49 replies to this topic

[jimshih]

Hi All,

I am working on a VW van with a 2.2L Subaru engine. This is a distributorless system.
The engine won't start and I see the condition of Spark #1 and #2 . No spark #3 and #4.
There are three wires going to the coil with 12V at the center. If I swap the other two then the condition becomes NO Spark #1 and #2 and spark #3 and #4.
From that I am guessing that the coil pack is good.

I am looking to test the ignitor and ECU but if anybody knows the cause of this problem please let me know, thanks.

[Cougar]

If the two wires go back to the ECU then there may be a problem with it. The crank angle sensor may be a problem also. Are there any trouble codes stored?

[Legacy777]

It does sound like the coil pack is good.

I'd say the ignitor or ECU is to blame. Test the ignitor and then the ECU.

If you don't have the factory manual, this will be of some help

http://www.main.experiencetherave.com:8080/subaru_manual_scans/1992_FSM_(Engine_%26_Electrical)/

[Cougar]

After looking over Joshs' thoughtfully provided diagrams I noticed there is not only a crank angle sensor but also a cam angle sensor. Verrry interesting; as Arte Johnson would say. Sometimes the magnet of the sensor can get broken off and this will prevent the signal for firing the ignition to happen. Before doing anything I would check the ECU for some trouble codes. My hunch is there is a problem with a sensor. Hopefully there is a code indicating that.

[Legacy777]

Glen,

The only reason I don't think it's either the crank or cam sensor is because he is getting spark at one bank of cylinders. If the cam or crank sensor was on the fritz, he probably wouldn't have any spark at all.

I'd still probably recommend checking the ECU for any codes it might have stored in there.

[Cougar]

Hi Josh,

You may be correct but my reasoning on this is; if two of the four CAS sensor magnets were damaged wouldn't that cause a problem like this? It seems to me that without those trigger signals you have no signal to the ECU to fire the ignition pulses. Or is my logic out of whack somewhere?

[Legacy777]

The The way the crank & cam sensors work is they emit a small electromagnetic field/pulse or whatever. The crank gear and cam pulley has marks or breaks in the normal pattern of the gear/pulley that interupt that electromagnetic field to let the ECU know what the timing is and when to fire.

I just really don't think you would get only one bank firing if you had a bad cam or crank sensor.

[Nug]

I'm leaning toward the ignitor also. I had one go bad in my bug, two cylinders just flat out died. It still had enough power to go 70mph.

[Cougar]

I'm with you on the pulses Josh but I think these are generated by four magnets on a gear for the CAS, one for each cylinder naturally. I have heard of these magnets being damaged and causing problems.

Snotrocket, you may be correct, I really don't know as much as I would like to on how the ignition really works. The ignitor is a little mystery to me. Anybody know the fine details on this to explain why one bank would be disabled? He did swap the leads to the coil pack and the opposite plugs fired.

[jimshih]

Here is a update of the problem:

I have tested the ignitor and that tested to be ok.
Two wires from the ECU going to the ignitor. One of them has no signal going to the ignitor. I tested by crossing the wires at the ignitor.

Now I need to research if the ECU has two sperate inputs for these two outputs or just one. If there is only one then the ECU is bad. If there are two seperate inputs for them then the ECU is good and a sensor or something else are bad.
Any help on this??
I have not pulled the code. Will do this now. Just need to figure out how.

[oddcomp]

not having new gen stuff.. but having some reasearch on dis systems for my old school suby i also think sensor problem of some kind
not knowing fully how the coil drivers on the 2.2 work .. are they seperate from the ecu?

a old needle style volt meter at various points on the ignition path from the ecu to the coil driver<igniter?> will tell the story i should think

maybe easier to see if your working solo would be a 12volt l.e.d light from radio shack on some long wires

[jimshih]

Ok I checked for codes:


I am reading the green led on the ECU as this is a engine swap.

with nothing connected:

it blinks 4 4 4 4 4 4 ..........

connect the two green:

4 4 4 4 4 4........but now the fuel pump relay click on and off and I can gear the fuel flowing sound.

Connect the two blacks:

I see 12 12 12 12 12 12 .....

Check that and it come up as starter switch open or short circuit so that maybe from the swap wiring and should not affect the ignition system.

Now still search??

[Nug]

Code 12 Starter switch-
Pin 10 on connector B56 of the ECU should connect to the start terminal of the ignition switch, or the starter solenoid. It needs to have 12v at it when the engine is cranking. Not while it's running, and not when it's off, just cranking.

Next, ECU outputs to the ignitior:
Pins 9 and 10 on connector F47 both provide a signal to the ignitor. Make sure pin 15 is connected to the ignitor as well, and is ALSO grounded. I had a bad connection right here.

