How is a fuel pump CEL generated?

[Talesin]

I'm curious as my ECU has started to throw one (DTC 42), though the car still seems to run all right. Looking at the replacement VFAQ, it appears to only have two lines going to the pump, which would be power and ground... nothing to carry back a diagnostic or trouble-signal.

So how does the ECU 'guess' when the fuel pump is starting to have a problem? Fuel trims dramatically out of whack? I'd rather look at 'upstream' parts first, given that the pump appears to be working just fine.

I've pulled out the pump to have a look, and visually nothing seems to be too far off from how it should be.. I was mistaken though, as the pump only has ONE wire going to it (power) and is grounded to the pump bracket assembly.. the other three wires are for the fuel level sender and low-fuel idiot light.
I'm probably going to slide the low-fuel sender down its adjustment rod a little bit, as I seem to get the warning far too early (assuming it's actually a 16-gallon tank, like in the manual and every other reference I've been able to find aside from word-of-mouth).

I will be picking up a new Walbro 255 on Monday.. hopefully this will correct the problem, but I still am curious as to how the ECU 'guesses' that the fuel pump is dying? Again, what causes DTC 42 to be thrown? I'd really like to know so that I can check other parts which may have come loose, before laying out a hundred bucks for a new pump, when mine might be perfectly fine.

 

[steve]

What's not shown on the FSM schematics is that ECU pin 13 is the sense line for the fuel pump.
If the ECU activates the fuel pump and it doesn't see pin 13 go high it throws the DTC 42.

Where you will find this documented is in the 1G Technical Manual.

There are a bunch of things that can cause the code to be triggered but it boils down to not seeing 12v at pin 13 when it thinks it should.

Steve

 

[Talesin]

Just pulled out the FSM Vol2 to check the diagram. I may be missing something, but it looks like the only things on that particular line are pin 13, the MPI relay and the fuel pump itself.

Given that the fuel pump does appear to get power (the car starts and runs, after a few tries in the morning to prime the fuel line) I'd say that the MPI relay would be less likely to be the culprit, and it's probably just a dead/dying pump at this point, or potentially a partial chassis short. No real loss as pulling the pump showed that I only had the stocker Denso unit... just a REALLY whiny one for some reason. I'll try pulling the fuel pump connector, grounding the check plug under the hood and hitting the power line with a multimeter to see if there's a significant drop compared to battery voltage.

Thanks very much for the help Steve.

Wouldn't happen to know the legality behind scanning a FSM and providing a version to others, by the way? My copy of volume 2 is pretty much falling apart at this point.. if I don't scan it for my own use soon, I'll be left with a sheaf of loose pages, and will likely lose the important ones first. :b While I'm at it, may as well do volume 1 as well... the only FSM CDs I've been able to find are for the '91 or '93, and I have a copy of the '90 version... though I'm lacking volume 3, covering the AWD system.

 

[steve]

Just pulled out the FSM Vol2 to check the diagram. I may be missing something, but it looks like the only things on that particular line are pin 13, the MPI relay and the fuel pump itself.

Nope, that's it.
I do see ECU throw this DTC from capacitor leakage damage but a bad pump, burned or bouncing MPI relay contacts, or a wire going open could cause it too.

Steve

 

[Talesin]

The caps in my ECU were replaced before I bought the car, though the person who did it did a semi-crappy job... left the complete legs on the new solid-state caps instead of trimming them down. Didn't see any damage the last time I opened it up.

In any case, new Walbro 255 is installed, and the ECU is still tossing DTC 42. Car runs and romps along (though still doesn't hold any fuel pressure while the ignition is off) with no real hesitation or stuttering. I'll need to pull and check the MPI relay next... fun-fun.

Background: This problem started after driving a buddy around all day in an area with rough roads and under-construction freeways that we hit fairly hard most times. On my way home that night, the CEL flickered on. I romped on it and it went away for a minute, then came back. This repeated for the rest of the ride home. The next day it came on about 30s after starting the car and would not go out again no matter what I did.

With that info in mind, I have to guess that either the MPI relay came loose from its plug, the MPI relay went dodgy, the ECU connector came loose from its plug (yeah, right), or possibly one of the new caps has a cold solder joint that finally fractured from the stress.
I have smelled a light burning electronics smell once or twice, but it seemed like it was coming from the area I was driving through, rather than something immediate in the car. Not the rotten-fish smell of a dying cap, but the melting-insulation-plastic smell.

Pulled and tested my MPI relay according to the Haynes, everything checked out perfect. Only 10V measured, but that's somewhat expected with the car not running, and is the same across the battery terminals.
Going to wait until I pick up a non-EPROM ECU from a local DSMer before I pull mine out, apply battery power to the fuel pump check connector, and use my multimeter to check for voltage on pin 13 to eliminate that and give in to the seemingly by-this-point foregone conclusion that it's a toasted ECU. :b No reason to pull the same stuff twice when the new ECU needs to go in, and the old one's already out.

 

[steve]

Only 10V measured, but that's somewhat expected with the car not running, and is the same across the battery terminals.

Has the car been down long. 10v at rest is pretty discharged and way too low for a battery that otherwise should have been charged lately.

Keep in ming that the MPI relay may look good with no current flowing but not so when there is 15 to 20 Amps passing through.

