No Start on a Subaru 99 Legacy SUS

[RetiredGearhead]

I swapped a 97(?) JDM 2.5L in for a 99 2.5L  recently. The 99 was driven about 60 miles with no water in it due to a blown hose, nuf said, it wasn't me. I installed the "new" older 2.5 with new timing belt blah blah blah. Timing set up nicely. Even though the JDM was injected, we transferred the inake and other appropriate items to the "new" block including drilling and tappping out the EGR hole in the head. I don't want to go into the 4 bolt vs 8 bolt "bell housing" issue as it doesn't seem to be an issue.

 

Where my problem is at, is that I can get spark to the cylinders #1 and #2 but get no spark to #3 and #4. The coil(s) check out, the ignitor checks out and there doesn't seem to be a signal from the ECU to fire the #3/#4 ignitor circuit. I have checked the wiring between the ECU and the Ignitor and it appears OK. The connector at the ECU is clean and undamaged. So I assumed it was the ECU as I have read that it is the main source of "no start" conditions. I put in a used ECU. I got exactly the same deal, no signal to #3/#4 circuit in the ignitor. I find it hard to believe that a second (albeit used) ECU would have the same problem ........ or does it?

 

What would cause the ECU to send signals to fire the #1/#2 circuit but not the #3/#4?  I am baffled but it is probably something simple. Any help? I am new to this site and this is my first request.

 

As far as my background. I have degrees in Electronics and Mechanical Engineering so I am no ametuer at troubleshooting and I am pretty much down to earth in that degrees don't mean spit sometimes. I usually find the simple things are the ones that trip us up.

[presslab]

I'm going to say that the cam sprocket (trigger wheel) for the cam position sensor is different.

[RetiredGearhead]

Though I have checked this between the two sprockets off their respecive engines, I wil check the air gap between all the magnets and the sensor. I suppose it is possible that a magnet representing the second set of cylinders may have been damaged somehow.

[presslab]

The sprocket doesn't have magnets, the sensor has a back bias magnet; this is why you see old sensors will metal filing stuck to them.

 

Might be just a flaky cam sensor. Hook a scope to it and check it out.

[RetiredGearhead]

I have swapped the cam sensor with no change in symptoms a week or so ago. I did not swap the the sprocket however as they seemed identical. The Subaru is at my shop and I will investigate further tomorrow morning. Thank you very much for your information. I have been stumped so far and it certainly could be associated with the sprocket/sensor interface.I do have an oscilliscope and can check that out. I have a picture of what the waveform should look like.

[idosubaru]

you have to check the ***back*** of the cam sprocket, did you?  the fronts will look identical, the backs will have different trigger marks, if that's the issue.

 

no spark condition can also be:
1. coil pack

2. igniter

[RetiredGearhead]

Thanks to all who have given suggestions, I really appreciate the help. I have been working on various cars and motorcycles for over 50 years and this really has me going a little nuts.

 

A little more background. This car sat for several years with the blown headgasket. It started right up with a little boost to the battery. But the fuel pump eventually gave out from our friendly ethanol gas corrosion  fun. So I installed a new fuel pump.  It is working fine as I ran it to empty the fuel tank and develops good pressure. The tank now has about 10 gallons of fresh non-ethanol gas which I keep on hand for my old Honda Goldwing.

 

After pulling the engine to look at a rebuild I noticed 2 egg shaped cylinders which meant a full rebuild not just head gaskets so I bought a used JDM engine which is what I am trying to make run now. Point of interest: I had moved the car around with the original engine on my property several times before I pulled the engine. I noticed that each time I started it up and moved it, the engine was a little harder to start, doing something similar to what it is doing now, but it always started and seemed to idle and rev smoothly. Maybe it means nothing, I don't know.

 

If I haven't already stated it, the current engine turns over briskly and fires #1 and #2 and outruns the starter sometimes, but just won't fully run because I get no spark to the other 2 cylinders. Fuel is clearly getting to the injectors. The problem is no attempt to fire the second half of the coil.

