Subaru Engine Swap Fuel Trim Issues - Seeking Subaru Gurus!

Joined
Jul 25, 2024
Messages
8
Hey folks,

Wondering if we have any Subaru junkies out here. Long story short - I am a practicing licensed automotive tech and this issue has me at a bit of stand still/brainstorming period for about a year. Not the end of the world - car runs fine, but if I could solve this issue I'd be very happy.

2001 Subaru Legacy GT Limited Sedan
Automatic Transmission
EJ251 (ORIGINALLY)
EJ203 Swapped Engine

My wife's car, typical Subaru EJ251 BS, head gaskets failed, coolant emulsion, I procrastinated it for too long and motor got sludged up pretty badly. No big deal, did some homework on the topic, although a bit of a horsepower loss, I swapped in a JDM EJ203 motor (proprietary to Japanese market - never produced for US market at all). Basically the idea behind this swap is that you get although a less powerful NA motor, a more reliable one, with a marginally larger deck surface for the head gaskets, and the JDM motors had MLS head gaskets VS the failure prone junk they put in the USDM 2.5L NA motors.

Anyways, had to have been the EASIEST engine swap I have ever done in my career, 1.5 hours to have the motor laying on my floor, stripped to a short block. As for the donor motor, basically the only thing you are after here is the long block assembly, the JDM variant has a vastly different fuel rail, injectors, a variable intake manifold, fly by wire TB, MAF vs MAP air metering, etc etc. The basics are that you swap the USDM thermostat assembly, cam/crank gear(s) as the reluctor patterns are different, and a couple other knick knacks, etc. So I did my thing, freshened it up accordingly, new timing components, plugs, etc. Got it in, thing runs perfect, zero issues with drivability, noticably less power but very very happy and it cost me very little money. I have had this engine in the car for a years time now.

After getting some drive cycles completed right after install I started to notice a P0170 - fuel trim "malfunction" DTC that would recur after 2 x drive cycles. Never had this issue for the multiple years I owned the car prior, and have previously logged data and know the fuel trim was never an issue prior. I followed Subaru's service information diagnostic flow path, that takes you to an ending of "replace PCM". Smoke tested everything, verified fuel pressure, A/F sensor, etc etc etc everything is working 100% as designed. I then started to log some data on it, and noticed my LTFT is consistently -20% indicating the ECU is pulling 20% of fuel out of the map to make lambda. This is the case at idle and through the RPM range anytime you are in closed loop. Then I do some digging into this topic and find that many if not all guys doing this swap are experiencing this. Then I think to my self hmmmmm, I just swapped in an engine with exactly -20% displacement relative to the original motor. So this code makes sense, 20% less displacement should in theory require 20% less fuel to be running stoichiometric. The DTC conditions for setting match the data the ECM is showing me (I believe it was a fuel trim in any direction in excess of 18% - pulling from memory here don't quote me). So I talk to a world renowned tuner who happens to live about an hour from me here in Alberta, Canada (Wester's Garage) and he is telling me sorry there is nothing we can do with the ECM to remap the fuel, it's too old. He suggested using the 2.0L JDM fuel injectors, albiet they are top feed style injectors like the USDM ones, they are of a different size and configuration and 100% not compatible at all. We had tossed the idea around finding an injector that would fit this car that is 20% smaller cc than the original 2.5L injectors. The issue I am finding is that the original 2.5L NA EJ251 motor is about the wimpiest thing Subaru ever produced of the era in theUS market and I am not savvy enough on Subaru interchangeability to know what injector would fit what. I am thinking I will have a very difficult time finding an injector SMALLER that these OEM 2.5L ones in the car currently especially ones that would be a direct swap.

Aside from installing a stand alone fuel management system (ridiculous expense for a wimpy stock car hah), I think that finding an injector of 20% less cc would likely work, as presumably the PWM of the injectors would then match the factory injector (when installed in a 2.5L engine) and I would think the fuel trim would come back into some normality or at least the typical +/- 10% area I would like to see it in where it will not trigger a fuel trim malfunction code.

If anybody has any experience with this swap and any ideas on combatting this, please advise. I had thought about an adjustable fuel pressure regulator (OEM one is vacuum reference and does work properly I did test it) however I don't believe this will help me in all areas (at idle as well as throughout closed loop RPM range). I don't think this will be an accurate enough modification to eliminate this issue. If I could tune the factory ECU I could easily solve this problem, and no the JDM ECU will not work either (much different engine management systems used on JDM motor).

