician to connect a scan tool so that any and all codes could be recorded before beginning the diagnosis. The scan tool showed a code B2107 in the CCN, which indicated a problem with the ignition switch status input to the CCN. The code would come back almost as soon as the key was turned ‘on’ even if the cluster was working at the time. The trouble tree for this code directed the technician to check fuses 1 and 27 in the Integrated Power Module and fuses 9, 19 and 34 in the junction block.

When checked, all of these fuses were good and all had power to them. While checking the fuses in the junction block, the technician noticed that even the slightest movement of a wiring harness or connector near the junction block would affect the operation of the instrument cluster.

The ignition switch inputs to the CCN were then monitored using the scan tool. The ignition switch run/start (F21) circuit status would change from ‘True’ to ‘False’ at times, yet the instrument cluster was working OK at the time. The scan tool was then set up to display data in a graphing mode of the ignition switch run/start circuit. This display was much easier to see than when just looking at the normal data display while trying to lightly move the wiring harness near the junction block. The graph on the scan tool clearly showed the voltage to the cluster was coming and going much like a square-wave pattern. See Figure 1.

When the technician unplugged all the connectors at the junction block and inspected all of the terminals, they looked clean and tight. The problem seemed to be in the junction block, but the technician could not pinpoint the cause. The fuses feeding the instrument cluster never lost power but finally the technician noticed something interesting when rechecking voltage on fuse 19. If the voltmeter probes were barely touching the fuse the problem was gone. Fuse 19 feeds ‘key on’ power to circuit F21 to the instrument cluster. He removed fuse 19 and looked at the terminals. The terminals in the junction block on one side were spread slightly further apart than the other fuse terminals. See Figure 2.

After carefully bending the terminals in the junction block to grip the fuse terminal slightly tighter, the problem was gone. It is important to note the instrument cluster would sometimes operate properly even though the power on the Run/Start circuit (F21) to the cluster was coming and going according to the scan tool. This is because the instrument cluster normally stays powered up for about 10 seconds after power is lost on the F21 circuit, which causes a delay of any symptoms to appear after the intermittent loss of voltage on the F21 circuit. This is what caused the difficulty in diagnosing this tough intermittent problem.

Imagine how many parts could have been thrown at this problem had the shop not decided to call us when they did.