Steve

 

[Talesin]

I'm using an Odyssey PC925LMJT dry-cell battery, recently purchased (within the last six months). No downtime, the car still drives and runs reasonably well, so it's still serving as my DD (fortunately).
And I'm going based on Haynes, as far as testing.. don't have any way to throw 20A through it and check. I also borrowed an MPI relay from a local (not the one he runs on, a spare he grabbed when he thought HIS MPI relay had fried) and popped it in, the CEL came right back with no change. Being fair it *was* a junkyard part.

Hmm. Now I'm concerned that the battery may just have been deep-discharged, and for some reason the alternator isn't putting out enough to top it up again. Normal operating voltage (as I'd said) is usually hovering right around 12V, about five minutes after starting the car. Prior to that it normally hits 13V or a bit higher and sits there. Those are from what my Apex TT tells me, and I verified it accurate at idle with a multimeter across the battery posts. I'll look into borrowing a battery to check if the issue goes away.. and if it does, have some harsh words for the Odyssey rep, and the buddy that recommended them to me over an Optima. Though I do really like the total lack of off-gassing... no need for a sealed battery box.

 

[steve]

You'll note that the chart on the page you linked to shows the time to 10.02v. Thats the point at which they consider the battery discharged.

On a 1G you should see at least 13.4v when the engine is running from the alternator at the battery.

Do you have the fuel pump rewired?

Besides the DTC is there any other indication that something is wrong.

Steve

 

[Talesin]

Hmm. I'd missed that. And no, I see 12V at the battery when the car has been running... the Apex TT seems to be pretty accurate. As I said, usually about 13.2V when I first start the car in the morning, then it drops down to around 12.0-12.7V while normally driving. It can dip down to under 11V while in heavy stop-and-go traffic though.
Autozone 75A alternator, for the lifetime warranty given how quick ours tend to die. I'll have to check the alt directly.

The fuel pump is not rewired (I'll be doing it soon though). The battery has, however, been relocated to the trunk. 0 gauge wire used for the relocation.. I'll visually inspect the wire if that might be the cause, as well as check my circuit breaker.

Other indications are hard starting in the morning... I can crank it for up to 30s with no effect, but the next time I turn the key, it tends to fire right up.. I can hear the fuel pump spinning down as the starter stops, but it might only be getting switched on as the key moves away from 'ignition' for all I know. Don't have anyone handy to watch my underhood fuel pressure gauge.
Applying battery power to the fuel pump check connector under the hood turns on the pump and provides my set fuel pressure. Would it be safe/advisable to drive the car with that check connector hooked up? I'd been toying with the idea of a 'fuel system prime' toggle switch, since mine will not hold the OEM 21psi while the car is off. Would make morning starts much easier.

 

[Talesin]

New alternator installed, Apex says 13.5V, though a digital multimeter said 13V when the Apex was reading out at 12.4V so I guess it can't be trusted at this point, other than for in-system readings, nor can my handheld meter. :b

Charging properly, load test shows that it's in good condition (can handle both fans, high beams, cabin fan, dome light, brake lights, reverse lights, stereo up high... it drops way down then creeps back up to 13.4V on the Apex, assuming most of the drop is coming from the mis-charged battery at this point).

Only things left to do are hook up the fuel pump check connector and look for power on pin 13, and swap in a new ECU if there's power there. :b Fun fun, and just what I didn't need.

 

[Talesin]

Update. Picked up a new multimeter, ripped apart my interior for ease of working.

I probed the wire going to pin 13 directly (sharp test probes are nice!) less than 2mm from the plug and it had 13V even, with the CEL still on. Testing at the front under the hood on the fuel pump check connector showed 13V as well. I've borrowed an ECU from a friend and tossed it in, and the CEL came on with the same trouble code (I'm starting to wonder if my logger cable might not be bad!).

At this point, I can only wonder if pin 13 gets power when the fuel pump check connector is hooked to battery voltage, to test at the pins' contact surfaces directly for possible damage (?), or potentially assume that the other ECU was damaged in the same way as my own... somehow... and locate yet another ECU to toss in and test on.

This is becoming bloody ridiculous.

 

[steve]

At this point, I can only wonder if pin 13 gets power when the fuel pump check connector is hooked to battery voltage

I don't get your question. Yes pin 13 see what you connect to the check connector so yes you could measure with the ECU cover off and see if the connector was good.

Have you tried forcing another CEL condition and see if it's reported correctly?

Steve

 

[Talesin]

That was the thought, so I wouldn't have to worry about accidentally shorting a running IC. But if the line was isolated from the ECU, just providing direct power to the fuel pump (without needing the MPI relay actuated), then the ECU might not see it at all. But if that pin gets power with the check connector hooked up, then I can test it from the end that plugs into the ECU.

I'll have to pop off the cover, apply power to the pin and trace it through the IC if it looks like power should still be coming through.

I'll pop off the MAS plug and check to make sure it comes up. Just weird that the second ECU threw a CEL as well, if it's the ECU that's the problem. :b

 

[Talesin]

All right, I've since replaced my ECU with a freshly rebuilt unit (mostly so I can send my current one off to ECMTuning for rebuilding and a socket service) and it STILL throws the fuel pump CEL. Pin 13 has 13V available, tested right before the ECU to a chassis ground.
Does anyone have an IC diagram for the ECU, or know where that specific circuit path eventually goes to ground? That's the only thing I can think of at this point. This is driving me absolutely bugnuts.