CarTalk.com Blogs Car Info Our Show Deals Mechanics Files Vehicle Donation

1997 Maxima - 14 igition coils failed over last 5,000 miles and they keep failing

When they fail, they get extremely hot and the insides melt out.

All six original coils and 8 new ones from several brands, including OEM have failed this way.

New ones have failed in as few as 3 miles, or as long as 3,000 miles. It is sudden and totally random.

The sparkplugs have been replaced with the correct part and gap to rule them out, coils keep failing.

The only known issue with the car is a supposedly bad diode in the alternator as shown on a test, although this doesn’t seem to cause any problems and charging voltage looks smooth on the oscilloscope.

I am at my wits end with this and have no idea what to do.

You said you changed the spark plugs, but didn’t say if you changed the spark plug wires. Did you?

They are coil-on-plug so no wires.

More than likely the ECU is bad.

Back in the day with point type ignition systems, there was a ballast resistor/wire to the coil. This was to prevent damage to the ignition coil once the engine started and the charging system came on line.

Then when things move to solid state ignition system systems, the ignition control module performed this function.

But now with direct ignition systems, the ECU controls the voltage to the coils once the engine starts.

Tester

I take it that the oscilloscope is only hooked up for a brief period and therefore isn’t reading over the road voltage. I would take care of the alternator issue before replacing any more coils.

If you have access to an oscilloscope I’d recommend looking at the pulse duration and voltages on the coil primaries from the ecu.

This is what the signal looked like, appeared to be identical on each cylinder.

The scale is 0.2 volts and 0 is the first line from bottom. So the signal has a peak of 1.2 volts and minimum of 0.6 volts roughly.

If this waveform and voltages are good or bad, I don’t know.

Well, this IS your problem. Part of the diode’s business is to convert generated A/C current to D/C, as well as preventing battery current to flow back to the alternator. With a bad diode you would be lucky your ECM is still alive.

The computer is able to tell if the spark plug gap is getting wider (which it does gradually with miles driven), and when that happens it increases the current through the coils to compensate. More current is needed to jump a wider plug gap. This process going wrong can burn out the coils, and that’s what I’m suspecting. What’s going wrong though, hard to tell. Make sure the spark plug is making good connection to the coil pack. Also that the engine is well grounded to the chassis. Look at the engine when it is dark to make sure there aren’t sparks jumping around. The coil packs have to be properly mounted to the engine to carry away the heat they generate of course. The only way to determine if the diode problem is contributing is to replace or fix the alternator. I wouldn’t try to debug this problem myself until the alternator was know to be working properly.

The car is aged since it’s a '97 with ??? miles. I would run a compression check; both dry and wet. Abnormally low compression can cause incomplete combustion which in turn leads to plug misfires (even subtle ones) and in turn take out the coils.

That waveform looks terrible. The backside of the waveform should be a relatively steep drop and the amplitude is not good.
Things to check- make sure the plugs are clean even though you’ve changed them recently. Not sparking at the right timing will build heat in the coils.
Make sure your 12v DC bus is clean and up to specs- you checked with the scope so this should rule out a bad diode causing this issue. Insufficient voltage will also cause overheating.
You could check one of the coil drivers with it disconnected from the coil and an appropriate load resistor across the terminals of the connector. It should be nice squared off waveform then.
Be sure you’re not chasing your tail, if the coils have all been damaged again, you may be getting bad results from that.
Finally, you need to assess the health of the motor by checking compression etc as mentioned by others.