Subaru WRX Forum banner

04 WRX power loss

8K views 23 replies 4 participants last post by  ms3wrx 
#1 ·
I just installed a VF39, STi pinks, STI top mount, and Walbro 255 fuel pump. It also has COBB intake, unequal length header, catless up pipe, catless 3 inch downpipe and COBB exhaust. The car pulled hard, but had trouble starting and ran a little rough. I thought the injectors were the problem, so I sent them off to be tested and the came back ok. While the car was apart I installed a Tomei turbo inlet, since the stock one had gotten soft. When I put the car back together I lost a lot of power. Just entering datalogs into virtual dyno as a reference, I lost around 50hp. I don't have any trouble codes.

The problem is when the hits around 5,000 RPM, it stops pulling hard. It's almost like the car is at part throttle. It doesn't misfire or hesitate. Just looking over my datalogs, the timing is advanced a lot more than before and the MAF g/s is much lower. The MAF g/s actually fluctuates some and stops increasing. I have some datalogs attached and I was hoping to get some insight.

I have done the following diagnostics:

Compression test. right side 120-120 Left side 130-130. Low, but consistent.
Had injectors tested
Changed plugs
Boost leak test
New crank position sensor. This actually fixed the hesitation I was having.
New fuel pressure regulator
All new vacuum lines
Checked for vacuum leaks
Reflashed ECU
Checked for exhaust leaks
Checked fuel pressure. 43psi at idle
Cleaned MAF and swapped MAF with one that is known working
Verified the throttle plate is opening 100%
Verified TGV's opening

Any help is appreciated.
Thanks.
 
See less See more
#2 ·
You are tuned for this correct? Cobb AP? Open Source?

I am sure you already know this but just in case.... ANYTHING that changes air flow (turbo, turbo inlet, intercooler, intake manifold, etc), exhaust flow (headers, up-pipe, down pipe, etc) or Fueling into motor (Injectors, Aftermarket FPR, etc) then you need to update your tune. With the mods you have this needs to be Protuned.

When you go to a aftermarket intake (pre-turbo) the volume changes form what the car thinks, Your headers changes the flow to the turbo which changes the spooling characteristics, the downpipe allows for less back pressure and better flow again changing the spooling characteristics, Injectors flow more at the same pulse as the stockers, That turbo is larger than what was there, the intercooler has better cooling eff...

OH and did you remove the restrictor pill? Are you trying to run a manual boost controller? Or stock boost controller?

I am hoping your reply is that you have some sort of protune for all of this... if you don't stop driving your car and do so... get an E-Tune at the very least!
 
#3 · (Edited)
It has the COBB VF39 tune, the restrictor pill for the VF39 is in place, and it is running the stock ebcs. It pulled hard before I installed the inlet, so if the inlet isn't compatible with the COBB tune, I have no problem getting a protune, but I don't want to waste the money if the car has a problem.

Here are a couple 3rd gear WOT datalogs.

https://docs.google.com/spreadsheets/d/1g5M-7lCTS0bhkGLlHctTOLZ4dE_XtEtHvDngMfW7_R4/edit?usp=sharing
https://docs.google.com/spreadsheets/d/13tuMrbzTJZe5dyHkKItz0MLIm08bv8bP8NZXy1uI1vE/edit?usp=sharing

Here is a datalog from before the inlet install

https://docs.google.com/spreadsheets/d/1Xf6Q8ZhnXr_R1-KmPN6s1RSxBE0CUEsMR_2_fP315JA/edit?usp=sharing

Also my fuel pressure is is pulsing when I check it. I made a short video of it. I have a new Walbro 255 in the car with about 2000 miles on it and a new OEM regulator.

https://youtu.be/4J6aUKNCKSU
 
#4 ·
Your headers and turbo inlet are the only things not compatible with that OTS Map. The other thing is that some OTS maps just really don't like some cars.

But that inlet is allowing more air to flow better to that turbo. This will cause it to not be able to calculate accordingly. Your car seems to be doing just fine from your logs from what I can see but I am no expert.

One more thing though is that you log shows 11.25 in your AFR cells. I have found that our front o2 maxes out at 11.14-11.25 and can't read any lower. Do you have a wideband on board? If not you should. When my cobb AP shows 11.25 my wideband might get as low as 10.1ish. If you are getting that low or even lower you could be running way too rich which would make it feel like you are losing power.

Otherwise you Boost is tapering off like I would think at that RPM. Really the OTS Maps don't pull much past 5500. That's where getting a protune really unlocks the potential your car has. Cobb has to be safe with the OTS and it leaves plenty of room for Tuners get more HP out of your car.

If you don't have a boost leak, fuel PSI on point, compression test is good... I personally would install some 1 step colder plugs and schedule a protune.
 
#5 ·
What do you think about the MAF voltage readings past 4,000 rpm? I have an Innovate wideband and it is actually reading leaner than the onboard o2. Its showing low 12's and high 11's. I just calibrated it yesterday, so it should be accurate. I'm changing out the PCV valve and related hoses right now. It all appears to be original and leaking.

