'Somewhere in automotive never-never land, there's a place called TPI Gulch. Walk to the edge, look down, and you'll see heaps of brand-new throttle-position sensors, mass airflow meters, manifold-absolute-pressure sensors, manifold-air-temperature sensors, coolant-temperature sensors, oxygen sensors, idle-air control motors, knock sensors, and electronic control modules. Among skeletons and empty wallets, a gentle stream flows around the discarded heaps of high-tech gadgets. The stream carries the blood, sweat, and tears of frustrated '85-'92 C4 and Gen III F-body owners.
We're not bashing on TPI cars, in fact, we love 'em for delivering massive amounts (330 lb-ft) of low- and midrange power while delivering fuel economy in excess of 18 mpg around town. But you can visit TPI Gulch if something goes wrong and you attempt to repair your Corvette, Camaro, or Firebird without the proper resources. That's what I did when the "Service Engine Soon" (SES) light in my bargain-purchase $6,000 '85 C4 lit up like a roman candle after I pulled its L98 350 to repair a spun rod bearing (read all about it in "Why Ask Y," Sept. '06). Fortunately, my trip into TPI Gulch was a brief one.
The trouble started after my Swiss Army Knife on wheels (open the hood, doors, and hatch for the full effect) was reassembled after the unexpected nuisance-and $657.89 expense-of the damaged crankshaft. In keeping with the low-buck theme, I removed, repaired, and reinstalled the engine myself, but instead of a clean startup and a tire-smoking drive-off, the darned SES light appeared for the first time in my ownership of the car. Something was screwed up again.
But what? Being a good car crafter, when I fixed the crank, I was very careful to completely document the engine removal process with no less than 250 digital pictures so I could put the unfamiliar C4 back together just the way it left Bowling Green, Kentucky, 22 years ago. I practically wore out a pair of shoes running between the car and my home computer screen to review the pictures during the painstaking reassembly process. The problem wasn't in the machine work either. The thing fired right up, and the oil-pressure gauge showed 65 psi. What's more, the 350 sounded great at idle and drove like new at part throttle on surface streets.
But at anything over half throttle, which is where Corvettes come to life, the motor would hiccup, the SES light would turn on, and the engine would turn into a pig. Oh, it'd run up to its 5,000-rpm shift point, but it was laboring and sure didn't feel like 230 net horsepower. Clearly it was in "limp home mode," a self-preservation tactic where the ECM triggers stinky-rich fuel metering and retards the spark curve to prevent damage. Worse, the dash-mounted instantaneous fuel-economy readout showed single digits. With a fresh set of piston rings, I knew the obvious gas wash would quickly dilute the oil, making every mile traveled equal to 1,000 miles of wear. I parked it after two frustrating days of repeatedly disconnecting the battery to clear the codes and defeat the SES.
Having little experience with ailing EFI cars, I searched the Internet and quickly found out that everybody has an opinion. It's the computer, it's the knock sensor, it's the mass airflow meter (MAF), the catalytic converter is clogged, the fuel injectors are bad, it's the oxygen sensor, it's the EGR valve, it's the fuel pump, it's Duntov's ghost. But the thing I heard most was nightmare stories of how TPI cars can nickel and dime you to death. If I followed all the advice, I'd join the parts changers, spend more than a grand on dead-end leads, and end up in Gulchville. I braced myself for the decision to either cut my losses and sell the thing as a fatally flawed parts car or dig in and make the needed repairs . . . regardless of the cost. Having come so far, I dug in. My first task was to read the trouble codes stored in the ECM.
Doing the TPI paper-clip trick, I jumped the A and B terminals of the assembly-line data link (ALDL) under the dash. The SES flashed a 21 (throttle position sensor voltage high), a 34 (MAF sensor reading too low), and a 43 (knock sensor). Wow, how easy is this? I whipped out my wallet and dutifully replaced the knock sensor ($51.95), checked TPS voltage (it was spot on at 0.54 volts), and hoped the MAF sensor was OK, but priced replacements just in case ($500 new, $250 remanufactured). Dry dive, here I come.
