Project BMW M3: Part 21

With the amount of track and performance testing Project M3 has undergone, it's safe to say I've tested its limits to a greater degree than most other owners of AA Turbo-equipped M3s. I'm now on the second engine block, fourth cylinder head, second turbo, third transmission and second clutch disc. Okay, it sounds a lot worse than it really has been. In truth, the only parts that actually broke were one head gasket off the factory 3-liter (my fault for turning up the boost and not the octane) and one transmission. We haven't quite figured out what happened to that tranny.

The current turbo is a slightly different unit from the original AA Stage 2 unit. Last year, I moved up to AA's Stage 3 turbo when evosport took out the stock 3-liter in place for a built-up, low-compression 3.2-liter with forged internals, even though the condition of the stock 3-liter block was fine. I was just going after longevity at that point.

So what about the four-cylinder heads? Back when the car was normally aspirated, the cylinder head was rebuilt after testing the Schrick camshafts in order to prepare it for AA turbo power, especially since the valve guides were a little shot. When that 3-liter was ditched, I was on my second cylinder head-one from a 325is--which came with the 3.2 block from Bavarian Engine already ported and polished. Suspecting the excessive oil consumption at the time was coming from this cylinder head's valve guides, I had it rebuilt once again by Bavarian Engine, only to have the same problem: a quart of oil gone every 400 miles, and it was three times worse on the track.

"Forget it," I said. "Let's get the old cylinder head back on." After its first build-up, it didn't lead to excessive oil consumption when it was turbocharged with AA's Stage 2 system.

The original cylinder head had been at evosport all this time, and I gave Vadim Federovsky the "go" to get it rebuilt before bringing the car back in for the swap. During this time I made sure to get my hands on the hard-to-find BMP Design cutting ring head gasket, which I finally was able to hunt down. After receiving word from evosport that the cylinder head was successfully freshened using my original stock parts, I brought in Project M3. Just to get a baseline, evosport dynoed the car at 15 psi, the boost I had been running with 95 octane (equivalent) fuel thanks to a Sunoco race gas mix.

I went into the test guessing that something might be wrong. The tires were getting plenty of grip under full boost in second gear, whereas previously I had been getting complete wheelspin. Maybe it was just me; the car still felt really fast, but I wasn't experiencing quite the rush I remembered.

The question was answered on the dyno. Expecting (and hoping) power to be near the 417-whp mark it had recorded at this boost level, I was shocked to find no more than 360 whp, and it wouldn't hold more than 14 psi after 5500 rpm. The engine was down nearly 60 whp from the last time we tested on this same Dynojet. Something was obviously wrong, and there was a "whistling" sound I had never heard before coming from the turbo.

Initially we thought perhaps the HIOP cat-back exhaust I was testing was crapping out and not allowing proper exhaust flow. The evosport technicians quickly removed the exhaust and tested the engine again, only to find no change (this goes to show that uncorking a decently working rear muffler may not yield much except a lot more noise, even on a turbocharged car). What kept me from worrying too much was that the power curve was still smooth, with no signs of detonation or misfires.

The one thing I didn't want to find out was that something was wrong with the block. But, I knew that when I pulled the crankcase breather and saw no smoke coming out of it that it was one good sign--at least the rings should be okay. And, another good sign, the compression tests were still good, with compression varying between cylinder to cylinder by no more than 2 psi.

After several hours of labor, evosport technicians pulled the head off to inspect it. Sure enough, even though I had been running the Chevron Techron Fuel System cleaner through the intake about every 500 to 1,000 miles, it was obvious some oil was still caking onto the pistons. Just like the first time we had this Bavarian Engine cylinder head rebuilt, we were suspecting bad valve guides to be the culprit. To check them, evosport pulled the valve springs and wiggled the valves. When new, the valves should have no more than .001- to .003-in. of play. To our surprise, the valves showed between .004- to .005-in. of play, which was still pretty good.

Our next place to check was the turbo itself. It was removed from the car and inspected, and it was soon obvious where that whistling sound was coming from. First, a small portion of the oil feed line to the turbo was crimped, definitely not a good sign. Next, when I manually spun the turbine wheel with my finger, it became apparent that the turbine wheel was touching the exhaust housing! This was obviously one of the reasons, if not the main reason, for the reduced boost pressure.

In order to find out more, I took the turbo unit to turbo expert Tom Rose of Performance Techniques. Tom has worked with Active Autowerke and other tuning giants from around the world, and Performance Techniques has been building and rebuilding turbochargers for all makes and models, including 18-wheelers, for several years.

Faster than I could reach for my camera, Tom took the turbo apart and quickly found several signs indicating it had suffered from just too much heat. First, the exhaust housing of the turbo had oil build-up that eventually got in the way of the free-spinning turbine wheel-that was the whistling sound. Second, the thrust bearing, which should have stayed its normal bronze color, was black and coated with oil along with its bearings. Tom saw that and said, "See all that oil there? That's very hard to do!"

According to Tom, there could have been a number of reasons for this turbo mishap, including too high exhaust gas temps, running incorrect timing or too lean, or even hot shut-downs. As I thought about it, I knew my fuel curves with the AA chips were pretty fat, and the egt, at least on the street, hasn't been extremely high, either. Perhaps I could have let the car idle a little bit longer after a drive every now and then, but this oil consumption problem started nearly since day one with this motor and Stage 3 turbocharger, soon after I broke it in.

A moment later I remembered about the crimped oil line and told Tom. "With a slightly crimped oil line, the turbo could have been getting enough oil to lubricate it but not enough to remove the heat, and the oil itself is what does the most in removing heat," Tom said. This can cause turbo bearing failure, which in turn can cause oil to be sucked through the turbo and into the motor. In hopes that my oil consumption could be traced solely to the turbo bearing failure, I asked Tom, "Well, could oil being sucked through the turbo and into the motor be enough to equal at least one quart every 400 miles?"

"Oh yeah!" Tom said. "I had a guy drop several quarts of oil from a bad turbo on the way here on a 60-mile drive!" Amazing that something as minor as a crimped oil line could possibly be the cause of so many problems.

In a way, I was relieved to find that a turbocharger failure alone could cause all of these problems, simply because it's a much easier to part to diagnose, fix and even rebuild than if it were the cylinder head or block. It bothered me, of course, that the original cylinder head had already been rebuilt and the current cylinder head had already been removed from the car when, in fact, we probably should have just checked the turbo first. My decision to go ahead and have the first cylinder head rebuilt for the second time was so that I could save the downtime on the car, when in reality it probably cost me a lot more money for no reason. But that's what I get for trying to get everything done so quickly. There's a lesson here. Regardless, rest assured Project M3 won't end on a bad note.

Just because Project M3 is temporarily down doesn't mean I'm forgetting about it. In fact, there is a real neat product you can buy for your BMW if you already own a Valentine V1 radar detector. As I had mentioned in Part 16, I never leave home without this radar detector. Although not the best-looking radar detector in the world, it works extraordinarily well. The only downside is it does take up some room on my windshield. Although a V1 concealed display unit is included with each Valentine V1 radar detector, I've been lazy and hadn't set the time to hook it up. But now it looks like I won't have to because of what Umnitza has made available.

Using a stock BMW mirror, Umnitza has masterfully placed a V1 concealed display unit behind the mirror. It doesn't restrict any of the view to the rear, and unless the V1 arrows are flashing, you can't really see the unlit arrows behind the mirror, making it visible to you only when radar is detected. An ingenious idea, Umnitza sells these mirrors for most BMWs and will have them available for Porsches as well.