GeneralDisorder

Failbucket, Photophucket, Image-hoseing.... lol. Only a matter of time before someone pulls the plug on that zombie. GD

2002 was the last year for reseal kits. They are no longer offered. GD

Amsoil ATF. Subaru HPGO for diffs. Amsoil 10w40 Premium Protection engine oil. GD

It was probably only running on two cylinders. The computer just isn't smart enough to throw 3 and 4 misfire codes under some circumstances. GD

95 Impreza is not OBD-II. 93 to 95 Impreza are similar to each other, but share nothing with Legacy models ECU wise. The car doesn't do any additional diagnostics unless you go through the entire d-check procedure. GD

The answer is yes and no. What will happen is that in mixed driving the temperature delta between hot and cold will be larger and that will cause unnecessary thermal stress on the gaskets - potentially leading to even more rapid gasket failure than if it were more tightly controlled and thermal stress was restricted as much as possible to warm up and cool down periods only. Thermal cycling, and head gasket thickness were the primary reasons for failure. Decreasing the operating temperature without also chasnging the software fan set points is pointless and only create more thermal stress on everything. GD

To be clear - I never said it can't throw codes for the crank sensor OR the cam sensor. What it will NOT do is throw a code for "crankshaft / camshaft correlation" which is a DIFFERENT code that indicates the two sensors are not in sync with one another. These early ECU's do not have the capability to detect this. GD

As far as filling voids in old bushings, yeah it will do OK with that. I've used it and I wasn't terribly impressed by it. But for making completely new bushings it's not liquid enough to flow and doesn't have the right durometer - you can also get the smooth-on stuff in any durometer you like. It flows out to provide a really nice finished look if you do it right using a vacuum chamber (you can make one) to pull out all the air pockets from mixing. GD

Up till about 2015, the STi spring rates weren't spectacular. 15+ STi's have fantastic suspension. Almost so good that you have to spend a LOT on coils to improve upon it significantly. We put $1800 Tein coils on a 2016 Hyperblue and although it lowered it, the handling and ride was frankly about the same. KING springs are VERY nice and are our go-to for performance springs. If you can find some take-off 15-17 STi suspension would say get those. They will drop your car about an inch. Yes they fit the GR chassis. GD

Unlike later model ECU's, these won't throw codes for camshaft-crankshaft correlation. They just cease to function if they can't sync the two signals. There is a lot of the OBD-II library of codes that were completely unused till newer models. Newer stuff will freak out and throw codes if it's off one tooth on most cars. Also - I have seen injector driver mosfet's fail, and ECU's do strange things in cold weather, etc (even not start at all), but rarely do they completely stop working. Usually they run like crap or lose a single injector driver, or have a circuit failure that causes a code even though the component is good.... I've only replaced a very tiny number of complete failures. Maybe 2 or 3 in a dozen years. And I see a LOT of these cars. Like an average of more than 2 a day. Whenever we get a no start - we start with checking fuel, spark, etc just like you did. At some point though we usually at least pull the outer belt covers and check timing - but as you noted - the trained ear can hear if it's out of time when cranking, or obviously has low compression on one or more cylinders..... the process isn't cut and dried because you use all those senses, as well as conversation with the customer to determine diagnostic path. GD

You just need a back probe, and some alligator clip test leads. Many tech's use these for back-probes: GD

That stuff isn't nearly resilient enough. The two part products by smooth-on are a much better option. GD

I'm fairly certain the H6 is sequential injection, not batch fire. Even going back to 1990 the EJ22's were sequential injection. Ohming the wire is not sufficient as the meter does not use sufficient current to test the wire's ampacity - you need to perform a voltage drop test, or wire the noid light directly at the ECU connector to verify that not only does the ECU ground that pin for the injector, buy also that the wire can carry a load and not drop all the voltage due to a poor connection. My money is on a pinched harness during the HG swap. GD

Well that's looking at two different sides of the connector set which are mirror images - what will really blow your mind is that schematics aren't consistent labeling pinouts as if you are looking at the front or back of the connector..... that's why you need to cross reference the pinout with the schematic containing wire colors and size. Quite frankly though, if you can't verify the waveform, then you can't confirm or deny an ECU failure. It's just not possible. So unless you can find another cause, you WILL be trying a different ECU for better or worse. GD

You mean 5 and 6? The hall effect waveform would likely show something in that range while cranking - the meter is going to average the waveform and give you some weird voltage that's.... well... meaningless because it doesn't tell you if the waveform looks correct of if the two waveforms are synced. GD

But, it looks to me like they are actually 1 and 2, not 15 and 16..... but I would need the pinout and schematic in front of me to tell for sure. I almost never work on models that old anymore and it's been years since I stripped a harness for a swap. GD

No 15 and 16 is correct. If you count only the positions that have wires, or have a HOLE where a wire/pin could be inserted, and NOT the solid plastic blank spaces, from right to left, then the large yellow/blue wires are 15 and 16. GD

You don't count the blanks. You only count the holes that have a wire, *or could accept a wire*. In your picture, 6 is either the white wire on the left, or the green wire on the right. My money is on the white wire as that one is next to a black wire (camshaft position negative), and white/black is the color for the hall effect sensor shielded pair cabling used in those harnesses. So yeah - it's the white wire on the top left. GD

You really need more than the pinout - you need the Engine Control System schematic to reference for wire colors along with those pin numbers. And you REALLY need a scope to verify cam and crank sensor signal integrity and sync. A USB one for a laptop is probably the cheapest entry point on those. Those two sensors, along with voltage drop testing (not just standard DVOM readings) of all the 12v and ground pins... .if all that checks out I would be then be confident replacing the ECU. I'm not a parts changer - I test, verify, then replace. ECU's do go bad, but it's actually pretty rare. This is why diagnostics aren't cheap - my scope and it's accessories cost $4,000. But I don't get to ask the customer to pay for an ECU they didn't need. If I call it out - it better fix the problem. GD

Now you need to scope the cam and crank sensor signals to the ECU to be certain it's receiving correct sync data. Do you have access to an oscilloscope? GD

Yeah if you have no pulsing at the injector that's probably a priority. Do you have 12v at the injector but no control signal from the ECU? 1LT has a point - check the connector carefully. That may have been touched/broken/disturbed in the HG replacement so it's on the short list of possible problems. GD

Get a new o-ring and suction hose. Should fix the problem. I use double Oetiker ear clamps on the pump elbow. GD

That never happens. They leak from the weep hole due to bearing failure. They don't stop pumping. This isn't a VW.

130 is pretty low compression. Don't rule that out as a cause. I would check against the other cylinders - should be up around 190. That's over 30% off from normal - way outside of acceptable 10-15% variance. Also head gaskets were just done so in all likelyhood it is related to that repair. GD

Thermostat change on a computer controlled engine will only confuse the computer and potentially create codes. The ECU isn't going to run the fans till it hits 204 regardless of thermostat. GD