Hello. I'm becoming TOO knowledgeable on my 1990 Honda Accord with 98K miles, it's been intermittent and not drivable for months now. Too expensive to replace the ECU or distributor (again) to troubleshoot.

Last year replaced entire distributor, had no spark, this solved it. About 6 months later the CEL began staying ON occasionally instead of going OFF after 2 seconds, car would not start if it stayed on. Didn't hear fuel pump. So replaced the Main Relay, that corrected the problems temporarily. Week later, same problem, wouldn't run for more than 2 minutes.

Suspected the ECU Crank Sensor input to the Main Relay coil #2 wasn't happening which would cause no fuel pump activation, clipped wire from the ECU at the relay connector and grounded that relay coil terminal. Car started right up, took it to the gas station, ran fine for 15 minutes. Next day, 2 minutes, then wouldn't start again, fuel now but no spark.

The relay seems to work fine, checked the voltages going to the injector resistors/ECU, just 0.3V below battery voltage input. From reading others' problems, next I would like to make sure the ECU is getting good voltage, anyone know what connector or pin on the ECU?

I guess what I'm asking, is there any way to separate whether it needs an ECU or distributor? CEL code checked, it says bad ECU (been saying that for months). Is that accurate enough to replace it for $300-400 at Autozone? Or I could go into the distributor and check the ignitor (at Autozone--unreliable) and the coil (resistances, also not very reliable).

Thanks in advance, appreciate any help. BTW, there seems to be a LOT of problems in this area after 20 yrs of operation...

Does the CEL come on for 2 seconds and then go out, when the ignition switch is turned to ON (Position II)?

About 50% of the time. Used to think that was a key indicator, if it didn't go out, no way would it start. But now, it doesn't seem to matter. Yesterday got it to start after about 25 attempts, ran for 1-2 seconds. Confirmed spark (at a plug) matches ability to start.

The source of your problem is likely a failing ECM. Perform the K-Test on the ECM:

The K-Test: Remove the MAP Sensor connector and turn the ignition switch to ON (Position II). Using a multimeter, check for 5 volts going between the MAP Sensor connector's reference wire (+) and ground. As you look at the connector, this is the socket on the right. Really press the black test lead into a cleaned main ECM ground on the thermostat housing. If the voltage is low, it's probably indicating ECM failure. Most failed ECMs will record a fraction of a volt. To me, the K-Test is simple, elegant, and accurate.

If it fails the K-Test, install a new ECM--ensure it's an exact Honda Part No. match. Shop for price and warranty.

TxGreaseMonkey, thanks so much for your help. I've been reading several of your responses, you are the "book of knowledge" on this subject for sure.

I will do this test tonight. I saw a reference to the "k-test" but didn't know the specifics, thanks. If bad, I was planning to match the p/n at Autozone, they said there's maybe 20 p/n's for that year/model. That to me sounds like Honda had some issues with their ECU design. I was hoping to get one at a junk yard, but with that many p/n variations, slim chance.

I'll report the results.

Yes, the electrolytic capacitors often start to leak--then, problems begin.

Well not exactly the numbers we were looking for on the MAP sensor connector:
Right socket: 5.02V steady
Mid socket: 4.86V
Left socket: 0.0V

But I guess it still could be the ECM? Is there another test, maybe put an oscilloscope or DVM on either the output of the ECM to the Ignitor, or on the Ignitor input, whichever is easier to get to?

The ECM appears fine. Next, I would check/replace the ignition switch for wear or pitting:

1. Disconnect negative battery cable.

2. Remove dashboard lower cover and knee bolster.

3. Disconnect 5-P connector from under dash fuse box and 7-P connector from main wiring harness.

4. Remove steering column covers.

5. Insert key and turn it to "0".

6. Remove two screws from switch and replace.

I would also check all under dash and under hood fuses with a test light or multimeter--be very thorough. Be sure to focus on the ACG (ALT) (S) fuse in the under dash fuse box.

Ignition Switch. This is a intermittent condition, it will spark maybe 1 second out of 10 tries (on those that the CEL turns OFF), but still no start. I can pull the ignition switch/wiring and jumper across for ON and then for START, if it starts then that's the problem.

I will check all fuses in the driver's kickpanel, under the dash and in the engine compartment for good voltage both sides with ignition switch ON. Good idea, we only checked fuse condition, could be an intermittent connection.

