legacycentral bbs

If it felt like fuel cut....it probably was fuel cut due to over boosting. I don't know if Rob has fuel cut elevated for all setups, eliminates it, or adjusts it to each ECU. Hopefully others can chime in with more knowledge on the specifics, but if it felt like fuel cut, it probably was.

The 550 tune is supposed to not fuel cut unless you specifically ask for a custom one.

I've felt what I thought was fuel cut after my tune. However, after I while I notice it was always when I was under hard accel in a right turn, even wide sweepers that don't pitch the car hard could do it. After some digging I found it was very likely fuel pump starvation due to our split tank setup. I was gonna do the plastic jug around the fuel pump mod, but had to store the car before I got a chance.

The odd thing is I've auto-X'd my car for a few years and I never noticed a problem with it until my Robtune. I don't know if its the huge increase in torque that allows it to easily throw all the fuel to the wrong side faster than the jet pump can move it back over, or the increased flow of the 255lph pump in comparison to the jet pump, or what (suspension stayed stocked and always ran all season tires).

Not saying yours is that, bit I am saying pay attention to any variables with how and where you're driving that are the same when it happens. Because before I noticed that pattern I was positive I was hitting fuel cut even though we don't (shouldn't) have one.

Josh, my gauge says I peaked at 19 so if the tune is working right it shouldn't happen. And it is sometimes a freak thing, usually I can WOT and no problems.

I also noticed the car hates driving cold. Unless the car is full hot, the throttle is super jumpy and jerky. Anything but smooth..

The fuel cut happened when driving straight so I don't suspect it to be a fuel delivery issue. The knock light was flickering so I'm just wondering if it pulled timing super hard or what. But the car seems to run and drive strong.

Isn't fuel cut a semi-power loss event? Not a total fuel cut off?

From my experience overboost fuel cut is a total fuel cutoff. Back when I legitimately would hit after my exhaust/MBC/TD04 install, the car would abrubtly lose all power for a second and then go back to normal. Its nothing like the cushy cutoff you get when bouncing the red line.

I thought that fuel cut was the same thing as redline fuel cut. Just prevents you from going too far. I could be wrong though.

-Boost fuel-cut (when the ECU cuts fuel due to excessive boost levels) is different as James mentions.
-I 'thought' Rob's default was to leave the 19 psi boost fuel-cut ON*, unless requested to turn it off.

Did Rob leave it on or off for your 550 tune ??

For the year I ran the 550 tune I hit something akin to fuel cut twice (I had Rob remove the 19psi boost fuel-cut on my tune - I think). I ran 18psi max, so it's possible either my car DID have the 19 psi limit still in place or there was something else causing the fuel-cut.

Edit: See below, in both cases I was also doing a right-hand turn @ 'speed', w/less than 1/4 tank of fuel...I remember thinking "Did I just run out of fuel?"

Td


*From what I understand, 19 psi is the max our 2 BAR MAP sensor can handle. If exceeded it also does a fuel cut or similar. To over ride this, Rob turns it off/removes its input to the ECU. You can then boost as much as you want or your engine can handle.

Maybe it changed depending when you got his tune, but when I got mine he said the fuel cut was completely off on my 550 tune unless I asked for something different. I also noticed there was a version number written on my chip when he gave it to me, so I wouldn't be surprised if little things like that have changed over time.

I too think it's on unless requested off for safety and the buyer decides what psi to set it at if they want it.

I bought it used from another member, so it's possible that if he set the fuel cut for over 19psi than I may have just hit fuel cut. But that still doesn't explain the random bucking and popping when I get on the throttle sometimes. It's more rare but It happens

Does the bucking & popping only occur when the engine is cold? YOu mentioned the throttle is super jerky when cold.

No, I never even boost before the temp gauge is in the middle. The throttle is super weird when cold and gets almost all the way better when hot but I an still feel it. Honestly I've had to relearn my car after the RobTune

Clearly that is not normal. When my old SS was healthy, it was as smooth as the stock 22t ecu, and I only saw the flicker on long hard pulls.

I may have missed it, but have you emailed (not pm'd) Rob? This is a lot of speculating we're doing, without hearing from the man himself.

I took a video at night so you can see the flicker. I know it's out of focus I'm sorry! I went out and after the car was hot I gave it a nice pull. The car cut out real hard and it says I only peaked 18 that time, and I wasn't into the gas very long. After that one time (first pull) all the other pulls were fine. It never did the cutting out. AFR's always in the mid 10's. Wondering if maybe my MAF sensor is to blame.


https://vimeo.com/115242643

Hi all,

Here is my first attempt at what will eventually end up on my FAQ for how my knock light works. Let me know if you see anything missing, incorrect, or just plain confusing.

