Hi,

I have a 1992 Honda Civic DX 5-speed that has been my daily driver for the last 18 years. Driving home last night, the engine stalled while waiting at a stoplight and the engine would not start up again.

After getting the car towed home, I have done the following:

1. Cleaned the battery connections.
2. Checked the under-hood fuses (all good)
3. Checked the under-dash fuses
3a. The starter signal fuse is not showing any power, all the others are good.

When I turn the key to the On position, the CEL comes on for 2 seconds, I hear the fuel pump, and I hear a click (main relay, I believe) under the dash that coincides with the CEL going out.

I notice a fuel smell at the exhaust when cranking the engine over, so I assume it's getting fuel. I checked the plugs and I also noticed a fuel smell.

I thought it might be the ICM or coil, so I had them checked and both check out OK. I realize this could be a false positive, but they're relatively new so for now I suspect that they are actually OK.


Based what I've read, I believe that with no power to the starter signal fuse my next step is to look at the ignition switch as a possible culprit. Is this the right way to proceed or is there something else I should check first?

The starter signal fuse will only show power, when the clutch pedal is depressed.

Since the CEL comes on and goes out after 2 seconds, when the ignition switch is turned to ON (Position II), the problem is likely distributor related. Usually, this involves the ICM and/or coil. Did you apply silicone heat transfer compound to the back of the ICM, when you last replaced it. If you didn't, they won't last long.

It's easy enough to check the ignition switch for wear or pitting; however, I think the ignition switch is fine. I suspect the ICM. Have you checked for spark?

EXAMPLE: Diagnosing an Engine That Cranks but Won't Start

Check for spark at the plugs. If none, check for spark at the coil, after removing the distributor cap, rotor, and leak cover. Ground the test light clip and check for spark from the coil, as an assistant turns the ignition switch to Start. If the coil works, there should be a spark between the spring on the coil and the test light, as the probe is brought in the proximity of the spring. You now have spark at the coil and no spark at the plugs; therefore, replace the rotor and distributor cap, since one or both may be shorting out.

If there's no spark from the coil, check the igniter with the test light. Take a jumper wire, with alligator clips on each end, and ground out the coil spring. Touch the test light probe to the negative terminal on the coil, as your assistant turns the ignition switch to Start. If the igniter is good, the test light should blink on and off. If the test light does not blink on and off, replace the igniter. Should the coil still not spark, after replacing the igniter, replace the coil.

In real life, everything can check out but the engine can still die on you, once it gets hot and the igniter and/or coil start to break down under load. That's why I'm not a real fan of testing igniters and coils. This diagnostic process may keep you from just throwing parts at the problem—just be aware of its limitations.

Bottom Line: ICMs and coils really take a beating in Hondas.

If the problem persists, I would install a new genuine Honda distributor housing. The internal sensors (CKP, TDC, or CYP) may be bad. Aftermarket ICMs and coils are fine, however, and will help keep the cost down.

Thanks for the excellent info. The test light flickers when I touch it to the coil spring, but I don't get any spark between the coil spring and the test light.

When I ground out the coil, I get a solid light on the negative terminal, so I'm going to start with replacing the ignition control module, then test for spark from the coil. I have some Arctic Silver 5 from a recent computer build project, so I'll use some of that between the ICM and heat sink.

Hopefully this will be enough to get it running.

Nice job. It sounds like the lack of heat transfer compound on the back of the old ICM caused it to fry.

Thanks TxGreaseMonkey -- My Civic is running once again. It turned out to be the coil that was the real problem - may have been both coil and igniter! It started up pretty rough the first time and seemed to be running very rich -- lots of smoke and such. That seems to have gone away after a short test drive.


For those that have the same problem, here are some things I learned along the way:

1. Always trust TxGreaseMonkey :)
2. Buy a cheap test light -- it will pay for itself in no time.
3. On a manual transmission car, the start signal fuse will only show current when the clutch is in and the starter is engaged.
4. Likewise, the high/low beam headlight fuses only show current when the headlights are on (and either high/low beam).
5. When checking for spark between the coil spring and test light, you will actually see the spark jump a gap of an inch or so. If you touch the test light and it lights up intermittently, the coil is probably bad.
6. When checking the negative terminal on the coil, be sure you're really checking the negative terminal.
7. Even if a part tests good, it may be bad.

Most of all -- you can do this!

Yes, parts testing good, that are actually bad, is difficult for many people to believe.

Thanks for the compliment. Here are some suggestions, so that your Civic lasts a long, long time:

https://www.askmehelpdesk.com/cars-trucks/faq-how-troubleshoot-repair-maintain-hondas-selected-other-vehicles-46563.html#post219991

Reflect on why you think the coil died. Keep in mind that distributor cap contacts can oxidize or corrode, creating high secondary resistance, which can result in rotor and ICM failure, reduced mileage, CEL coming on, catalytic converter failure, or failed emissions test. The secondary ignition system will destroy the ICM and coil, if you do not properly maintain the ignition system. This means replacing the distributor cap, rotor, and spark plug wires every 60,000 miles.

Here's a great way to quickly test all fuses, without having to depress the clutch or worry about anything else:

Set your multimeter to audible ohms, DISCONNECT THE NEGATIVE BATTERY TERMINAL (or you could blow the fuse in the multimeter), and touch the red and black test leads to the test points on each fuse. When you hear the meter buzz, you know the fuse is good. This is my favorite method. If your meter doesn't have audible ohms, but it does have ohms, you can still use this method--just read the meter instead. This method performs a continuity test on the fuse. Just remember to de-energize the circuits, by disconnecting the negative battery terminal, and record any radio activation codes first. This approach only tests the fuse, not "power through the fuse."