Thanks for the replies. Sometimes you just need to talk to someone to get your mind back on track.
 
#6 ·
If you have about 100k on the OEM o2/MAF you should replace them soon. I did the MAF and my FBK dropped in general, and the front o2 helped my MPGs. Have you checked the TMIC-to-manifold connection for oil? Maybe the VF if starting to leak, which would cause your MAF to get nasty coated with oil vapor and overtime preventing accurate readings at high RPM, as well as your front o2.
 
#7 ·
The front o2 sensor has around 10k miles on it. The MAF is original with 150+k on it. If I change the MAF, how long will it take for the MAF voltage to be correct? I swapped MAFs and I was getting the same readings, but I didn't drive far. I figured MAF voltage would change immediately and not need to be learned. The turbo does appear to be leaking, which sucks because I just put it in and it looks relatively new. The problem persisted after I cleaned the MAF though.
 
#8 ·
Just looking at your logs, your car is running very well. That does not mean that it behaves as normal on your end, just based on looking at those logs.

I do not see any issues in them at all, your AFR is just a little bit on the lean side, but those are similar numbers to what I run and my setup is able to run it knock free with my WAIC.

Your ignition timing is likely higher simply from environmental factors. It isn't a bad thing so long as Fine Learning Knock is not introduced and DAM (IAM) stays at 16 (16 bit ecu) or 1 (32 bit ecu)... you're on a 16 bit.

I don't believe having a turbo inlet installed would make your car lose power. The intake itself drastically changes how the MAF reads the value due to the shape and size of the intake tube itself and where/how the MAF is installed. I am not so sure that an after market turbo inlet would make you hit a wall in power or make that much of a difference at all.
 
#9 ·
The car runs great. I haven't had any driveablitiy issues since I changed the crank sensor. I just don't understand how the MAF was reading 254 g/s and 4.48 volts before with a smooth increase, but only 227 g/s and 4.3 volts now with choppy readings after 5,000 RPM. I know part of it has to do with ambient temp, but that doesn't explain the inconsistent readings.
 
#10 ·
You said the turbo is leaking? Where, the inlet, the compressor/turbine seals? If your compressor seal gives out you will notice oil in you TMIC/post turbo intake. If so then your front o2 and MAF are getting oil film on them causing inaccurate readings. Let is go too long and the car won't start, ask me how I know :p

As soon as you change your MAF it will see the new readings, but using a sensor not from Subaru has been proven to been slightly of on voltage and should get a retune to compensate, the voltage difference as around 0.05v at the high side IIRC from the OE MAF.

Do you still have the EGT sensor installed in the UP? If so you should see what it's readings are when the issue happens, too hot/cold and that can also help determine said issue.

MAF readings are usually only accurate between .5v-4.5v. That .v5 tolerance is for sensor change learning and other ECU factors.
 
#11 ·


Look at this portion of the datalogs. The before datalog was the day after I installed the turbo. The after was done today. I have driven around 1500 miles between these 2 data logs and changed the turbo inlet pipe and there is approximately 30 degree Fahrenheit difference in ambient temp: the before logs being the lower temp. All the other data in the logs is normal, so lets just remove it completely and focus on what appears to be a problem.

Full before datalog:
https://docs.google.com/spreadsheet...N6s1RSxBE0CUEsMR_2_fP315JA/edit#gid=703099106

Full after datalog:
https://docs.google.com/spreadsheets/d/15UCD8OwqLJQkHjHcucCB9eEF1Kk_Hnsdr6z-yfkiw7k/edit?usp=sharing

The turbo is holding pressure in a boost leak test, but there is oil in the intercooler and the silicone coupler to the throttle body. The new turbo appears to have a leaking oil seal unless the PCV valve was allowing oil to get in the turbo. The PCV valve has been replaced, so I will have more incite on that when I pull the intercooler off next time.

Would pulling the front o2 and cleaning it be a bad idea? I have heard You shouldn't try to clean them.

The second MAF I'm using is an OEM Mazda part out of my Mazdaspeed 3. The readings I'm getting are consistent between the 2 sensors.

I still have the EGT installed. I can check those numbers with my next logs.
 
#12 ·
You should pull the TMIC and rinse it with distilled water to make it easier to see if the turbo is leaking, my bet is it is leaking through the compressor seal. You CAN sometimes take some MAF cleaner and carefully from a few inches away spray the front o2, just do it in short pulses as you rotate it. While this may work, I'd still get a new front o2 and keep the other as a spare. As for the MAF, try a new one from amazon, as long as it's the same P/N and it's a Denso you should be okayish (I'm using an Oriellys Denso MAF) but it did need a slight map tweak for the voltage changes I mentioned before, nothing too drastic.
 
#13 ·
I appreciate the help, but I can't justify spending $240 on parts that are testing good. There is no oil residue in the pre-turbo intake, and the o2 sensor has less than 10k miles on it. Surely the car would be smoking out of the exhaust if the turbo was leaking enough oil to foul the o2 within 400 miles.

