Jump to content
Ultimate Subaru Message Board

Flashing CEL and P0328 on 99 legacy L 2.2L


Recommended Posts

Hey, so I have a 2.2L engine from a 1999 legacy L swapped into my 1984 vanagon.

 

But from day one, I get a flashing CEL after about 5 secs after the engine starts. On ecuexplorer  it show a P0328 code. So I checked out the knock sensor/wiring and its all good, 2.8V when running. Also tried another knock sensor that's in spec and another ecu with no success. Reset the ecu and within 5 secs the CEL is flashing and P0328 is back.

 

My friends scanner also picked up a P0302 code, cylinder 2 misfire. So checked the leakdown on the engine and changed the plugs. wires are new....

 

Engine idles smooth, but on ecuexplorer I can see its pulling most of the timing out when accelerating. Which makes sense since it should be in limp mode.

 

I changed the timing belt before putting the engine in the van, but I double checked and believe its good.

 

Any Ideas what would make a Flashing CEL come in so fast?? and no matter what I do P0328 keeps coming in???

 

Thoughts? idea?

Link to comment
Share on other sites

Do you have 2.8 V at the sensor when its plugged in? (Back-probe the connector)

 

Flashing CEL indicates a constant misfire, but it SHOULD be setting a code for which cylinder is misfiring.

 

Kinda seems like you may have an ECU issue.

-5v off the ecu and 2.8v when its plugged in.

-1999 models seem to have a hard time communicating with a lot of obd2 readers... found that out after I got this one as a donor. have to try my friends expensive scanner again

-I got a second ecu from pnp and it does the same thing.... so either its a common issue with the ecus or I have a funky wiring issue i'm thinking. All the grounding looks good, every voltage I've checked looks good...

 

-all the plugs looked abit lean too when I changed them. maybe I should check fuel pressure. its a new pump, but maybe the regulator is wonky.

Edited by Vanagoon
Link to comment
Share on other sites

I would be interested to see what the scanner says.

 

Could be running too lean, pinging and driving the knock sensor crazy, which makes it pull timing. Lean also causes misfires at idle. If it pings at idle that can make it set codes for the knock sensor. What kind of fuel pump are you running? Have you checked fuel pressure to see if its in spec? Should have 35psi at idle. About 45 if you pull the vacuum hose off the FPR.

 

Entirely possible you just have the cam timing a tooth off, might double double check that again for gits and shiggles, even if you already have.

Link to comment
Share on other sites

I have 3 knock sensors to choose from, all in spec and all the same issue.  100% sure its not the circuit or sensor.

 

adjusted the valves and changed the valve cover gaskets/seals. Then changed the plugs, no oil to be found.

 

The pump is the stock bosch vanagon pump, which worked flawless with the old engine.

 

I'll be checking the timing belt and fuel pressure asap.

 

Thanks for the ideas so far.

Link to comment
Share on other sites

well timing belt is definitely good, was almost hoping it was off. But that would of been to easy....

 

Fuel pressure looks good too, wasn't the best gauge but im sure its good...

 

when ecuexplorer is talking to the ecu, I can see that the knock sensor isn't pulling timing. but the ecu must be in limp mode, because there is very little timing when under load.

 

still have to check the harness connectors, ill do that tomorrow.

Link to comment
Share on other sites

Timing can drop back significantly under load. Back to around 10-12 degrees in some cases.

A while ago Presslab (I think) posted a graph showing base timing and another showing maximum possible advance beyond base for the whole RPM range. It was for a phase 1 (92 or 94 IIRC) Ej22, but it would be similar for the 99. At least similar enough to give you a general idea of where ignition timing should be.

 

 

Do you have a vacuum guage?

 

Does your setup have a MAP sensor? Could look at MAP values to determine vaccum.

Link to comment
Share on other sites

carbon build-up in the cylinders could lead to knock - but if the sensor isn't pulling timing, it isn't knocking.

 

one experiment to try, clear the computer with a battery disconnect after filling with high octane. maybe try a coupla tankfuls. If the problem seems better, may be a 'real' knock issue.

