Project Frankenmiata: Failure is ALWAYS an Option
by Dave Coleman
It would be easy to mistake the Eyesore Racing 24 Hours of LeMons juggernaut for a polished, finely honed endurance racing machine. Yes, we've won a lot of races, and consistently finished well in ones we've lost. Yes, the car is powerful, reliable and easy to drive, but careful analysis of the stats reveals a dark streak of failures. The last year brought destruction form every direction, but through dumb luck, we managed to clump 6 races worth of mechanical misfortune into just two races. This is the story of all our shit going wrong.
A Mazda5 is a surprisingly soundproof cocoon for sleeping between stints at a 24-hour race. Soundproof enough that I never really heard the mechanical cacophony when our car made an unscheduled 2 AM pit stop a few paces from my mini-minivan. Even without sound, enough of a vague sense of panic wafted into my semi-comatose consciousness to get me to pry open my eyes, slide open the door, and shoulder-roll off the taco-shaped queen-sized inflatable mattress I had shoved atop the folded back seats.
Until now, this race had been almost routine enough to be boring. The Frankenmiata is so shockingly reliable, we think nothing of running back-to-back endurance races on consecutive weekends. We had just run 14 hours of LeMons at Sears Point the weekend before, and then, with nearly zero mechanical prep, were running 24 hours of ChumpCar at Streets of Willow Springs. We hadn't even changed the oil.
The sound coming from our engine made it clear the days of simple reliability were over. The noise would fade away at idle, but revving the engine resulted in a bone chilling jackhammer-on-cranium noise. The once-per-revolution frequency, load dependence and deep tone all pointed toward rod knock, but the sound was unmistakably coming from the cylinder head. But cylinder head noises should always happen at half engine speed (camshaft speed) and are generally a tinnier, higher frequency tone consistent with small things behaving badly. We were quickly and thoroughly baffled.
To his credit, Alex Vendler, guest driver and Gnome builder, didn't have the intimate knowledge of our turbo system when he immediately concluded, "It sounds like something got sucked into a cylinder." Sure, it sounded like that. Only some foreign object being rammed, once per revolution, into the cylinder head could generate the brutal combination of tone, frequency and location we were hearing.
What Alex apparently didn't know was that this was clearly impossible. Any foreign objects that managed to slip past the air filter and through the turbocharger would clearly get stuck in the intercooler before ever making its way to the cylinder. There was just no way.
In addition to the noise, the engine had the rough-running lope of a 3-cylinder or a Subaru. Unplugging one spark plug wire at a time, we quickly found that nothing changed when we unplugged #1. Clearly that was the dead hole. Pulling the sparkplug showed Alex was clearly right, but failed to explain how. The ground electrode of the spark plug was bent down, touching the center electrode and leaving no gap for a spark to fire. Something had clearly hit it, but how could something big enough to do that get through the intercooler?
"Doesn't matter," yelled Alex. "I'll just go back out and keep driving it. Whatever it is will find its way out."
With five engineers on the team, that flagrant disregard for data would not do. We pulled off the intercooler instead and shook it. Nope. Nothing loose in there… Then, during reassembly, someone grabbed a flashlight and happened to see this:
Only later would we learn that it's "common knowledge" that extensive high-rpm operation has a tendency to cause Miata throttle shafts to crack like this, sending their screws down into the engine. Common knowledge…
"Great!" said Alex, "now can I drive it and shoot the screw out?"
Nah. That might fuck up the turbo, and where are we going to find another free Mexican Dodge Stratus turbo? Instead, we poked around in the cylinder with a magnet, which failed to do anything. Then we pulled out the #1 spark plug, unplugged the injector, started the engine on the other three cylinders and revved the everlovingcrap out of it. The sound was horrible, and the jet of air above the spark plug hole was impressive, but the screw stayed stubbornly in place.
Finally, 45 minutes after he pulled off the track with a commanding lead, we let Alex back on track with a screw still bouncing around in the cylinder and our 1st-place position a distant memory. 30 seconds later he radioed in that the sound was gone and the engine was running strong. Guess he should never have come to the pits in the first place…
About 8 hours later, I'm fed, caffeinated, recently peed, wrapped in Nomex and ready to race. As Jay Kavanagh rolls in at the end of his second 2-hour stint, he mentions that the brake pedal has been getting longer, and the car is making some kind of grindy, rubbing noise in left turns. This is obviously a wheel bearing going south, since those are the exact symptoms I felt in the middle of the night exactly one year ago at this same track when our last wheel bearing failed. The long hours have turned Jay's mind to oatmeal, though, so he considered it an intriguing mystery and never bothered to radio in a warning.
Last year it was our first bad wheel bearing, something we were completely unprepared for. Since we started discussing the problem over the radio last time, we had time to buy a spare bearing, already re-packed with Red Line grease, from another Miata team in the middle of the night, and plan and strategize the whole wheel bearing-changing pit stop over the radio without losing any track time. Thanks to this preparation, last year's repair was handled in a pit stop where we managed a driver change, fuel, and a new wheel bearing in just 7 minutes, never falling out of the lead.
Thanks to oatmeal-for-brains, we had no time to plan this one, and we still weren't smart enough to bring our own spare bearing. Luckily the bearing change itself was muscle memory from last time. "Bitter, grab a jack!" "Sara, find the 29mm socket and breaker bar!" "JayKav, go beg for a bearing from Emilio!" Break the lugs loose while Bitter jacks up the car, put the tire under the wheel well behind the brake, 14mm impact wrench on the two caliper bolts, grab the screaming hot rotor and caliper with Nomex-gloved hands, keeping the caliper on the rotor to save reassembly time, and set it on the tire to keep the brake hose happy.
"Screwdriver!" The bearing dust cap is too tight to get a screwdriver under its lip. Give up and puncture the cap with the screwdriver, prying the cap off. Sara shows up just in time with the 29, break the bearing nut loose, run it off with the electric impact, yank off the hub. The bearing is so destroyed, the inner race sticks to the spindle and balls explode everywhere. Luckily, the race moved about a quarter-inch before stopping, so there's room to jam a screwdriver behind it and pry it off.
As soon as the race his the ground, JayKav sprints up with a new bearing from 949Racing, who is running exhibition class in their Miata and using this race as a test session. Slide it on the hot spindle, ram the nut home. "Someone find me a torque spec!" 160 lb-ft. "Hammer!" Stake the lock ring on the nut with the screwdriver, hammer home the dust cap, slide the brake assembly back on, 14mm bolts, wheel, torque, let's race!
Even with the complete lack of preparation, the entire stop takes less than 9 minutes.