Automobile ignition systems have never been more reliable, but when not, how does one troubleshoot? and is it really an ignition issue? This simple test setup costs only about $1 and may easily be taken for a road test to help isolate mind-blowing intermittent problems. It consists of 2 LEDs and a meter. The standard means of troubleshooting is for a mechanic to change boxes and pass the cost onto the owner even if it does not solve the problem—and some systems have multiple boxes—and it might not even be a box issue. This method will help isolate the problem before making large expenditures.
The occasion for this discussion
When my car (86 Dodge Aries) developed intermittent problems, I took it to my mechanic. It ran perfectly while he checked it out. His response: “If it ain’t broke, I can’t fix it.” So I kept driving it until it stalled in pouring rain. I then pulled the wire from the coil to the distributor and checked for spark and it was DEAD. While that was useful information, it did not come close to isolating the problem. Eventually, it restarted after healing itself for another week.
I recommend white and green LEDs because they are daylight visible.
Green LED: Power to the coil—cannot run without power!
White LED: Dwell—this term I borrowed from the old breaker point dwell time setting—it simply means that the switching transistor is sinking or conducting current. It should blink at low speeds—at least blinking was visible on my 4cylinder engine when at idle. The LED must be protected with a 1000V series diode because this voltage swings up to 250V or so when the transistor turns off.
I located the LEDs at the end of (3) wires and slipped the assembly under the edge of the windshield wiper so the LEDs could be viewed while driving or sitting in the car. The power LED in my case was red, but I found that it was barely visible.
The meter circuit simply peak detects the collector voltage that occurs when the switching transistor turns off. On the 500V range, the meter draws only about 25uA at mid-range and that will not interfere with normal operation. Again, a 1000V diode is used for the rectifier. Mine ran at 250V. If the voltage is too low e.g. 100V, the transistor could be defective. If too high, the spark plug gaps may be excessive or there may be a problem in the distributor. If it wobbles, there could be defective spark plug wire(s)—with typical carbon string wiring, an internal arc turns the carbon conductor into CO2 and blows away thus creating a gap inside the wire.
Interpreting the data
The power LED is obvious; however, if there is no power it may be because the ignition key is on, but the engine is not running—the microcontroller disconnects the power in the event the engine is not running to prevent the coil from overheating or discharging the battery. When cranking the engine, it must turn on! The means (relay) of disconnecting power during this condition is suspect. If it never turns on, there could be a connector or wiring issue.
The dwell LED indicates that the switch transistor is conducting, but it does not fully indicate normal operation as it must open briefly to generate the primary voltage—a leaky transistor cannot generate full voltage. If there is power, but no dwell current, the transistor is either open, is not being driven by the microcontroller or there is a wiring or connector issue. Note that if there is no power, there can be no dwell current.
Proper meter voltage (e.g. 250V) indicates that the transistor switch is doing its job.
In a previous life, I did aircraft electrical systems. Per my experience, half of all electrical problems were connectors. I believe that automobiles are similar. Inspect all pertinent connectors for corrosion etc. Use silicone lubricant on mating surfaces to seal out moisture. Do not use abrasive material to clean gold plated contacts as the gold may be very thin.
The jury is still out with my old car. I did learn that it was indeed a power issue and that the power to the fuel pump is interrupted at the same time—I normally can hear the fuel pump run quietly for 2sec when the ignition key is turned on. One curiosity with this particular vehicle is that it tends to flood the engine with fuel under slightly abnormal conditions—the only way to ever get it restarted is to unplug the throttle body injector and crank the engine until it starts and starves to a stop. In this case, it was NOT flooding and now I know why.
I am now awaiting a chassis wiring diagram so that I can dig deeper.