2025 Mid Year Review - Impreza STI Hatchback

With the hatchback left to rest during the latter part of 2024 whilst efforts were concentrated on getting the hillclimb car up and running, it was time in early 2025 to solve the hatchback's piston ringland issues and to get the car back up and running.

Naturally there was only one machine shop for the job, and this time it was time to up the ante and to purchase and install a set of 100mm Mahle 4032 grade pistons (which we also sell from our online shop). Following some superb machining from Roland Alsop on the EJ25 block and W25 heads, it was time to assemble the engine from the ground up, with the ARP 2000 head stud kit providing assurance that plenty of pressure was put on the Cosworth head gaskets.

After the engine's careful assembly, it was time to get to the bottom of the hatchback's reluctance to start. After some investigation and sending datalogs to Link ECU's very helpful technical support team, two sensor signals seemed amiss. Specifically, the RH intake and exhaust cam sensors were reporting each other's cam positions! The cause was found after removing the Link ECU and inspecting the wiring harness - somehow during the car's past, someone had swapped over these wires, which were the cause of the car's reluctance to start. With the wires swapped back over, the car started first time (and since has always started on the first try).

Running in the EJ25 engine was an exercise in patience, with the uprated VF30's low down boost necessitating extra throttle pedal restraint to ensure that the running in was completed in a gradual manner. With the 1000 miles over, it was then time to remap the car firstly for the turbo's wastegate pressure, and then for a higher boost setting.

The hatchback was performing nicely after the remap, and it was believed that the Type RA shortblock's problems were solved. A small exhaust manifold or intake manifold leak were believed to be the cause of the slightly later than expected boost response when commanding a higher boost solenoid duty cycle from the Link G4X ECU. The decision was then made to visit the Whifbitz 4WD dyno in Caldicot, to see how much power the hatchback was making, and to try and pinpoint the cause of the lower than expected boost. With the car set up on the dyno, the investigation could begin. With some pulls on the wastegate showing a healthy but modest 237WHP, it was then time to turn up the boost.

Unfortunately, it was at this point when things started to go awry, with the coolant temperature showing upwards of 110 degress after our first pull on the high boost setting. After letting the engine cool down and checking the oil and coolant pressures, it was decided to go again, this time keeping an extra close eye on the coolant temperature. As the engine speed got over 6000 rpm, the oil pressure limiter kicked in and stopped the engine completely. With the coolant temperature at 112 degrees, and evidence of some oil blown from the catch can, things seemed to have taken a turn for the worse! After waiting a few minutes for the engine to cool down, the engine was restarted and after a quick check over, nothing obvious seemed to be wrong, although the event was suspiciously similar in symptom to the broken ringland scenario.

To make sure that all was well, an engine inspection was required, which meant a complete disassembly for the EJ25. Removing the pistons from the block halves showed that no piston ringland failures had occurred as the ringlands themselves were intact. However, the pistons were worn heavily on the ringland areas and were difficult to remove, and required some gentle hammering for them to be pulled from the top of the bores. The main bearings were also found to have been excessively worn, but a positive from the inspection was that no evidence of detonation was found, meaning that the timing and fuelling were likely as commanded by the Link G4X.

Following a consultation with Steve from Roland Alsop, the likely cause of the hatchback's trouble at the dyno was that the bores themselves had overheated, which caused them to warp. This was a bit of a lesson to keep an eye on the coolant temperature during dyno sessions, which hadn't seem to have been a problem when mapping the hatchback on the road. Yet again, a decision needed to be made with the hatchback's future, as the original intent was to sell the car after installing the Mahle pistons. Thankfully, due to the flexibility afforded by the Link G4X, and the strong 6 speed gearbox fitted to the newage STI Impreza models, it was decided to keep the hatchback as a show car, and an engine test bed for Pleiades Performance. With the decision made, the car's first task was to run in a freshly built closed deck EJ20, with a 2.1 stroker kit installed.

This engine was machined and slowly assembled following the disassembly of the hatchback's EJ25. This chapter of the hatchback's story will be continued in the latter half of 2025 so please tune in at the end of the year for an update!