Toyota GT86 Would Not Start After an Engine Rebuild: One Bank Was 180 Degrees Out

Good Compression, No Start? Count 720 Degrees

The Toyota GT86 arrived after an oil-starvation failure and a major engine rebuild. It cranked at a normal-looking speed, occasionally coughed, and produced an odd note through both intake and exhaust, but it would not run. The workshop had already checked compression and mechanical timing. Both had reportedly passed. That combination is exactly why this case is useful: two test results can each be technically true while the conclusion drawn from them is wrong.

Once the evidence justified opening the timing area, a steerable inspection camera was a useful documentation aid rather than the deciding test. A tool such as the AUTOOL SVB308 can help record marks, guides and cramped areas without pretending the image establishes phase. SVB308 can document accessible chain marks, guides and confined mechanical areas after safe access is created; it cannot establish cam/crank phase, see through a closed timing cover or replace an in-cylinder pressure test and the Toyota/Subaru timing procedure.

This was one post-rebuild GT86, not a shortcut for every FA20 no-start. The safe takeaway is to ask whether compression occurred at the correct time across the full four-stroke cycle—not merely whether each cylinder produced a respectable peak number.

The rebuild history belongs at the front of the job

The rebuild changed the prior probability. A no-start on an untouched engine might begin with codes, fuel pressure, spark and immobilizer status. Here, the engine had been out, both cylinder heads and timing drives had been disturbed, and the complaint began on the attempted return to service. That did not prove an assembly error, but it moved synchronization and valve timing high on the list. Stop repeated cranking while fuel washes the cylinders or an unknown mechanical condition remains; a newly assembled engine should not be spun indefinitely just to see whether it catches.

Why good peak compression did not clear valve timing

A relative-compression capture suggested two cylinders were weak. Four individual in-cylinder tests then measured roughly 13 to 15.5 bar, apparently supporting the workshop’s statement that compression was good. The apparent conflict disappears when peak pressure and event timing are treated as separate questions. A cylinder can trap air and make pressure while doing so on the wrong stroke relative to the rest of the engine. Peak readings alone do not describe the sequence in which the events occur.

Count events across the complete 720-degree cycle

A four-stroke engine requires 720 degrees of crankshaft rotation to complete intake, compression, power and exhaust. When all four pressure traces were synchronized to one ignition reference, the GT86 produced four compression events within only 360 degrees. Opposing cylinders were compressing together. That unusual arrangement also explained why average cranking speed looked believable: paired compression loads were followed by unusually free sections of rotation, and the average concealed the wrong event spacing.

Synchronize every cylinder to one reference

Cylinder identification mattered because the FA20 firing order is 1-3-2-4, while some test software at the time assumed a different four-cylinder order. The technician used one ignition trigger as a shared reference and overlaid the four pressure captures. This prevented a software label from choosing the bad bank. Whenever a tool assumes a firing order, confirm that assumption before naming a cylinder. The waveform shape and its physical synchronization matter more than the bar-chart label.

Evidence	What it actually proved
13–15.5 bar on all cylinders	Each cylinder could build peak pressure
Four events inside 360°	The pressure events were timed incorrectly
Intake-cam pattern displaced from known good	A bank-level phase error justified inspection
Right-bank cams 180° advanced	Physical cause confirmed

Compare cam and crank only in their base positions

All four cam sensors were then compared with the crank signal. Variable-valve-timing controls were disconnected so the cams returned to their base positions, and a known-good GT86 supplied the reference that the waveform library lacked. The exhaust patterns looked close enough to mislead because their identifying features repeated every 180 cam degrees. The intake pattern revealed the larger displacement. This was enough to justify the labor of removing the engine and timing cover—not enough to guess which colored chain link had been misread.

Let the waveform earn removal of the timing cover

With the cover off, the left-bank chain marks initially appeared correct. Rotating the engine into the prescribed position for the opposite bank exposed the real problem: both right-bank camshafts were 180 camshaft degrees advanced. An inspection image can preserve this condition for the customer and rebuilder, but the marks must be judged against the factory sequence and crank position. A bright link lining up somewhere is not proof that both banks are timed on the same four-stroke cycle.

Reset the right bank rather than replacing more parts

The right-bank valve timing was reset according to the workshop procedure. Only gaskets and normal reassembly consumables were required for the correction; the evidence did not call for another short engine, sensors or control unit. This distinction matters to the owner. The repair addressed phase, not the engine’s ability to create pressure, and the invoice should say so plainly.

Verify the start, sound and synchronization together

The reported result was a normal start after reassembly. A complete close-out should also compare the post-repair intake and exhaust note, cam/crank relationship, fault memory and oil-pressure behavior. Because the original failure followed oil starvation, that lubrication history remains a separate reliability concern. The timing repair solved this no-start; it does not certify every reused part in the rebuilt engine.

The turning point was not a more exotic sensor. It was asking a more precise question of data already called ‘good.’ If a rebuilt GT86 has compression but will not start, keep the 720-degree cycle visible. Good pressure at the wrong moment is still wrong—and it can look surprisingly normal until every cylinder is placed on one timeline.