GeneralDisorder

I vote for the steering joint being frozen from rust. Make sure that it has free movement on all the u-joints and in both directions that each joint can travel. GD

The tach signal is a generated signal that is produced by the ECU from the crank sensor input - it is present on all ECU's so you should be able to run a wire if it's not in the harness. The speedo head incorporates the odometer so you would have to swap the whole speedo head. It will likely work but I've never tried it personally. I do know the speedo heads on those years were seperate and can be replaced without replacing the entire cluster so it should be possible. GD

Lap the valves - could have carbon debris in them from removal, etc. Retest with solvent in the ports if you have some. Generally water is not the best choice around valves and valve guides that are steel. GD

'96 to '99 phase-I EJ25D's are the engine with the HG's that experience fire ring failure at a high rate - approx. every 100k miles. No other engines before or since have fire-ring failures that result in random overheating. '99 to '05 phase-II EJ251,EJ252, EJ253, and EJ254 have a decreasing rate of coolant and oil leakage from the head gaskets to the exterior of the engine. This is annoying and smelly but as long as fluids are topped off regularly they will not overheat or otherwise experience damage as a result of these leaks. I've OWNED a '99 EJ253 with over 250k miles on the original HG's. Yes they leaked a TON of coolant and oil - the coolant can be mitigated or stopped with the Holts Radweld that Subaru sells for $1.49 for this purpose and the oil leaks are just annoying.

They leak oil - primarily from the drivers side head at the rear but often from both. That wasn't solved till the 2010 model year. You can simply replace then with the STi "642" gasket and you'll never deal with the leaks again. Maintenance has NOTHING to do with this - the coating on the gaskets washes out and a leak is formed. Plain and simple. I guarantee with 100% certainty that at the first timing belt interval you WILL have seepage evident from the back of the drivers side head. You can replace them or live with the burning oil smell - your choice. GD

They build the idler at a (small) bearing factory. It just has non standard races. There is no difference between an EJ Subaru idler and any off-the-shelf bearing from a manufacturing standpoint - just different shaped races. I should point out that *some* of the early single-row bearing smooth idlers do have seperate off-the-shelf bearings. The bearing is about $10 and the idler is about $18.... it's not cost effective from a labor standpoint and in any case even if the car used two of those for the smooth idlers you still have to buy the cogged one. Subaru no longer uses the single-row bearing idlers. All of their engines have double-row idlers now and those do not have a bearing that can be replaced. If you care about longevity you should be using the double row bearing version for both of the smooth idlers anyway - thus making it again not possible to press out the bearing. They are not. They are stamped steel idlers with pressed in bearings. Different era and different technology. EA and ER idlers can be serviced - though it is not currently cost effective to do so as replacement components are availible already assembled. I do replace bearings in lots of things - the older EJ AC belt idlers for example. The idler is $35 - the bearing is $5. This makes sense. Sometime in the early 2000's they went to a molded plastic idler that cannot have it's bearing serviced. GD

This is all that's used in my shop: http://www.inlandtech.com/catalog/products/brake-prep It works well for general cleaning also. I have a brake parts washer that runs a 3:1 mix of brake-prep and water - it's on a rolling cart and rolls under each wheel to clean all brake components. It's very safe and limits exposure to harsh chemicals - it dries completely clean - cleaner even than aerosol brake cleaning products. I use it to clean valve covers, heads, timing belt cover, engine blocks, etc. A real pleasure to use and all the guys that I've introduced it to swear by it. GD

The bearings are not replaceable on the EJ idlers. The inner and outer races ARE the idler. There is not a separate bearing that can be pressed out. Get the mizumoauto kit from ebay with the Aisin pump. Mitsuboshi belt is nice but I haven't had any quality issues with any of the belts I've used and your '90 is non-interference. GD

Shouldn't make any difference. Sensor tip temperature only need be over about 600* F to operate correctly. Which should be easily obtained since the sensor has it's own insulation and heating element internally. I've never encountered an externally insulated sensor and in any case if the sensor wasn't reporting or the heating element was bad then that would set different codes - not the 420 code. GD

Yes - call me and either I can take care of the HG's for you for a better price or I may be interested in taking it off your hands, etc. I'm in Portland but travel/pickup is no problem. 5o3-88o-4o84 GD

