Hey, long time no post.

I'm trying to help my buddy with his 1994 Accord and 67,000 miles (just 67, not 167 etc) and think we need some help.
First, is there a way to get check engine lights?
Second, it has been randomly stalling on him. Now it's doing it like immediately after sitting for a week in my garage.
So far, a shop looked at it and replaced the distributor, which had all the engine speed sensors in it - there is no separate crankshaft position sensor.

We put a fuel pump meter and inline spark tester on it and it conveniently stalled for us right away. We replaced the main fuel pump relay and the fuel filter, which was SUPER dirty inside. I've never seen so much come out of a fuel filter. It must have lost a full teaspoon packed with mostly corrosion. I thought that might be the trick because it ran for like 30 minutes no issue last night, but I never realized that it has a check engine light on or that the D4 flashes about 2 times per second when it stalls.
When you first put the key in and turn it to run (before starting), the D4 light comes on and stays on. Turning the key from off to on today made the main fuel pump relay behave inconsistently. Sometimes I got nothing, sometimes it primed for the full 2 seconds, and sometimes I heard the relay and fuel pump turn on off on off.

It reminds me of when I worked on this terrible motor home and the ECU failed. That case had the main fuel pump relay just buzzing. This on/off business is weirder to me.

The fuel pump is in the tank and there is no access port. I found a youtube video of this exact issue and the fix was replacing the ECU.

Can anyone help with this?

Thanks,
Volvodude (still need to reassembly my 02 civic ex after the head gasket job)

 

Update: I found and read the codes, but the Check Engine light didn't flash - the D4 light flashed long short short short short, which I read as 14. I found one site that says idle control valve OR bad ECU.

 

A bit of info - ambient temperature is about 90F in my garage.
Update 2, I sat and fiddled with it for a few minutes and I think it's a bad ECU or bad wiring to it.
After I checked the codes and disconnected the jumper with the key on, the CEL went off. I started the car and left it running. I checked on it 20 minutes later and found this:
Engine off at almost fully warmed up. Check engine light was on and D4 was flashing 2 times a second.

Pulled the ECU fuse and inspected the socket and fuse again - looks great, just couldn't remember checking it before.
Turned the key to Run without cranking and the CEL was off.
Main relay turned on to prime, then turned off early. Flickered on and off a few times, then the CEL came on.
Main relay continued to flip on and off sometimes 5 or 10 times, sometimes it stayed on for a second, sometimes it just stayed off - but it kept doing this for more than a minute. The pump shuts off after 2 seconds, but the relay always has power in KP2. The 2-second prime/off/wait is all inside the relay.

 

If you have inconsistent functionality of power, turn the key to ON and jiggle the key around, if power drops out then your ignition switch is worn out and should be replaced.
~$20 for the switch and a few minutes to get it out.

Not to be confused with the key lock cylinder($$$).

If power does not drop out it may be a wiring or failing main relay or faulty main relay solder joints.

Normal operation of car when turned to II(ON) is the warning lights will all come on(prove out) then certain lights will turn off. This is done in time with the main relay turning on to prime the fuel system and then turning off.

To properly check codes on a 94/95...

Verify the CEL lamp on the dash is functional, at prove out it will illuminate, if it doesn't most likely the bulb is burned out. Replace bulb.
Can borrow a neighboring lamp if need be.
Each applicable warning lamp has its own codes. CEL is one set, D4 is another, as well as ABS.

CEL code 14 is NOT the same as D4 code.

You will need to grab the two wire blue connector from below/behind the glovebox door, jumper it and then turn ignition to II(ON) to see any codes. Usually goes in order of CEL>D4>ABS if applicable codes.

Rarely is the ECU at fault. It's just the easy blame.

If the only code you have is D4 Code 14 then that is the FAS wire, which is the communication wire between ECU and TCU. This code can be tripped if there is inconsistent power.

Fix all warning lamps, verify the ignition switch is making proper contact(key jiggle test), and then move onto main relay.

 

Good points.
CEL comes on at key Run, then goes out after a few seconds as expected. Doesn't come back on until the engine dies. Shortly after, the D4 light starts flashing twice a second. Recently, the (brand new) main relay started to flicker on and off.