 

but from what you say - ECU seems suspect doesn't it? Or the connector to it w'ever

Link to comment
Share on other sites

so almost no resistance on any of the ecu grounds to the engine, then I turn the ign on and I get close to 40ohm of resistance!

 

then I turn off the ignition and the resistance stays at 40ohms for abit before dropping down again.

 

bad connection thats getting warm and increasing resistance?

Link to comment
Share on other sites

well, I don't know what instrument you're using but, an ohmmeter supplies its own current and isn't used on a powered circuit. Sounds like the meter is just confused with the power in the line and the slight drop across the connection? kinda weird as there shouldn't be a lot of current with the ignition just at ON. I'd be tempted to do all my resistance testing with the neg batt connection removed.

 

Also, try reversing the leads on your testing. Sometimes corrosion can act like a diode. not commonly, but possible.

Edited by 1 Lucky Texan
Link to comment
Share on other sites

Even with the key On there should be little or no resistance on the ground side because there should be no potential (voltage) on the ground.

Check for voltage on the ground side. More than 0.1V means you have a bad ground.

Preferably you want your meter grounded to the battery negative post, so make a jumper wire to attach your meters black lead to the battery if it isn't long enough to reach.

 

The ECU originally grounds through the wiring harness to the intake manifold. Not sure if you kept that when you made your harness, but if so make sure its clean and tight.

Also make sure the engine is well grounded to the frame. You might even add a ground wire from the manifold to the battery.

Link to comment
Share on other sites

Even with the key On there should be little or no resistance on the ground side because there should be no potential (voltage) on the ground.

Check for voltage on the ground side. More than 0.1V means you have a bad ground.

Preferably you want your meter grounded to the battery negative post, so make a jumper wire to attach your meters black lead to the battery if it isn't long enough to reach.

 

The ECU originally grounds through the wiring harness to the intake manifold. Not sure if you kept that when you made your harness, but if so make sure its clean and tight.

Also make sure the engine is well grounded to the frame. You might even add a ground wire from the manifold to the battery.

I'm going to have to check for voltage again, there is a bad connection or something in the harnass. The resistance is from the ecu to the ground on the engine. Its in the harnass some how.

 

The engine harnass is unchanged and the ecu to engine connectors have had very few cuts. Its mostly removing the unused stuff and just hooking the power, ign, and those items back into it.

 

I'll dig into it again tonight.

 

and no the ecu case isnt grounded, but I did try using a jumper and grounding it once and it never made a difference.

 

T

Edited by Vanagoon
Link to comment
Share on other sites

The reason you can't take a resistance reading when power is applied to a circuit is because the meter supplies power to it in order to make the test. Adding additional current in the circuit will throw the reading off. Checking for voltage across a suspected bad ground connection while there is a load on the circuit is the best way to find a ground problem.

 

I suggest you try testing the circuit operation for the knock sensor. Remove the connector to the sensor and install a 510k ohm 1/2 watt resistor on the connector end in place of the sensor. If there is a singe wire going to the sensor then you will have to ground one end of the resistor. See if the CEL light goes out then. If it doesn't then the ECU may have an internal problem with that circuit, though having the correct bias voltage is encouraging.

Link to comment
Share on other sites

Did you use the venting stuff from back of car ? on 98 they ran power and signal to rear of car for gas tank venting control must be in wire harness or defaults was 4 or six wires to rear of car fuel temp senser / vent valve / vacume signal senser

 

Yeah....mine runs fine without that stuff.  Throws codes for Fuel temp/pressure and a "Mnfctr. Cntrld. Emmisions" whatever that means.

 

Just a though......Did you plug in the "test" connectors?  Green plug from ECU?  it should be disconnected or else the ECU will be in a constant test cycle.  That could be the flashing light.

Link to comment
Share on other sites

I have a little board that gives the proper voltages for the unused circuit so I don't get codes.

 

I went to pnp and got another ecu to engine harnass and swapped it in. No more limp mode or p0328 cel!

 

when I get a chance i'll be ripping the old harnass apart and see if I can spot something. 

 

thanks for the ideas and help

Edited by Vanagoon
Link to comment
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now
 Share

×
×
  • Create New...