ECUFlash - 2012 Subaru STI Roughness Monitor count increasing, no noticeable misfire

Hello everyone!

I have a Subaru STI 2012 with the following problem : Roughness Monitor #1 and #2 count start increasing around 3100rpm, in cruise, low throttle. The thing is customer and I can't feel any misfire or hesitation while driving the car on the road. But as the count keep increasing, the check engine start to flash as it should. Car isn't heavily modified, has an aftermarket FPR, 40k kms on the odometer. ECU is already tuned by a reputable tuner.

Here's what we did so far :

-New Plugs

-Swapped Ignition coils with known good ones

-Replaced Crank Sensor

-Replaced Injectors

Customer said his injectors were 750cc, so we replaced them with Bosch 750cc but trims are adding a lot of fuel now so I'll need to rescale them in the tune, but I haven't put the car on the dyno yet. I did a road test anyway and decided to let the close loop strategy work harder than me lol. Don't worry, I'll rescale them when I will do more troubleshooting on the dyno.

Any input on what could cause this? Cam sensors maybe? I'm out of ideas, maybe it'll be better when the injectors are scaled?

I have attached 2 logs, any input is appreciated.

Thanks,

Frank

@Frank - Couple things to look for / at, just to confirm (though you probably already looked into these) the factory injectors on 2012 STI are 565 - did the injector tables look like they had been scaled for 750 from the previous tuner? Was this issue the reason the car was brought in for tuning? If so - how long ago was the original tune done?

If you haven't yet - I would do a leak down + compression test to rule out rings / bearings, if the oil filer has at least 800-1600 km on it I would also recommend replacing the filter and cracking open the old one to inspect the filter / oil for ferrous material IE metal shavings or apparent silvery glitter in oil residue. At the least do the compression and filter if there isn't access to an air compressor - otherwise if that's all been done and with the checklist you mentioned being done, my immediate reaction would be to check the wiring harness for the usual lose-wire / connector culprits that are common in this generation - I attached a couple files that will give you a run-down of where to look as well as another (easier to check) spot that comes up from time-to-time as causing random misfires NASOIC Misfiring Issue - Culprit connector.

Double-check the ground straps as well - you'd be surprised how much a broken strap can cause headaches on the EJ25.

If none of this yields you any success, then take a look at the ignition tables - particularly at how much ignition advance is setup in the 2000-4000 RPM range, and see if bumping down the ignition timing in the problematic RPM range / load a smidge improves the roughness at all - the EJ25 platform can definitely be picky when it comes to ignition timing - especially the last 2 generations of the STI (09+) where the total ignition compensation formula starts getting a little more complex with additional adder tables in Dynamic Advance (or Ignition Advance in your case) - I copied a section from the Cobb Tuning guide that has some really good details about the do's and don'ts for ignition control as well as what to monitor when tuning to ensure any wrinkles get ironed out. I also included an excel doc that has 3 tables, one that compares the stock primary ignition table to that of an average "Stage 2" (Catback 3" + high-flow Cat Down-pipe, Intake, potentially fuel pump and up-pipe or full exhaust manifold) from a 2012 WRX I am working on currently - the tables aren't an exact 1:1 with the STI ignition tables from the factory map, but the same idea behind how much advance to apply across the rpm band at "X" engine load remains the same since both models have essentially the same engine - the STI has slightly different heads and the 08+ WRX intake system is essentially the 05-09 LGT setup with minor changes.

My guess is - if the previous tune was applied a while ago (16,000km+) and this issue recently started to occur, and it was a smooth drive until then - and no modifications have been made since the last tune... then it's likely mechanical in nature - however if the car has less put on less than 1,600 - 3,000 km give-or-take since it was previously tuned, then I would re-evaluate the tune he has on there and start picking away at the ignition/injector/fuel/knock tables, in that order.

Thanks for the reply James, I was suspecting ignition advance too as it seems pretty high, I will try to lower it down a bit to see if it improves. I will also ask the customer about some background info of the car.

Thanks,

Frank

Following up on this matter, retuned the car using Cobb instead of ECUflash and all is good now.