Is there any possibility the fuel pump could be causing these issues? I have been suspicious of it since I installed it. I have been experiencing long crank times and the fuel pressure is pulsing when I test it. The fuel pressure gauge needle is flickering.

See this video of the fuel pressure:
https://youtu.be/4J6aUKNCKSU
 
#14 ·
A bad PCV won't let that kind of oil into the intake as they 99% stick closed. As for the front o2 being bad, I never said it was bad just coated in oil if the compressor side seal is leaking, which introduces oil into the combustion process and cannot all be burned, therefore coating your spark plugs, front o2, and MAF (less than the rest but still).

How do I know this? I had my TD04 leak through the compressor seals and not only did it leave a streak of oil in my intake but within 1000miles (or maybe less) the car stalled while I was pulling out from a red light and getting into 4lbs of boost. It stalled due to the front o2 having oil coated on it and it couldn't get readings within the allowable parameters so after the ECU couldn't find a way to fix the issues, so it stalled from not being able to burn that excess oil. After a day of head scratching and testing other issues, I swapped the front o2 and within 5 seconds of cranking, bam...engine cranked and I drove it up the road to tow it home and swap the turbo.

As for your fuel PSI, that pinging can be caused by your injectors creating a low pressure each time they fire and the gauge not being dampened. Did you hardwire the pump? The ECU uses PWM to control the pump normally to increase pump life by not running at 100% all the time. And as much as I hate to say it, Walbro 255s are hit or miss sometimes. Not only do they occasionally show up DOA in the box, but many ebay sellers are listing the as real 255s when they's knockoffs for a similar price. Even my real 255 from a Subie builder was DOA, but they warrantied it like a fat kid eating cake.

Remember the bigger the pump, the shorter it lives. Even the Bosch 055s are known to die early, that's why people keep a spare with them when they run aftermarket pumps.

Waiting for more info/answers.
 
#15 · (Edited)
I could blow straight through my PCV valve in both directions. I think it was just worn out. It appeared to be the original part. I replaced the valve and the hoses. I haven't had a chance to pull the front o2 and inspect it.

My fuel pump had a bad check valve. I didn't know about the knock off Walbros until after I purchased mine, but as far as I could tell after I pulled it back out, it was authentic. It's starting normally again, but had no effect on the power loss I have been experiencing.

I'm really starting to think the turbo inlet is causing MAF scaling issues. I emailed the guy who tuned my Mazda asking to look over the datalogs. I'm hoping I can get him to retune the car and get everything sorted out. I'll update this post when I hear back from him. It usually takes a couple days.

I just remembered that I dropped a wrench or ratchet and it hit directly on my front o2 sensor when I was putting in the injectors. I wonder if I damaged the o2 sensor. I know the sensor I replaced worked fine until we pulled it off to get the heat shield off the manifold during the up pipe install. After pulling it out and reinstalling it the heater circuit went bad and it eventually stopped working all together, so they seem kinda fragile. I'm getting a "hunting idle" condition where the afr, vacuum, and RPM are fluctuating and people are saying that is a sign of a bad o2 sensor.
 
#16 ·
Yeah a failing check valve combined with a shot front o2 will definitely cause a wandering idle and other drive-ability issues. Swap in a new front o2 and pump and go for a 20 minute drive. It won't want to start normally at first until the new o2 is noticed but it will crank.

And get yourself a subie tuner lol

Keep us posted!
 
#17 ·
I unplugged the o2 sensor and let the car idle and it idled smooth and my AFR reading on my wideband was steady. I think you might have been right Robo. I'm going to order a new o2 off Amazon so I should have some new info in the next couple days. I think dropping the ratchet on the sensor might have damaged it internally. I noticed on the drive home today it's about 1 point off when comparing it to my wideband.
 
#18 ·
The OE front o2 will be about 1afr off compared to a wideband, especially in boost. It's not a true wideband but not narrow either. Though I have issues with my rear o2 failing every year. Lmk how it goes.
 
#19 ·
I got the o2 sensor changed and it didn't help. The sensor I replaced still looked new. I unplugged the MAF and started the car and it idled smooth, so the car just idles better in open loop. Does anyone know what inputs the ECU reads in closed loop at idle? The o2 sensor and MAF are good, so what else could it be? I'm not sure if this idle issue is related to my power loss, but I'd like to rule it out.
 
#21 ·
Does the idle issue happen more during a cold start (4-8+ hours of not driving) vs a hot start (less than 4 hours) of not driving?

During closed loop the ECU reads info from other sensors (besides the front o2/MAF/MAP) to determine if changes need to be made. If your IDC or Injector Scaling isn't perfect down low you can have idle issues which is also common with massive injectors, not 565s. Most OTS/Dyno based protunes don't focus in rpms below 2000RPM, and they try to get a decent idle, but if you have other mechincal issues/mods it may not work great. Light Weight FlyWheels (aka LWFW) are known to cause idle/p0300 if not tuned for.

Can you get some logs of idle at both cold start and hot start logs and share them in XML format? 30-60sec logs should do, I'd bet there's some F*ery going on in the map.
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top