Thanks in advance!
 
@Trav will be along with some help soon I bet.. he's really good with injector applications, sizes and numbers

So I get that the Japan market injectors were a different style, but why could you have used them with the Japanese rail and swapped connectors to adapt to your harness?
 
Dude, you might be kinda boned. Have you ever visited the NASIOC site? The people there have done some amazing Subaru things, post the same question with all the same details there.


1763153097849.webp




NASIOC
 
Maybe you could find an adjustable fuel pressure regulator that's able to reduce the fuel pressure.
 
Are you using the EJ203 intake manifold? Will the EJ251 manifold fit? Can you post a pic of the injectors you are using now that you need a smaller injector for, are you using the 251 injectors? There are a few injectors that work that are a bit smaller depending on if they are JDM or USDM. Automatic or stick for the JDM engine?
 
Hey guys, to clarify:

I am using everything off the EJ251 motor including the ECM. Only thing I took from EJ203 was long block assembly. Intake manifolds are dramatically different - EJ203 having a variable butterfly valve in the intake manifold, fly by wire throttle body, MAF + MAP setup etc. As mentioned above by SubieRubyRoo the EJ251 USDM cars have only a MAP sensor, no MAF in this application (for this model year anyways). I don't think the ECMs would be anywhere close to plug and play and would be a massive deal to adapt the EJ203 JDM ECM into a USDM EJ251 as there are a bunch of "extras" on the EJ203 that the USDM EJ251 does not have. Would be a repinning and rewiring operation that wouldn't be worth it I don't think. Also - the fuel rail was very specific and routed in a way it would work with the EJ203 manifold but not the EJ251 I remember that.

I am using the original EJ251 fuel injectors for my application (2001 Subaru Legacy 2.5L) which has an automatic transmission (transmission is a variable for this application) with the EJ203 motor. I would need an injector that is 20% smaller than a EJ251 OEM injector for this car.

I had thought about a manual adjustable fuel pressure regulator but like I say I just don't feel like that would dial it in well enough to combat this issue @ base idle and across the closed loop RPM range.

@The_Eric @slacktide_bitog @fozzdesy2001 @twX @Trav @SubieRubyRoo
 
I see it and that is the problem, the automatic car could use Toyota 3.0 v6 injectors, a 4 cyl would be 2.0 ltr. I will look into a bit tonight, I have not seen one you have in a long time.
 
I see it and that is the problem, the automatic car could use Toyota 3.0 v6 injectors, a 4 cyl would be 2.0 ltr. I will look into a bit tonight, I have not seen one you have in a long time.
Thanks Travis, I’m hoping that my theory of downsizing injectors is the most practical and simple way to achieve my goal and “fool” the ECU.
 
It doesn’t help that the 251 is speed density. But yes, you should be running a 2.0 ECM and 2.0 injectors, IMO.

Cool swap though! Always felt the 1998-2009 or so Subarus were effectively LEGO cars, as long as you weren’t trying to swap N/A cars to turbo’s.
I think the speed density set up is what’s killing me here yep
 
I think the speed density set up is what’s killing me here yep
Speed density, when the engine is moving the correct amount for the program, is a fantastically simple and effective system. The problem comes when you have too much or not enough air for the fuel that’s in the tables.

Maybe you could use ROMRaider or something similar to fix your tables? It’s been years since I’ve messed with those programs, and some are not supported. But that’s where I’d say to look for helpful hints.
 
Speed density, when the engine is moving the correct amount for the program, is a fantastically simple and effective system. The problem comes when you have too much or not enough air for the fuel that’s in the tables.

Maybe you could use ROMRaider or something similar to fix your tables? It’s been years since I’ve messed with those programs, and some are not supported. But that’s where I’d say to look for helpful hints.
Yes 100%. Assuming in my scenario a MAF car would adapt itself, maybe with a different code but presuming the fuel trim would stabilize and learn. I’ll have to see if I can find something that’s supported, I know my tuner I deal with is very very good with 30+ years behind him and he said it’s just too old he can’t touch it. If we could fix the fuel tables this would be an easy resolve for sure.
 
Back
Top Bottom