My first thought is to use my Chrysler DRB3 scanner to view "Inputs / Outputs" data and see if there's a signal coming from the crankshaft or camshaft position sensors. If there is, I would replace that sensor as a test. Some aftermarket scanners will not display that information.
If you do try to replace the crankshaft position sensor, be aware the air gap is critical. Your old one most likely had a paper spacer stuck to the end to set the gap, and it slid off as soon as the engine was started. You must install a new paper spacer on a used sensor, otherwise the sensor could hit the flex plate and be broken. Some replacement sensors have a plastic rib molded onto the end. It partially wears away during engine operation. To reuse one of those sensors, cut the remainder of the rib off, then stick on a paper spacer.
If there were no erroneous signals, I think I would measure the voltages on all of the pins in the connector for the Engine Computer when the problem is acting up, and again when the engine will start to see what changed. If, for example, the voltage was low on one of the 12 volt power supply pins, there might be a lose or corroded terminal in a connector in that circuit. You might also find a corroded ground wire where it bolts to the body. There should be 0 volts on all of the ground wires. Typically there are four of them. Two are labeled "signal ground" and two are labeled "power ground". "Power ground" doesn't exactly refer to power for the computer. It means those are the two ground wires for the high-power switching circuits for the injectors, coils, and various solenoids. The second ground wire is just redundant in case there's a problem with the first one.
The "signal ground" wires are for the various engine sensors. 0.1 volt change in one of the sensor readings is significant and the computer will adjust fuel deliver accordingly. If there is just a tiny amount of resistance in the power ground circuit, pulses of current flow when things are momentarily switched on will develop pulses of voltage drop. Those voltage pulses might only be a few tenths of a volt, but if the sensors shared the same ground wires, those voltage pulses would cause the sensor readings to pulse too and adversely affect engine performance. That's why there's two separate ground circuits.
Very often any resistance in those ground wires is way too small to measure accurately, but you can see the RESULTS of that resistance in the form of voltage drops. So, ... 0 volts is ideal on all of the ground wires to the computer, 0.2 volts is not uncommon, but anything higher or especially if it is pulsing means that wire needs closer examination.
When you measure the rest of the voltages to the computer, it is helpful to have the service manual showing the drawing of the connector with all of the wires, their colors, and their circuits listed. There can be up to 60 wires, but you can disregard unrelated things such as the wires going to the cruise control servo and emissions system solenoids.
The problem you described could also be caused by the Engine Computer itself. GM has a huge problem with their computers and it is the first thing people try, but Chrysler has very little computer trouble, especially with these older vans, so that's probably one of the last things I would suspect. Unfortunately there is no easy way to test it other than to replace it and see what happens. Before you get that far, unplug the crankshaft position sensor, then see if the relays stop clicking. If there's no change, unplug the camshaft position sensor and see what happens. If the relays keep clicking, at least you will know the problem is not related to those two sensors. The crankshaft position sensor is on top of the transmission bell housing, right next to the back of the engine. The camshaft position sensor is inside the distributor.
Friday, September 17th, 2010 AT 5:23 PM