 

Now to the suggestions:

 

I have carefully examined both sprockets (from the original engine and the swapped in one) and they appear to be identical in every way. Though I have no precision tools to measure the angles of the magnets (and they are magnets on both sprockets) with respect to the alignment pin between the sprocket and the cam, they appear identical.

 

I do not have an automotive scope so I can not take a look at the waveform coming out of the sensor, however, since I get a steady signal out of the ECU for cylinders 1 & 2, it appears that the info from the cam sprocket sensor is getting to the ECU.

 

I have checked the sensor and wiring and all appears well. If I pull the connector off the cam sensor I get a Code P0340. When I replace the connector, clear the code and crank again the code does not re-enter. The same is true of the crank sensor which returns a P0335 code when disconnected.

 

So, in summary: Both sides of the coil check out OK as I can make them both fire by hooking the Cyl 1/2 lead from the ECU to the ignitor to the Cyl 3/4 ignitor input. This also verifies that the ignitor is working on both sides. (I did this test with the plugs removed and grounded so as not to fire #3 and #4 ad in appropriate times with any residual gas in the cylinders).

 

The wire from the terminal pin on the ECU to the ignitor checks out in a resistance test ok and is not shorted to ground (or power for that matter). A close inspection of the terminals on the connector and the ECU appear in good working order.

 

I have swapped the ECU with a used one with the same result. I suppose it is possible that the used one has the same failure, whatever that is. A new one runs between $700 and $1000. I can not seem to find any place to test the ECU. The Subaru shop manager all but laughed at the request saying that none of the systems can be checked out of the car and it must be brought in to troubleshoot it. Sounds like a new car would end up a cheaper deal.

 

Since I don't know how the firmware in the ECU works and it shows no codes when the sensors are connected, I can't figure out why the ECU would fire #1 and #2 but won't attempt fire #3 and #4.

 

As an after thought, I should mention that I have not put the radiator back in yet as I was concerned that I may have to re-do the timing belt, so there is no coolant in any part of the engine. Is it possible that it requires some coolant sensor to detect fliud to run? And if so, why would it fire any spark plugs at all?  The air sensors are hooked up and everything else that I can think of (or find) is connected.

 

(looks like I wrote a novel)

[heartless]

i would begin by verifying the timing is set correctly - even one tooth off can cause a no start situation.

[RetiredGearhead]

The timing is correct. All markers are lined up perfectly. The problem is that there is no signal sent from the ECU to the ignitor for one half of the coil and that is from two separate ECU's. I may have to borrow an automotive scope or a digital oscilliscope to get a clear look at the cam position sensor waveform output.

[johnceggleston]

any CEL trouble codes?

 

mis-fire / no fire on 3 & 4 is a coil issue,

or maybe an injector issue.

 

i have no idea what an injector problem might be.

 

 

So I assumed it was the ECU as I have read that it is the main source of "no start" conditions.

i do not know where you read that but it is not true.

hands down the most common cause of a no start after major work is timing.

but since you are confident of the timing,

i would look else where.

besides, a 1&3 , or a 2&4 no fire would be a timing issue.

not a 3&4.

 

long before i swapped an ecu,

i would have swapped

coil,

igniter,

wires,

plugs,

whatever.

 

again, i do not know enough to give a good answer,

but no spark on 3 & 4 is a coil / igniter issue.

 

just because they appear to fire correctly when tricked,

does not mean they are firing correctly when the engine is running.

they could be weak.

 

imho.

Edited July 31, 2014 by johnceggleston

[presslab]

Ah I didn't think about the plastic pulleys, those must have a magnet in addition to the one in the sensor. I know the metal pulleys don't have a magnet.

 

Anyway hope you figure it out.

[Dj7291993]

Did you do continuity tests on the wire?  If it fires when given a signal from the other one, then your coils should be fine.  Have you back probed the signal wire at the connector, to see if it's sending from the ECU?  Also, I'd double check that you have no shorts (to ground or power), and check your ground from block to bat, bat to body, & block to body.  Bad grounds can cause some of the weirdest codes.  Easy way to check it do a volt drop between the two points while cranking.

