OB99W

I hope what happened didn't involve someone getting hurt. However, I'm not sure that disabling the ABS is a good idea. Food for thought: 1) Very few drivers have the skills to do a better job of stopping a vehicle without ABS than with it, under certain road conditions. If you're one of those who can, congratulations. However... 2) How can you anticipate when having or not having ABS would be better? You might decide that you can do outperform the ABS under the conditions you're observing, but what if you then hit a spot where the ABS being active would be an improvement? Would you have the time and presence of mind to switch the system on? 3) What would happen if you had an accident under road conditions where the police and/or insurance company felt ABS would have prevented or minimized damage or injury, and it was found that there was a modification to switch off the system (whether or not you claimed the ABS was active at the time)?

I wasn't being "tuff" -- just pointing out that a '97 Subaru could have an "Outback" badge on it, not necessarily be a Legacy, and have a 2.2 (from the factory). Blame Subaru for the possible confusion.

The 1997 Outback Sport (Impreza) did.

How did you identify which connector was which so that you knew they were swapped?

You obviously aren't an idiot; you took on a fairly involved task that most people would never even attempt, and overcame several impediments along the way. If there are any visitors from a distant planet, please raise your hand now -- okay, the rest of us will have to admit to being human, and making occasional mistakes. It's great that you've had the fortitude to stick with this and are continuing to make progress. Congratulations on being one step closer to resolving things. Hopefully, you're near the end and complete success is near. I don't see any reason to apologize. Another human characteristic is pain avoidance; if it hurt anyone that much to be involved here, they probably wouldn't do it. In fact, I'd bet that most USMB members actually enjoy participating (or they have a slightly masochistic streak ).

For anyone wanting to see the above quoted info in original context, see this thread: http://www.ultimatesubaru.org/forum/showthread.php?t=82540 .

At no time did I rule out a mechanical problem, nor only consider an electrical one. Modern vehicles typically have several sections that include both electronic and mechanical components; anyone who thinks a problem in one of those areas can only be attributed to the mechanical or electronic aspect is setting himself up for failure. The scope made it possible to verify that a suspected timing problem actually existed; I never said it would point directly to the cause. On the other hand, if the scope had indicated no timing discrepancy, that result could have helped eliminate certain possibilities and suggested the focus should be on others. Sometimes mechanical work leads to electrical problems; see: http://www.ultimatesubaru.org/forum/showthread.php?t=82614 As to your six-point approach to oscilloscope usage, I hope you'll forgive me if I won't "play". After using sophisticated test equipment almost daily for more years than you've been alive, I've developed a pretty good sense of when to follow convention and when to be unorthodox, and I've successfully completed repairs that certain techs have been stymied by. Unfortunately, doing diagnosis by "remote control" such as on USMB sometimes has its limitations. Finally, no one should forget that, after all, the ECU is just a dedicated computer, and like all computers, GIGO applies to it. See http://en.wikipedia.org/wiki/Garbage_in,_garbage_out . Connectors that didn't get plugged in fully (or at all), damaged or mismounted sensors, swapped plugs, etc., could each play a role in causing the ECU to spew "garbage" just as much as a mechanical problem can.

Yes, injector timing has been an issue in a few threads recently. Thanks to 91Loyale, who evidently trusted me when he read my previous post (#11) in this thread, and apparently benefited from it. See: http://www.ultimatesubaru.org/forum/showthread.php?t=82614

You're welcome -- glad to help.

Sublute, I'm sorry that you're having difficulty with the procedure I suggested; sometimes things that I do on a regular basis aren't easy to have others accomplish. It did seem we were making progress, and that there appears to be an electrical timing issue between injector pulse and spark. That can be due to the ECU not getting proper sensor data (among other things), which could be caused by something as simple as a poor connection or improper mounting/spacing of a sensor or what it's sensing. The ECU may then go into a default mode, "assuming" parameters like timing. You should probably recheck the contacts at any and all connections that were removed or disturbed during the engine work, as well as relative positioning of sensors or possible damage to them (tips, for example). Since you're frustrated over my approach, if you don't mind I'm going to extract myself from this. I hope that nothing was damaged by the "slip"; I can assure you that using 1k-ohm resistors as loads and monitoring the pulses as I described wouldn't in themselves cause a problem -- I've done that many times without difficulty. Perhaps you should just reconnect the injector and coil wiring, and try suggestions offered by others. A set of "noid lights" may be useful to verify injector "signal". I wish you best of luck in rapidly resolving the problem. By all means please help Sublute resolve the problem. If you've followed the situation from the beginning, you know there were numerous issues that were dealt with during the work, making diagnosis of the no-start more difficult than usual. I leave it in your capable hands, and those of others who may be able to direct him with good result. Note to Skip: Is this where I insert some self-deprecating comment?