The flowchart from the FSM suggests that if you do not have a signal from both terminals 9 and 10 from connector F47, then either there is a poor connection between the ecu and ignitor, or the ecu is faulty.

[Nug]

When wiring mine up, i used an analog voltmeter set at a very low voltage to check various sensors. I wasn't interested in numbers, but movement of the needle. A pulsating needle at the crank, cam sensors, and to and from the ecu/ignitor proved all of these components were functioning.

[Cougar]

I think the trouble he is having is because either pin 9 or 10 has no signal output. He tried reversing the connections and the other bank fired then.

[Legacy777]

There are no magnets on the gear or the pulley.

Here is a picture of the crank gear. You can see the little tabs I was talking about. Unfortunately I didn't get any good pictures of the back of the cam gear. But it has a similar set of irregularly patterned out knotches.

http://www.main.expe...lt/DCP_2716.JPG


It'd be worth checking the sensors to verify they're actually outputting some sort of signal, but as others have mentioned, definitely check your wiring and grounds.

Also.....did you put the ECU in the engine bay? That's not really the best place for it, especially since it's not water tight.

[Cougar]

Thanks for the pictures Josh, they are a big help. I guess my confusion was about the tabs that are on the outside edge of the small gear. It is these tabs I was thinking of that can get broken. These are what the sensor uses to check the timing, the way I understand it anyway. Is that correct?

[Legacy777]

Yes those tabs can get broken, and basically cause the motor not to work.

As I mentioned the sensor puts out a small electromagnet field and those tabs break that field. That pulse get's sent to the ECU. The ECU is programmed with the fact that the tabs are irregularly placed around the gear. Because of this, it knows the exact position of the crankshaft. Same is true for the camshaft.

[Cougar]

Thanks for the clarification on this and your help Josh. I was a little confused on this alright. Hopefully some of this side discussion will help Jimshih also.

Another question I have since we are on the subject is, it seems to me that this timing cog is for the cams; or is it really for the crank? I would assume the crank sensor would be located on the flywheel. I am going to have to actually work on one of these newer style engines someday so I can learn a little more. So far I only have touched the EA82T. It really helps to have hands-on experience with these things and I know you've been there.

[Legacy777]

No problem Glen

The gear in the picture is for the crank. The bigger pullies are for the cams.

The crank sensor is located on the front of the engine just above the oil pump.

Take a look at the pictures I took when I did my injectors and timing belt. I labeled a lot of the sensors when doing the injectors. The timing belt, I don't think I labeled the pictures, but you can get an idea how everything goes together.

http://www.main.expe...ages/timingbelt
http://www.main.expe...mages/injectors

[oddcomp]

this actually explains how teh dis system i adapter to my 87 turbp wagon works
but the suby dis setup works teh same it just has a cam and crank sensor

but the operation is still basically the same
http://www.bgsoflex....is_summary.html

[jimshih]

http://www.bgsoflex....is_summary.html

Looked over the Ford EDIS 4 and tried to match that to the 2.2 Subaru.
Here is what I see different:

EDIS: crank signal goes direct to the EDIS unit and EDIS has driver inside to control coil fire. EDIS also sends PIP signal to ECU to adjust for timing advance and ECU outputs SAW signal back to EDIS after it determined timing adjustment needed.

Subaru: Crank and cam sensor signal are feed to ECU not to ignitor(EDIS).
After ECU determines the necssary timing it send signal to the ignitor which there are two wires. one for each bank. So on the Subaru system the ignitor performs different function then the Ford EDIS unit. My unit has no signal from one wire and is causing one bank not to fire. I am suspecting the ECU defect at the point and looking to get a replacement to test.

Please correct me on the above if I am wrong, thanks.

[Legacy777]

No, it sounds like you're on the right track. Did you test the signal at the ECU pins or at the ignitor?

If you tested at the ignitor, you may want to test at the ECU to verify wiring is not an issue.

[jimshih]

No, it sounds like you're on the right track. Did you test the signal at the ECU pins or at the ignitor?

If you tested at the ignitor, you may want to test at the ECU to verify wiring is not an issue.

I cut the two wires going to the ignitor and use jumper wires to connect them.

I will test signal at the ECU pins since that would be easy test also just to get any wiring issue out of the way.

[Cougar]

Jimshih,

You might check the components in the ECU if you do electronic repair and the problem turns out to be with the ECU. I would suspect that since the two output leads are the same then the circuits would be a mirror image. There may be a bad driver transistor stage at or near the output connections. I have fixed some ECU's with this kind of problem before. I still wonder about the crank sensor tabs though as the cause of this problem.

You might try doing a resistance test on the ECU leads while connected to the ignitor to see if they read the same. Doing a diode check may help also. If the resistances are the same then the problem is elsewhere I would say or at the input to the ECU.