jvaldovi

Hello,
I have a 1995 Honda Civic DX, driven about 680,000 miles. It originally had a 1.5 liter engine with manual transmission. I swapped the engine and transmission a few months ago with a 1998 Civic LX manual 1.6 liter. I don't think my issue is related to the engine swap, as I've had the problem before. Under normal behavior, the following happens. I turn the key to the on position without trying to start it. The fuel pump will make a whirling sound (priming) for a second, then the check engine light will turn off. Then I can start the engine fine. Sometimes the following happens. I turn the key to the on position without trying to start it. I do not hear the fuel pump prime. The check engine light stays on. I try to start the engine. It cranks, but does not start. I wait a while and try again. It eventually starts after about 3 attempts. I have tried and checked the following:
1. Replaced the main relay with an Autozone relay
2. Sanded the ground contact surfaces at the thermostat
3. Checked the spark on two spark plugs with a screwdriver. There is spark.
4. I checked the voltage at the connector above the fuel pump. I got battery voltage, a little over 12 volts.
Item 4 leads me to believe I should replace the fuel pump. However, I am hesitant because it's only been a little over a year ago that I had a shop replace the fuel pump and fuel pressure regulator, both original Honda parts. That old fuel pump had lasted about 500,00 miles. Is there any other checks that I can do? I noticed that the connector at the fuel pump had some yellow grease. Is that supposed to be there? What is it for? There was another connector near the fuel pump. What is that one for? Do I need to check the voltage there? If so, then how? Additional information: I originally thought the issue was related to rain because it would happen after it had rained and it would take a long time to start after it rained. However, now it happens even in hot, dry weather. Thank you.

fcm

How is the fuel pump ground?
If good , try tapping/banging on the fuel pump cover [or gas tank] when key is turned to run, if fuel pump works replace it there is an intermittent connection problem in the motor winding. 94

jvaldovi

What do you mean "How is the fuel pump ground?". Do you mean the black wire on the connector to the fuel pump? If so, how do I know if it is OK? I will try banging on the fuel pump cover, but I will be surprised if this makes it prime, as it is not too old. Thank you.

Former User

Maybe a bad ECU.

jvaldovi

Is there a way to test the ECU? This past weekend I tried sanding the transmission ground. So far so good.

mk378

The fuel pump has a ground wire that grounds to a bolt under the back seat.

If you have 12 volts on the fuel pump wire (large green and yellow) at the tank, the pump should be running. Generally if the main relay or ECU is bad and the CEL staying on, there will not be 12 volts on the fuel pump wire.

Former User

No.

But you can test for the key power pins (A25, B1) and key ground pins (A26, B2) for the ECU. If good, swap in a known good ECU.

Also measure voltage on the MAP sensor pins with the key in ON(II) and engine off.

jvaldovi

I tried banging on the fuel pump, which did not work. I tried checking the Check Engine Light codes. With a paper clip, I connected the two pins on the two pin blue connector with green wrapping under the glove box near the door. Whe I put the key in the II position, the Check Engine Light stayed solid (no pulses). The SRS light kept blinking short pulses without pauses.

fcm

A solid CEL normally indicates a bad ECU. 94

mk378

CEL should come on and go out if you turn the key on and don't crank, with the test jumper NOT in place. With the test jumper IS in place, it will come on and stay on if there are no codes, and if there are engine codes they will blink out on the CEL.

Do you get 12 volts between the two wires to the fuel pump, at the gas tank? One is a ground that is grounded behind the back seat, and one comes from the main relay. If you have 12 volts at the tank but pump does not run, clearly there is a problem inside the tank.

95civichatchguy

I had this problem in my old 95 hatch and I changed everything then it ended up being the ition tumbler and key cuz it wasn't letting the key turn 100% try replacing the tumbler and key and see if that works cuz I mine was messed up and did the same exact thing than I soldthe car a year later and it never did that to me again while I owned it

jvaldovi

How exactly to I check the key power pins and key ground pins? I'm a novice with electrical, but I have checked a few things with my digital multi meter. Don't you mean if they are BAD, then try a new ECU? Do you think a computer repair shop may be able to diagnose if the ECU is bad?

Former User

No, the tests I mentioned are done on the wire harness, not the ECU.

Have you read your multimeter manual to see how to measure DC voltage and resistance or continuity?

Only if a component is obviously blown or some other obvious damage, which you can do all by yourself easily by opening the ECU and inspecting the circuit board.

jvaldovi

I opened up the ECU. With the ECU connectors connected and the key in on (II) position, Check Engine light consistently solid on:
I put the positive meter lead on the A25 wire inside the ECU and the negative meter lead on the chassis. I got 0 volts.
I put the positive meter lead on the B1 wire inside the ECU and the negative meter lead on the chassis. I got 0 volts.
I put the positive meter lead on the A26 wire inside the ECU and the negative meter lead on the chassis. I got 0 volts.
I put the positive meter lead on the B2 wire inside the ECU and the negative meter lead on the chassis. I got 0 volts.