It sounds like you're describing a windshield side reveal molding. Go to a site of one of the online dealers that use TradeMotion and "drill down" in the body/collision parts; it should show up there.

Thanks for the front panel photo. Assuming everything on this scope is functional, I can probably talk you through setting it up for our puposes if you aren't successful yourself. First thing I'd do (assuming you're not using EXTernal TRIGgering) is put the TRIGGERING SOURCE switch in one of the INT positions.

Different scopes have varying functionality (delayed sweep, adjustable delay time, horizontal multiplier, adjustable trigger level, etc.). I can't readily advise without knowing the scope. Seeing a train of more than one pulse would be very useful; if practical to do, a time exposure might help. You could try switching the triggering to whichever channel you're not currently triggering from. You could also try adjusting the sweep variable control, rather than the timebase switch, or experiment with both. It may even be that the scope isn't functioning correctly. Perhaps if you gave the scope make/model, I might get a better idea. Yes, and that seems to verify what was seen and anticipated before -- namely that the spark (which would happen when the igniter waveform rises, if the coil were attached) is occuring at the start of the opening of the injector, not after. It certainly seems to be a timing problem.

That brings up the good point that just swapping the ECU might cause issues if a paired TCU isn't installed at the same time. "Let it be" if it works okay.

Sometimes they can be shortened just a bit. For example: http://www.autotrader.com/fyc/vdp.jsp?car_id=212957233 Or, there's http://tinyurl.com/ .

No problem, but I want to clarify that I didn't say the above, although the way it was quoted might make it look that way. However, I did say something very similar in post #63, when the issue was a bit less defined. Yes, the injection-spark relationship seems wrong, which is why we're investigating it further. If we can get a clean scope shot, I'm hoping not only to verify the order, but perhaps the actual time difference; that could be even more revealing.

What vertical sensitivity do you have that channel set for? It should be the same as what the channel the injector is connected to is at. With the resistor between 2 and 1 (or 3), there should be battery voltage on 1 (3) until the igniter pulls that pin near ground (beginning of dwell period). At the end of the dwell, the pin should go high again, and at that point the spark would normally occur.

I think the operative phrase here is "It runs well". "Don't fix what ain't broke" also comes to mind. Since you mentioned efficiency, how's the gas mileage? If it's decent, and the engine starts easily cold/hot, idles well, accelerates without hesitation, has good power, etc., I'd leave the ECU alone.

Previous use of a timing light apparently indicated ignition advance of 20-30 degrees during cranking, which would be excessive. Usual base timing until engine start should be closer to 10 deg BTDC. Not quite; the scope shots weren't definitive, possibly due to inductive effects of the coil. Sublute already has a dummy resistor in place of one of the injectors, and is going to do the same at the output of the igniter (in place of the coil primary).

Yes, let's try the 1k across pins 1-2, scope hot lead on pin 1 and ground to chassis, as before. (It won't bother the igniter that 2-3 are unloaded.)

If you use a Fumoto valve and off-road, I have two words: "be careful". From http://www.lubricationspecialist.com/front/showcontent.aspx?fileid=105 : Warning: For users under extreme conditions: If you are using the Engine Oil Drain Valve on off-road vehicles such as farm tractors and logging equipment under extreme environment (i.e. corn fields, brush areas), it is recommended that a hose clip as shown below be used for extra safety and security. The hose clip, which can be purchased at any hardware stores, will firmly keep the lever in the locked position at all times. Safety: Drive carefully at all times, and if you run over any foreign object on the road pull over and check the valve for any damage or leak.

There are certainly a few possible causes, but there's a simple one to investigate. Verify that the injector connectors for #2 and #4 are in the right place (not reversed). The engine will start and run with them swapped, just not very well. There should be a piece of identifying tape around the harness right at each injector connector; it's white on the front cylinders (#1 and #2), and black at the rears (#3 and #4). If the tape is missing, you can check the wire colors. That's not encouraging. A simple thing to check is the PCV system. You probably don't want to think about the other possibilities, which tend to relate to the quality of the work that was done.

The first scope photo looks a bit like my ASCII drawing, but the igniter pulse appears to be inverted. (By the way, what was the timebase setting for that first shot?) Is there any possibilty that the scope has a "trace invert" function that was on? On the other hand, the last photo seems to show the correct igniter pulse polarity, but the timing seems wrong. As you said, it does seem to be erratic, perhaps due to triggering difficulties. The ringing on the waveform is typical of a coil, as is inductive kick-back, and are reasons I prefer to use resistive dummy loads when checking these timing relationships. Can a slightly lower sweep speed be used, so that more than one pulse could be seen?

Seems like you need to get the battery fully charged before proceding. I've attached an ASCII drawing that roughly depicts what you should see as the timing relationship between the injector and igniter pulses.

You're welcome. I hope you and your daughter get well soon. (And eventually, the same for the car .)