When you turn the ignition switch to ON (Position II), does the Check Engine Light (CEL) come on and then go off after 2 seconds? If not, the problem is with the ACG (ALT) (S) fuse, in the under dash fuse/relay box; ECM (perform the K-Test); main relay; or the ignition switch. Perform tests in that order. If the CEL does not go out, most Civics will not spark and the fuel pump will not run. That's why we begin our diagnosis here. Therefore, all effort should first be directed to correcting that problem. That's as simple (a plan) as I can make it. Unless you have jerry-rigged something, you should be able to get the CEL to go out.

Be sure to remove and clean the contacts to the ACG (ALT) (S) fuse. Also, remove and clean the main ECM ground, located on the thermostat housing.

Will do all, thanks. All is stock BTW. I have to go out of town for a few days, will pick this back up next week and report results.

The CEL comes ON and turns OFF after 2 seconds reliably lately (8 repeats). But does not start. Fuel pump is working fine (bypassed ECU input to Main Relay), you can smell gas.

The kickpanel fuse box has no "ACG" fuse that I could find. It does have an ECU fuse which I tested and pulled then put back. I pulled and put back all of the fuses in this panel for good measure. I don't know how to attach a picture file, but I have one.

Next I checked/tested all the fuses in the under-hood box, especially the one labeled "ECU".

Next I tried to locate the thermostat grounding terminal, none found. I did find ground wires attached at the other end of the engine which I removed, cleaned, tested and put back.

Tried to start it, CEL is fine ON then OFF, 2 sec. No start, tried several times. Pulled the distributor apart, removed the coil and ignitor. Tested the coil at Autozone, it's fine (0.9 ohms primary, 55.6K ohms secondary). None of the Autozones have the "Gen 3" tester required for the ignitor. Autozone checked, the distributor I bought 2 years ago is still under warranty, they suggested I just replace it to help isolate the problem. This is next.

If the distributor replacement still doesn't work, I'm considering buying a rebuilt ECU on-line for $60. Or, if I can find a wiring diagram, put an oscilloscope on the ECU ignition out and then the Ignitor input to check the signals.

Since the CEL is working properly, and the fuel pump runs, you want to focus on the distributor. Since your distributor is under warranty, try installing another one from AutoZone.

FYI, my experience has been bench testing coils and ICMs is virtually worthless. I've seen too many that test fine that don't work in all sorts of vehicles.

I hear you on the bench testing. 2 yrs ago I had the ignitor tested at Autozone, tested BAD. Bought a new one, installed, still no spark. They accepted the return if I bought a new distributor, I did, installed and the car ran fine till now. Now about the same problem. I'll replace the distributor next. Thanks.

To end the suspense, the car is now RUNNING fine!

I reassembled the distributor and pulled it, Autozone gave me another remanufactured unit, installed it, started on the first turn of the motor. Good acceleration and smooth at idle, just like it was for 20 yrs. So either the coil was bad dynamically (couldn't do a hi-pot test), or the ignitor. I reused the 2 yr old rotor, distributor and wiring which were is very good condition (these didn't come with the distributor).

So it wasn't the ECU after all. But what caused the distributor to "degrade" over time, could a poor signal from the ECU cause degradation say, in the ignitor? Just curious.

Thanks for all your help and guidance TxGreaseMonkey, I surely appreciated it.

Intense heat is the killer of Honda distributors. Over time, the igniter, coil, and internal sensors (TDC, CKP, and CYL) start to break down due to heat. It's particular bad on Hondas, since they place the igniter and coil inside the distributor, where there's little heat dissipation. Degradation isn't caused by the ECM at all.

Have you fixed the problem?

Yes, I can confidently say the problem is fixed now after 1 month of good running. Best I can determine is that the refurbed distributor that lasted only 2 years (original lasted 20 yrs and 100K miles) had an already weakened Ignitor unit, coil tested good.

Awesome, I'm about to get a new distrib. Myself. No spark, cranks fine. Had to fabricate my own ignition switch assembly. Coil tests good igniter tests good but I'm thinking looks are decieving. Main relay tests fine as well. When the key is left in the run position the coil heats up quite a lot to be almost too hot to touch and kills the battery. I even tried to get spark right from the coil itself, no luck there either. CEL stays on with code 5 but if I understand correctly the MAP sensor doesn't prevent sparking? MAP circuit tests good as well Leaning towards igniter or sensors inside distrib..