SHORT VERSION - My 2.0 Subie Turbo developed a mechanical knock/rattle sound (not pinging/detonation) and doesn't want to idle. Anything I can do/check before tearing down the motor?

LONG VERSION - I was out in my buggie today and was not running hard, when I had a problem with my Subie. When I ran up the RPM's I heard a medium loud noise (or rattle) that I thought sounded a lot like a rod or bearing knock. I heard the noise a couple of times and pulled to over to check it out. When I came to a stop the engine would not idle and stalled. When I tried to restart it cranked a little hard and fired, but would not idle unless I applied some throttle. I proceeded to limp home for a couple of miles trying to keep the rpm's under 2000. The engine did not miss on the way home, but if I raised the rpms I started to hear the knocking sound (which sounded like a rattle on deceleration). When I came to a stop the engine would stall.

Here is all that is known about my motor (I was not the owner during the build)

* 2.0 EJ20 long block (exact year & model unknown, assume late 90's JDM)
* Subaru factory turbo
* Outback intake and exhaust
* no intecooler
* tuned by The Dyno Shop, Santee for 8 psi low boost, 12 psi high boost on 111 octane leaded
* 2 years since installation
* Super low mileage since install - 85% road, 10% hardpack, 5% sand
* Have heard detonation once or twice in the past, even when running 111
* I usually run with the dial-a-boost down, keeping boost below 10
* I occassionaly run 91 octane when cruising around town, and turn the boost down to 5-6 psi max.

When I got home I determined the following:

*coolant level was normal
* drained some coolant and it looked fine
* oil level looked high
* drained some oil and it appeared to be thinner than normal (maybe just hot?)
* I could not detect any metal in the 10 oz (or so) of oil I drained.
* oil did not appear to have water in it.
* pulled all 4 spark plugs and they looked dry and burning properly
* no sign of external leaks on abnormalities on the engine
* restated the engine when cooler, and still did not want to idle without extra throttle
* didn't rev the engine much on restart and didn't hear the noise
* couldn't check the compression yet, need to get gauge tomorrow


Last time I used the car was in Glamis. Thinking back I remember that I accidentally buried the tach running up a hill behind Olds. Never did that before, but the car ran fine the rest of the day after the red-lining.

I know it hard to diagnose a motor without tearing it down, let alone over the internet, but hear are my questions.

1) Are there any other useful diagnostic steps I can take before tearing down the engine? i.e. Oil pressure? Drop the oil pan and look for metal? Oil analysis? (I'm going to do a compression tomorrow).

2) Is there anything else to check other than engine internals? i.e. If my timing suddenly went off, could I hear a mechanical knock/rattle? (not a pinging sound, more mechanical)

3) If I was running too much boost or timing, what is the typical failure mode for these engines? i.e rod bearing, main bearing, broken ring etc?

4) If I overeved the engine what is the typical failure mode? i.e spun main bearing, rod bearing etc?

5) From what I've heard it's cheaper to get a new short block or used JDM long block than trying to rebuild the bottom end. Does that sound correct?

6) What do I need to check to see if I can bolt my 2.0 heads, manifolds, etc onto a 2.5 short block? Will I need bigger injectors or turbo if I am shooting for 300HP (or less) on 91 octane?

Thanks to anyone can assist!!!

This got a little long, sorry, I'm taking my frustrations out by pounding on this keyboard...

Sorry I can't help you with your problem, But I do have a 2.0 long block for sale. CHEAP
P.M. subie4me

the first thing i would check in your situation is to see if your timing belt has jumped a tooth.just pull off the big plastic cover on the front and refer to a manual to make sure everything is where it needs to be.the downfall is that if your belt jumped it can cause interference between the pistons and valves and do some serious damage.

Thanks for the reply iscream.

I just pulled my valve cover and found that the left side exhaust cam sprocket has jumped a tooth on the belt.

I'm looking online now for timing belt alignment instructions for the proper way to re-align the sprocket.

After I get the sprocket re-aligned, I going to check compression for any damage to the valves.

Thanks again!

I bet dollars to doughnuts you bent a bunch of valves, I just did a side job a couple weeks ago where the guy did the exact same thing, sorry to give you the bad news, but he needed 4 exhaust valves and 6 intake valves.

not a good thing to do at all, i wish you luck.

I took the timing belt and idler pulleys off today, and re-aligned with "jumped" cam cog to the proper position. The cogs and belt looked fine, but a couple of the idler pulley bearings were worn and making noise, perhaps because of the grit and sand that had worked it's way behind the covers. I'm wondering if the bad idler pulley bearing were enough to make the belt jump one cog at high RPM's?

After I got the timing belt back on I ran a compression check which looked good - 138 to 140 across all 4 cylinders. I know a leak-down test would be a better check for bent valves, but I don't have a leak down gauge. Since the compression is good, and the engine tuned over easily and smoothly, I hope that my valves are not bent!

After the compression test I put everything back together and tried to start the engine, BUT NO LUCK. I carefully rechecked the crank, timing belt, and cam cog alignments, which looked good, but still wouldn't start. I checked for spark (on one plug) and it was good. I thought maybe it flooded during the compression check, so I waited a while, but it still wouldn't fire. It just cranks and cranks, without catching or firing at all (like there is no spark).

If I have spark and gas, and the cam timing is correct, what should I check next?

Are you sure that you have spark, check it by pulling the coil off and putting a spark plug in it and crank it to see if you have spark because if you have the timming off just a bit the cam sensor and the crank sensor if they are not in sync they wont trigger a spark. Oh and sort off the subject you can run easy 10 to 12 psi of boost on 91 octane on a stock 2.0 subaru I ran mine at 14 for along time and never had a issue and I sold it and it is still running Outfront says 91 is good for 12 psi on stock 2.0 Hope this helps good luck!!

You mentioned you checked alignment but did you rotate the assembly the 14 times to check out the sync?

Maybe you have a stuck lifter? That's what I had. Keep in mind that these 2.0 subies are pulled from a Japanese car, shipped overseas, and sit in some guys warehouse for quite a few years. The noise comes from the Head and sounds like a bad knock. It didn't idle propely unless you gave it some gas then once the engine warmed up the knock goes away. Not sure if this is your problem, but it was mine.

Hey thanks for the reply.

Turns out is was the worst case scenario. A badly spun main bearing and the contaminated oil did some damage to the heads as well. I'm half way thru the new engine built using a 2.5 STI short block this time, along with upgraded turbo, injectors, etc.

Can't wait to be back out there!

We did the same thing in our 2.0 with the stock turbo. We went to the 2.5 with the T3/4 turbo. Night and day difference, you'll love it.

If you put those 2.0 heads on a stock sti block, I'm pretty sure that the c/r is gonna be a bit on the low side. You might want to pick up a set of 2.0 twin turbo heads, or even a set of 2.5 dohc n/a heads.

If they are Phase II heads and he uses an STI head gasket the compression will actually be a half count higher, 9.0:1 instead of 8.5:1. Late nineties would indicate phase II. If they are EJ20K heads then I believe you're right, compression will be lower, I'd have to check to be sure.

Hey thanks guys. I had the old EJ20G hydraulic rocker heads with the 57cc combustion chamber. With the thin STI metal headgasket I believe the C/R would be around 8:1.

I decided to upgrade to twin turbo heads with the smaller clover style combustion with the 46cc combustion chambers. With the thicker 1.36 mm headgasket I think the C/R should be around 8.3:1.

I'm sticking with the Phase 1 heads so I can use my current intake manifold, etc.

Thanks!