I tested other wires inside the ECU on the A and B connector, just to be sure. Other pins did have 12 volts.

With the ECU connectors connected and the key in on (II) position, Check Engine light consistently solid on, MAP connector disconnected:
I put the positive meter lead on the left female socket in the MAP connector and the negative meter lead on the right female socket in the MAP connector. I got 0.02 volts.

Former User

Again, none of the tests I suggested is done on the ECU.

jvaldovi

I am assuming that the tests need to be done with the ECU connectors disconnected and the DMM test leads touching the connector pins. Is that right?

Former User

Yes

jvaldovi

I tested the key connector. It failed the ground blue/blue wire connector continuity test, with the key out, as specified. I unplugged the ground connector at the key. It has a pin that bobs in and out. If it is bobbed in, there is continuity. If it is bobbed out, there is no continuity. If I put a small piece of metal (from a fuse) in the gap, then assemble the connector, it is always bobbed in and it passes the test. However, I don't see how it is supposed to pass the test, since there is nothing to push that pin in with the key out. In any event, with or without the fuse fix, the fuel pump still won't prime. I'll try testing the ECU connectors. Thanks for your patience and help.

Former User

What specific pin/connector failed the continuity test? How are you doing the test? Post pictures.

jvaldovi

I had blown a fuse while testing the wires inside the ECU. I put in a new fuse and got the following new results.

With the ECU connectors disconnected and the key in on (II) position:
I put the positive meter lead on the A25 key power pin on the left 26-pin connector and the negative meter lead on the chassis. I got 12.3 volts.
I put the positive meter lead on the B1 key power pin on the middle 16-pin connector and the negative meter lead on the chassis. I got 0 volts.
I put the positive meter lead on the A26 key ground pin on the left 26-pin connector and the negative meter lead on the chassis. I got 12.3 volts.
I put the positive meter lead on the B2 key ground pin on the middle 16-pin connector and the negative meter lead on the chassis. I got 0 volts.

I tested other wires inside the ECU on the A and B connector, just to be sure. Other pins did have 12 volts.

I tested the ignition switch connectors using the procedure in the 92-95 Civic manual I bought at Autozone. With both battery terminals disconnected and the two ignition switch connectors disconnected at at the fuse box, I checked continuity between the following pins on the connectors:
Key position___Wire Colors Connected
I White/Black (Acc) - White (Bat)
II White/Black (Acc) - White (Bat)
II White (Bat) - Black/Yellow (Ig1)
II Black/Yello (Ig1) - Yellow (Ig2)
III White (Bat) - Black/Yellow (Ig1)
III Black/Yello (Ig1) - Black/White (St)
Key out Blue/White (Key) - Blue/White (Gnd)

It passed all tests except it failed the last test for ground. When I put a tape a piece of metal to keep the pin bobbed in, it passes the ground test. The attached pictures show the connectors for this test and the small connector with the pin that bobbs in and out. However, the check engine light still stays on and the fuel pump is still failing to prime.

jvaldovi

Hello,
My car eventually started. Can anyone draw conclusions from my test results? Did the pictures show up?

Former User

Problem solved?

safety switch

My '92 DX with a 1.5 refused to start with a solid CEL and no fuel pump. I tried the entire FSM decision tree and followed a lot of RonJ's other excellent forum posts (thanks dude), and then finally came to the conclusion (in no small part thanks to fractions of a volt at the MAP pigtail) it had to be the ECU.

Pulling the P06 ECU looked fine at first glance, but upon desoldering the C14 power-regulator(?) capacitor it turns out the cap took a little leak on its own feet, eating through two traces.

I replaced all of the electrolytic caps with new ones of similar quality from Digi-Key and repaired the traces as best I could, and the problem has not reoccurred.

At the age of these cars, by now it is almost certainly bad capacitors in the ECU, but luckily it is not an expensive thing to fix. It cost me just under eight bucks in parts and a lunch hour. The hardest part of the job was getting the ECU back in without breaking the door trim.

Former User

Nice job with the circuit board repair!

This test^, referred to as the K-test by TxGreaseMonkey, is a great way in some cases to identify a bad ECU for 92-95 Civics.

https://www.askmehelpdesk.com/cars-trucks/1989-honda-ecu-test-541702.html#post2659372

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

jvaldovi

Hello,
The problem reoccurred this morning after a long 100 mile drive, after leaving it off for about 1 hour. I will look into the K-test.

Are my test results OK?

What is the purpose of that small "connector" with two blue-white wires? Taping down that pin that bobs in and out had the effect of chiming when I opened the driver's side door, which turned off when I closed the door.