You can do a "wet" test by adding some oil to the cylinders and checking for any increase. A significant increase would indicate a ring issue rather than a head gasket (or valves). But, it sounds like classic head gasket. Pumping air from a compressor into the cylinder might give you bubbles in the cooling system, which also helps confirm a head gasket. That works best if you rotate the crank so that the valves are closed on the cylinder being tested.

 

I finally got the head off, the gasket had been sealing perfectly, no loose head bolts, it seems my problem was bad valves.

The intake side was barely hitting the top of the pistons, as you can see in these pictures. Notice the shiny ring around the edge of the valves, and the little dents in the tops of the pistons. What would cause this? Timing chain off a tooth? Loose timing chain?

 

"Barely hitting the top of the pistons" is way too much. As you stated, it could be caused by the intake cam timing being off.

It could also be caused by a previous repair where the head was machined (skimmed) too much. The WIS lists cylinder head height as 139mm (5.47244") minimum after regrinding. IIRC, this is the distance from the surface of the top of the head (valve cover side) to the sealing surface of the bottom of the head (cylinder side).

Finally, using the wrong head gasket could also cause this... Ron

 

Overrevving could also do it, but it would have to be a massive downshift at too high a speed to accomplish that. Seeing how clean those marks are, it would be something very recent that you'd remember.

 

The only way for that to happen and the engine still run would mean the intake cam was opening too early on the exhaust stroke. As the chain guides wear and the tensioner extends, the timing should be going in the opposite direction advancing the timing, not retarding it. Am I correct?

My guess is bent valves causing the poor compression. A good head shop would be able to pull a vacuum on each port for a quick answer.

On my current project post here:
http://www.saabcentral.com/forums/showthread.php?t=469946
I just found the pistons to be hitting the exhaust valves, which is characteristic of a stretched chain. My tensioner was extended 17.4mm.

Did you check the timing before pulling the head? Did you check the extension on the tensioner?

 

The only way for that to happen and the engine still run would mean the intake cam was opening too early on the exhaust stroke. As the chain guides wear and the tensioner extends, the timing should be going in the opposite direction advancing the timing, not retarding it. Am I correct?

No. A worn chain always retards the timing. Its increased length is the culprit. Tensioner is on the non-driven chain side, won't affect the timing in relation to crank position, it only takes up the slack to keep the chain from flopping about, and potentially skipping a tooth on a cam gear.

 

Thanks for the replies.

I did measure the head immediately after removing it, it was about 5.482 inches, or just over 139.2mm, did not look like it had been machined or anything.

The timing is an interesting thing, I wish I had taken a photograph of the cam and sprockets while intact. The notches on the sprockets were to the right of TDC by a few degrees when I zeroed the crank pulley (matching the pulley's notch with the mark on the timing cover). The notches on the camshafts themselves also did not quite line up with the notches on the cam bearings.

The timing chain seems to be in good shape, although I did not measure the tensioner.

 

Here's a photo from Anders (aka Swedecar) over at saabnet. He's an Indy in southern California... Ron

http://saabnet.com/tsn/members/gallery.html?memberID=112&do=show&id=180

 

Good resource, thanks. When the gasket and bolts arrive I will see if I can time the engine properly. I believe both intake and exhaust cam timing was advanced.

I loosened all of the cam bolts and sprayed WD-40 down the valves, most of the intake side were leaking.

I also just remembered, I used to hear a pinging sound from the engine during WOT pulls that I thought was from gasoline, that may have been those poor valves.

 

Both the T5 and T7 ignition systems detect pinging electronically by using a low voltage pulse through the spark plug after ignition rather than a common knock sensor which is basically a small microphone attached to the block in many cars. That's why it's important to use OEM recommended spark plugs to detect pinging... Ron

 

To determine whether the chain is extremely worn, I think doing a mock rebuild with the old head, reinstalling the chain tensioner, and turning the engine over a few times will give me the appropriate value, right? I have no way of knowing without doing this as I reinserted the spring into the tensioner when it was out of the vehicle.

 

Just an update, got the car put together and I think the cam timing is correct. The Mitsubishi TD04 turbo is far superior to the Garrett T25.

However, it stumbles and misfires only on very hard acceleration, and "CHECK ENGINE" flashes. The code is P0302, cylinder 2 misfire. I hope it's something simple like the ignition cassette or a vacuum leak.

EDIT: I had considered the possibility of the mixture not quite being right with the new turbo, but wouldn't that cause misfiring in all cylinders?

 

Its possible you are getting spark blow out, try gapping the new plugs a little closer (maybe around 0.032") if that doesn't help a new DIC may.

 

Spark blowout?? Oh the wonderful world of turbos...I think that's *exactly* what's happening. These NGK plugs were in the junkyard head when I bought it, and I foolishly did not check the gap. They all have a much bigger gap than the Bosch plugs from my old head. I'll try replacing them immediately.

 

I cleared the codes, dumped the Bosch plugs back in (they are resistor plugs!), and took it for a test drive. No hesitation, turbo boosts very reliably. I'll buy some new NGK plugs soon.

This car is a completely different beast now. The low compression was obviously a problem since I bought it, as engine braking is a very different experience now. The needle for the turbo gauge drops all the way to the bottom of the green, it never did that before. Fuel economy is much improved at the same time as performance.

 

I checked the tensioner after installing the head temporarily with the old gasket and turning the crank a few times, and I was very happy to find that the chain was not worn and well within limits. Apparently it was just replaced not too long ago and the timing wasn't quite set right.

Now I'm having other issues, an embarrassment of riches so to speak - the engine outputs far too much power to the wheels, WOT through first gear causes the tires to start spinning at around 4000-5000RPM and I can hear the distinct sound of engine mounts failing (clonk). I just replaced the transmission mount and apparently the others were all on the verge of failing.

So next on the list is definitely re-flashing the T5 to a more appropriate fuel map, right?

 

How long have you had the car?

Sounds like someone tuned out the torque limits in first (and likely second) gears or put in an MBC to control boost (same effect).

It doesn't take a lot more HP to break the front tires free in these, especially if you have a little left/right or front/back weight transfer going on. Even after suspension work, a little dampness on the road, or some tar on the asphalt, etc, and just a little extra HP tuned in can do it.

 

HI,
Im not quite following here- You say you put your car back together and its running real nice and strong... What repairs or work did you perform before re-assembly of head,etc?...

From what ive read and just experienced on my car, the camshaft timing marks will be slightly off, both showing slightly toward the firewall side of marks..

As Bob suggests, sounds like you may have a MBC installed on that car, or some other sort of tune/remap

 

Completely forgot to mention this somehow - this whole project started as a turbo swap from Garrett T25 to Mitsubishi TD04HL-15T 5cm^2. I've had this car since June 2014.

The second cylinder head I used had nice tight valves. The timing marks on the camshafts were a little towards the firewall side, but the cam sprocket marks were exactly perpendicular to the plane of the head surface.