Maybe the flickering main relay causes the CEL/D4 flashing.

I get no feedback from the CEL when I try to check codes. When I put the key into Run with the service plug jumped, the CEL stays on and D4 flashes the 1-4 code. It repeats without stopping. The CEL never changes and stays on steady.
Ignition switch also occurred to me, and I tried jiggling the key, but there was no correlation between me not touching the key or moving/tilting/pushing it in any direction. I didn't pull on it though. Do you know what the pinout is so we could bypass the switch? I do have a Hanes manual, but I don't know if I should trust it.


In 22 years, I have seen 2 bad ECUs. One had water damage from a leaking windshield and made the fuel pump relay buzz with a high-pitched whine. This one gets so hot that the case is uncomfortable to touch, like 140F. There are only 2 components that use the ECM case as a heat sink, so I'm pretty surprised that it can get that hot without one of them faulting. In open-air, it was at 105F with the ambient temp being about 95F.

What do you think of not being able to pull codes?
I did try to resolder the 3 components that look like transistors or SCRs. Had a little difficulty because the entire board has a conformal coating maybe of rosin. Any solder that touched it stuck and needed desoldering braid to remove it. Resoldering the components didn't help.

A replacement ECU will be here tomorrow. Do you think this new main relay could be bad? I need to pull it out and backprobe it and substitute in the original.

Thanks for your thoughtful reply.

 

We dropped in a repaired/rebuilt ECU. It resolved the issue. It's under the passenger side carpet, which you need to cut about 3 inches off of up on the left side of the footwell to fold it back. It's on a plated held in with 4 10mm nuts and mounts to the plate with more 10mm bolts. The TCU has to come off to reach one of the bolts to the plate.

The big difference between the two ECUs when running was the temperature of the case. The old one was too hot to comfortably touch for more than a second. The new one was about body temperature. Both after running for about 40 minutes in similar ambient temperature.
There are 2 IC components that use the case as a heat-sink. I think that one of the voltage regulators/transistors failed.
It has 3 of these (only one touching the case) - https://www.digchip.com/datasheets/p...-3001S-pdf.php - and a replacement is https://www.digikey.com/en/products/...-0-NOPB/148141 . Critical features are 1A/10V output and 12V-26V input. It's a "low dropout" type, which is very important to match. TO-220 package with 3 leads in PNP outputs.

It has 1 of these https://www.digchip.com/datasheets/p...-3001S-pdf.php and a replacement is https://www.mouser.com/ProductDetail...2I1IuvxQ%3D%3D . Critical features are the 5-pin setup (2 pins are soldered together on the back-side, 1 is cut off) providing 5V/1A or more output and input voltage 6-30V but 26V max input should be fine.

We may try replacing these 4 components on the old one.

No check engine lights, stalling, weirdness, or transmission issues have followed replacement.

 

You probably have blown caps in the ECU. It's more common than you think.

I had a 93 with similar issues, no CEL, but flashing D4 light. Replaced the caps and sold the car 100k miles later with no further issues.

You'll be able to tell by dark spots near the capacitors, on and under the traces and board coating. Most caps are used to filter, or stabilize power. This could cause the FETs to heat up.

 

Ooh, blown caps? Like visibly blown electrolyric caps? I did take photos and I don't see anything obviously blown. I've seen that a few times on mother boards from 2005-2009.

[img alt="Inside bad ecu angle 1. 1994 accord
ECU model 37820-p0b-a51. Possibly a serial number: 381-189438"]https://cimg1.ibsrv.net/gimg/honda-tech.com-vbulletin/2000x1504/20250826_192231_b3e7a4b368d34935bc315cec02ff830aff 7b4434.jpg[/img]
Inside bad ecu angle 1. 1994 accord ECU model 37820-p0b-a51. Possibly a serial number: 381-189438
[img alt="Inside bad ecu angle 2
There are no components on the back."]https://cimg2.ibsrv.net/gimg/honda-tech.com-vbulletin/2000x1504/20250826_192227_aee9f1159a8cc3e7b5590ace3d55682ffe 29b7f8.jpg[/img]
Inside bad ecu angle 2 There are no components on the back.

 

They usually leak, but I don't see any evidence of that in the pics.