You can use stock '99 struts for the '04 springs. Or '02 to '05 WRX struts would also work and there's probably many more options for those. I have '02 WRX KYB's on my '91 SS - using '06 STi springs - ride height didn't change appreciably. They have the same cone-shaped rear springs that match my stock strut tops. But '04 would be conical and would match your '99. GD

The gasket ports will not line up. That's why trimming is needed. Ivan is saying he used 2.2 Turbo head gaskets which are composite not MLS style gaskets. They would be much thicker resulting in even less compression. But they would be easier to trim for the cooling ports being graphite composite gaskets. The gasket thickness will space out the heads farther than normal - resulting in the timing belts being tight. So it makes sense he had to modify the belt tensioner. He stated the year - 2003 Impreza. Which would be an EJ251 - so that's the heads he used. Seems like a lot of work but if Ivan says is works then I would be inclined to believe it. Though an Imp is going to weight a lot less than a Legacy GT so the decrease in performance would be less noticeable. GD

The new CVT is a variable pitch pulley design and is quite good. You are simply wrong about it using more fuel - it's more efficient than any other transmission type due to it's ability to always be in an optimal gear for fuel economy. You can't make judgement calls like that based on a single test drive. You do not have a large enough sample data set from which to derive any useful conclusions. Your speculation is unfounded and biased. Kia (and Hyundai - same thing) is Mitsubishi CRAP. I've worked on plenty of them. Take a look at consumer ratings of their cars. Hell their dealer tech's can't even fix them. The engineering and implementation is extremely poor. I work on several Kia's at my shop and when I research my customer's complaints I am astounded by how much information is out there on Kia and Hyundai problems that seemingly can't be fixed. I've read several reports of brand new Kia's being declared lemon's and having to be bought back by the dealer. Having worked on them - I can't say that I'm surprised. GD

McMaster-Carr. GD

90 to 94 rear top hats use a different bolt pattern than 95+. 90-93.5 used cone shaped springs and 93.5/94 used conical but the bolt pattern stayed the same and you can't use 95+ top hats without modification so it won't work the other way either. GD

All that does is put the ECU into "d-check" mode for diagnostic purposes. It cannot damage anything. They should be disconnected for normal driving. Check to see if your distributor is turning. You may have lost the drivers side timing belt. Check the screw that holds the rotor to the disty shaft also. If neither of those solve the problem and the coil, plugs, and wires check out then you may have a bad ignition amplifier on the coil bracket. GD

With low compression it's going to be rediculously underpowered. The '00 2.5 had 165 HP. The 2.2 has 137 and with low compression it's probably going to struggle to hit EJ18 performance which is 110 HP. You should find another EJ25. The EJ22 in that car with low compression heads is going to be scary to drive. GD

Holley 5200 for sure. Holley licensed copy of the Weber DFV. They kinda suck - different venturi sizes depending on application make them a chore to tune. You can't just use what someone else used as a baseline jetting.... They should accept standard Weber jets though. GD

Both carry away heat. Oil also lubricates. GD

The phase-II 2.2 uses the same style HG as the phase-II 2.5. They both suffer from the same external leakage due to failed HG coating. There is no appreciable difference in reliability as neither suffers from the fire-ring failure of the phase-I engines. There is NO compelling reason to prefer the 2.2 to the 2.5 in phase-II trim - less power and less common. That's the whole reason Subaru dropped that engine. GD

What transmission? GD

Thermostat isn't your problem Find the leak..... GD

'99 to '01 Last years they made them. Frankly though - they aren't any more reliable than the 2.5 IMO. GD

'96 2.2 heads aren't even an option because the manifold won't bolt up to them. Regardless of condition they will not work. '96 2.2 is not the engine you need for that car. Nothing about that combo is going to work well. You need a '99 to '01 2.2 GD

Compare the 2.2 HG to the 2.5 head - there may not be enough material to seal against the fire ring. You can usually put a smaller bore head on a larger bore engine but to go the other way invites problems with combustion chambers too large to seal against the HG properly. Also the 2.5 heads will probably have some differences in the cooling ports that may not line up well with the 2.2 HG. Even if it will work with the HG.... I'm not sure of the compression ratio. So you will have to do some research there. It will likely be a bit too low. The tendancy was toward larger combustion chambers in later years so you might end up with an engine great for turbo-charging but completely gutless on it's own. Generally crossing the phase-I and phase-II boundary with blocks and heads is not done. There are some exceptions but all of them involve smaller heads on larger blocks. You may be able to use phase-II 2.2 heads. Those would be a better fit and would mate up to your manifold. GD