When I turned the Vette's plastic-coated key, the L98 sprang to life with no SES activity. That changed as soon as I mashed the gas. Hello, SES! This time the paper-clip trick revealed only a Code 43 knock-sensor problem, the other two codes were not seen again. It was toying with me. I felt like a gambler on a losing streak. Lacking the experience to go any further, I sought professional help from Rich Sjoberg at Corvette Specialists.
I told Sjoberg about the rod-bearing hassle and convinced him that my mechanical work was not at fault. Connecting his Snap-On scan tool, he noted a Code 42 (electronic timing-control problem). Naturally, all the codes I had witnessed refused to reappear. This got Sjoberg curious about the many electrical connections to be found under the C4's clamshell hood. After a thorough examination, nothing seemed out of place, and all 11 ground wires were present and connected. I felt so proud. But the Code 42 persisted. Relying on his instinct to take customer claims like mine with a grain of salt, Sjoberg went back to basics.
On any TPI car, checking the basics consists of confirming minimum idle speed, TPS voltage, and base ignition timing. The Vette's idle speed was correct at 550 rpm, TPS voltage was still spot on, but when Sjoberg checked the ignition timing, he discovered the distributor was installed 10 degrees retarded after top dead center (ATDC) instead of the factory-specified 6 degrees advanced before top dead center (BTDC). The thing was waaay retarded.
The established process for setting static TPI ignition timing after engine disassembly involves positioning the harmonic balancer at 6 degrees BTDC then stabbing the HEI distributor with the rotor and No. 1 tower perfectly aligned. The computer takes it from there once the engine is running and imposes a preset timing curve that's factory calibrated to suit the engine.
Even though I had taken time to make sure the harmonic balancer mark, No. 1 piston, and TDC notch on the timing tab were in harmony-easily done when the engine was out of the car with the heads off-my error was a botched effort at maintaining proper rotor and cap alignment when installing the HEI. I must have bumped the distributor and screwed up the rotor-to-cap alignment during reassembly.
Fortunately, Sjoberg's scan tool detected the ignition problem in real time, and with a quick twist of the distributor housing, baseline timing was corrected. This gave the ECU the proper starting point from which to establish a workable ignition-timing curve. The amazing thing is how the ECU was able to make the best of my lousy distributor installation and that the car ran so well (at part throttle) despite my screw-up. In fact, it ran so well, I totally discounted the possibility of ignition trouble in my description of the problem to anybody who'd listen. It had me fooled. But now the car runs great; no trouble codes flash and life is good. It's time to get back to the job of reclaiming the C4 from the yo-baby gold chain crowd and enjoying it for what it is ... the ultimate budget supercar for the Car Craft masses.
Troubleshooting Tips
* A classic TPI goof is mixing the wires for the knock sensor and coolant-temperature/fan-switch sending unit on the passenger side of the engine. The blue wire goes to the knock sensor, the green wire goes to the temperature/fan sensor. Cross 'em and you'll trigger a Code 42 electronic timing-control problem and/or a Code 43 knock-sensor alert and bring on major driveability problems.
* Sjoberg said, "These computer-controlled cars will do funny things if you don't have clean, tight electrical ground connections." After major surgery like engine removal, make certain all ground wires are intact and properly routed.
* Lots of mysterious TPI problems are caused by old batteries that can't maintain a full charge. Check the battery before you dig in. If the battery cables are bad (checked with a volt-drop test), replace the whole cable. Don't cut the ends off and use clamp-on replacements because the strands deep inside the cable may have resistance-causing corrosion you can't see.
* Scan tools are not the same as code readers. While inexpensive code readers eliminate the hassle of counting the flashes on the SES light, they don't read and display TPS voltage, oxygen-sensor cross counts, manifold absolute pressure, manifold air temperature, coolant temperature, and other helpful, real-time diagnostic data. Code readers don't always identify the root cause of the problem. Take care not to jump to conclusions and randomly start changing parts.
* The FSM is a great alternative to the expense of a scan tool. But when you navigate the FSM's diagnostic trouble trees, you must be certain to follow every instruction. Don't cut corners or skip steps, which could trick you into changing parts you don't need to change.
* If you've thrown a bunch of new sensors at your TPI car and it's still not right, reputable shops like Corvette Specialists offer inexpensive diagnostic services that can save you money in the long run. Don't be afraid to get help.