[RetiredGearhead]

Thanks for the thoughts USMB is life!. I have done the continuity tests. I find no signal at the ECU on the #3/#4 wire but srtong signal on the #1/#2 wire at the ECU connector. The wire at the input of the ignitor is not shorted to either power or ground.

 

However, it just occurred to me that somewhere in the harness it may be shorted to another wire which could negate the positive pulse coming out of the ECU. Tomorrow I will cut that wire an inch or so from the connector and check for the pulse at the wire stub. It can always be spliced back in either case just as it have isolated it at the ignitor connector. Voltage drop test is a good idea. Wires that are frayed somewhere along the line can read 0 ohms across their length but still not be able to carry any significant current. It is a bit rare in the electronics world I worked in but not entirely unheard of. The automotive world is a much harsher environment from temperature, vibration and corrosive elements standpoint.

[Gloyale]

I would look very closely at the connector pins inside the Igniter
connector.....greasy dirt can get easily shoved in there and will cause a
poor connection.

 

It's gotta be in the wire from ECU to Igniter.  It's doubtful the ECU failed....let alone 2 of them.

 

Try running a jumper to bypass.  Like you suggested cut an inch or two up from the ECU and check for the pulse.

 

 

 

Also do the Crank pulleys match?  they can be different too.

 

This is why I always pull covers on JDM engines to see what setup is in there.

Edited July 31, 2014 by Gloyale

[RetiredGearhead]

I cut the wire at the ECU connector and still had no signal coming out of the wire going to the ignitor for cyls 3 and 4.

Further testing, if I remove either one of the crank or sprocket sensor connectors, I get no signal to cyls 1 and 2. This

tells me that those sensors appear to be working fine back the the input of the ECU. Back to the Service Manual I 

guess....... 

[Gloyale]

Did you verify the correct crank sprocket?

 

there are 2 kinds.....one with 6 tabs......not all evenly spaced......

 

and another that has like 30 tabs and on;y one gap.

 

Verify that

[Fairtax4me]

X2 Check the reluctor teeth on the crank sprocket. One is probably chipped off and the ECU is not getting the proper signal for 3/4 to fire.

[RetiredGearhead]

I do know that the crank sprocket doe not have more than 6 or 7 tabs. I don't remember exactly how many. (at this point I get excited to find almost any new piece of information) Since  both cam sprockets have 7 magnets;[ 3 in a group, 1 at about 90 degrees, then 2 at about 180 degrees then 1 at about 270 degrees] I would expect that the crank sprocket would have exactly the same amount ot tabs.

 

My understanding is that the ECU uses the CAM sprocket as the primary position indicator and falls back to the crank sensor in "limp" mode if the cam sensor fails. Maybe this only works if the emgine has been running and some operational data is present. That why the experiment of removing the sensor connectors one at a time yeilded a no spark attempt for the #1/#2 side of the coil. I would have thought that either one would still attempt a spark.

[Fairtax4me]

Cam sprocket has 7 tabs to differentiate cam position at TDC.

Crank sprocket has only 6 tabs.

 

ECU needs input from both sensors in order to time both spark and injectoin properly. Failure of either sensor will give a no-run condition.

I haven't tested this myself, but its generally believed that the crank sensor is responsible for spark, and the cam sensor triggers fuel.

[RetiredGearhead]

Have been out of town a few days. Back to the grind.

 

I have verified that the CAM Sprocket and the Crank reluctor ring are identical on both engines. Cam sprocket 7 magnets; crank 6 tabs.

 

The ECU must at least be seeing part of the crank and cam sensors or it wouldn't fire the #1/#2 circuit. It may not be able to see one or both reqired to fire #3/#4 if the signal is weaker.

 

I am trying to find a storage scope that will give me an image of the sensor inputs at the ECU connector to ensure that the signals for ALL sensors at the ECU are healthy and consistent. This may take a few days to locate something I can borrow or can afford and acquire. Thanks for all the input.

[Gloyale]

I've got an Oscilloscope down here in Corvallis.

 

Too bad you can't drive it down here eh?  If you get a trailer I'd be happy to ook it over.

 

Hell.....I've got an 96 harness that will run that engine standalone to figure if it's in the enigne or in the Harness/ecu.