Latest update: We tore down the engine this weekend looking for a noise that some thought sounded like a rod knock, some thought sounded like piston slap, and some (me) thought we should just ignore and keep racing.
It's been making a subtle, but increasingly persistent and nasty noise at part throttle and it's been filling its oil filter with metal ever since we first assembled it from a stack of shitty old broken Miata parts. We did about 3 hours of track time before the first time someone thought they heard rod knock. We couldn't duplicate the noise, so we wrote it off as in-helmet acoustics.
We pulled the first oil filter at about 7 hours and found this:
Still couldn't make the rod noise, and we were out of time, so we went ahead and raced.
Halfway through the race, we started noiticing the rod knock noise from pit lane as the car went by on the straight! But only during yellow flags. On boost, the engine was silent and smooth. When we pulled and cut the filter at the halfway point, there was a lot less metal than before, so we decided to go ahead and drive it like we meant it (we had been short shifting at 6000 ever since the first filterfull of metal). Finished the race with no drama (other than almost winning and almost setting fast lap).
One more filter, a little more metal, and finally we could duplicate the noise. If you revved the engine, it would make the clatter for just a second at the top of the rev. None while revving up, none while coasting down, but a little clatter right at the top. Ron Bergenholtz was standing nearby for one of our clatter tests and he said it sounded like piston slap. That actually made sense, since a turbo motor always has positive pressure on the top of the piston when its on boost, but when there's vaccum in the manifold, the pressure delta across the piston goes negative on the intake stroke and positive everywhere else. If you have an ancient block and pistons, maybe the piston-to-wall clearance would open up enough that the pistons could rock back and forth?
After this diagnosis, we did another 6 hours on track at Laguna Seca.
In total, we think we've got 33.5 hours of track time on this engine, at least 30 of which were after the first time we thought we heard rod knock.
So, this weekend's teardown? No cam bearing damage. Piston skirts look fine. Cylinder walls look quite worn (hone marks gone in 30 hours!), but nothing that would shed metal. Rod bearings? oh...
All four look like this.
How he hell did we manage 30 hours of track time with bearings that looked like this?
1: dumb *** luck.
2: during assembly, we picked up a washer off the floor and slipped it under the oil pressure relief spring, kicking up oil pressure some random, unconfirmable amount.
3: Before the race, we went through the oil shelf in my garage and dumped every half-used bottle of oil into a jug and poured the concoction into the engine. Since we usually use the good stuff in our other cars, a lot of those partial quarts were synthetic. I think the mix resulted in something like 7W39 quasi-syntheitc. That probably saved our asses...
Those bearings had less than 60,000 miles on them when we installed them, and we put them with the crank they originally knew (though not necessarily on the same journals...) So maybe you shouldn't move used bearings around?
Thanks to wholesale parts discounts from our lone sponsor, MD Automotive, we just spent $15 on a head gasket, $25 on new rod bearings, and $3 on oil pan gaskets and put it back together. After a race, if your car survives, you can ask Jay Lamm for a value on your car. He said ours was worth $275, leaving us $225 to spend before the next race. Since this weekend was half-price weekend at Pick Your Part, we ran over there, found a Volvo 240 turbo, and snagged the oil cooler ($8), lines (2x$4=$8), and sandwitch adaptor ($free! they didn't notice it). I think we'll be back on the track for $59.