-Rob

Robtune Knock Light behavior

My custom knock light code is triggered off of a specific software flag in the subaru knock detection algorithm. This flag is triggered only when the ECU has determined it believes that actual engine knock has occurred. At this point the ECU has a few options to how to handle the knock. It could do nothing, change the long term knock correction, drop the IAM, and/or immediately retard ignition timing. It is also important to note that a side effect of my knock code is that when there is a real CEL error, the CEL will flash very fast to the point it almost looks dim. If you see this, you need to check your CEL codes and fix and appropriate. With the CEL flashing like this you will get no knock indication.

Now the question is, what to do if you see knock activity. It is always good safe practice to immediately let off the gas when you see knock activity and plain imperative if you see more than a couple flickers at a time to prevent engine damage. These are the logical steps that I suggest for tracking down the source of the knock light activity.

1) It is important to understand that occasional and brief knock activity should be considered normal. The ECU will continually try to reach maximum ignition timing. It will either max out all possible positive adjustment or hit some areas of knock that will affect the LTKC map (localized correction) or the IAM (global correction). This will be more obvious with lower octane or a motor with higher knock tendencies.

2) To determine if you are having real knock issues, you have to judge that you are not having "occasional and brief" knock activity. Obviously this isn't always easy to figure out. I suggest looking for patterns of when you see the knock activity (the occasional part) and how long the knock activity lasts (the brief part). If you can reproduce the knocking at the same load/RPM consistently over a couple tanks of gas then I would not count it as "occasional". If you get more than a few knock light blips in a row, then I would not count them as "brief".

3) If you think you have real knock issue, then the first step is separate false knock from genuine engine knock by characterizing under which conditions you are seeing the activity. That will lead to understanding the best course for correction. Now we need to talk about false knock. My knock code is triggered after all the Subaru hardware and software filtering, but even with that, false knock gets through. False knock is most typically due to mechanical vibration, faulty or loose knock sensor or bad sensor wiring (shielding or grounding). If you see knocking in low load situations (free rev, or low load), then you can almost guarantee it is false knock. Also if you see consistent knocking at the same RPM even after continuously turning the boost down eventually to WG level, that is also a good indication of false knock. I have also recommend running the current tank empty and putting in 1/4 tank of race gas. If the knock activity continues (again after systematically turning down the boost) then you are seeing false knock.

4) Eliminating false knock, then I would say there is a knock issue to track down asap. There are many many things that can cause real knock issues. Although, not an exhaustive list, here are the most common ones I have seen;

• Lean AFRs: There are also many many reasons why this can happen. In terms of troubleshooting a knocking issue, identifying a lean condition is the first step. You will need a good WB02 setup with logging capability.
• MAF issues: The ECU depends on the MAF to determine engine load. If the value seen by the ECU is incorrect, then it could be applying ignition advance values from the wrong part of the map
• High CR: My tune is based on a factory 8.5:1 CR. Anything higher either from the wrong motor (late model ej20G motors) or based on component choices during an engine build, all bets are off. Another issue is excessive carbon deposits in the combustion chamber.
• Fuel issues: My tune is based on 92 octane. There is enough adjustability to run on 91, but below that, again all bets are off. Also it is possible to get low quality fuel. In this case, knock light activity should change with the change in fuel type or source.
• Ignition issues: Incorrect or failing plugs and a failing ignition system can also cause knock issues.
• Poor intercooling: Basic Ver1/2 WRX intercoolering should be enough.
• The tune: Given the number of hours run on my current tune by both myself and my hundreds of customers, I feel ok saying the ignition maps should be considered lowest likelihood for causing knock issues. It is possible for something to go wrong with burning the chip, either human error or some kind of data corruption.

Yep this is a real problem. I just had another customer with this issue. I ran into this years ago during rallyx. Luckily my Defi fuel pressure gauge showed me exactly what was going on. I don't think I had the problem running the stock ej25. The difference is a much higher fuel demand (pump and injectors) as well as likely carrying more speed (more corner Gs).

Best info on the subject here;
http://www.rs25.com/forums/f128/t49108- ... -help.html

-Rob

James614 wrote:The 550 tune is supposed to not fuel cut unless you specifically ask for a custom one.

I've felt what I thought was fuel cut after my tune. However, after I while I notice it was always when I was under hard accel in a right turn, even wide sweepers that don't pitch the car hard could do it. After some digging I found it was very likely fuel pump starvation due to our split tank setup. I was gonna do the plastic jug around the fuel pump mod, but had to store the car before I got a chance.

The odd thing is I've auto-X'd my car for a few years and I never noticed a problem with it until my Robtune. I don't know if its the huge increase in torque that allows it to easily throw all the fuel to the wrong side faster than the jet pump can move it back over, or the increased flow of the 255lph pump in comparison to the jet pump, or what (suspension stayed stocked and always ran all season tires).

Not saying yours is that, bit I am saying pay attention to any variables with how and where you're driving that are the same when it happens. Because before I noticed that pattern I was positive I was hitting fuel cut even though we don't (shouldn't) have one.

Rob, if you take a look at my video, what do you think is happening?

I'm going to replace my knock sensor. I see the knock light flickering in free rev sometimes here and there. Also when my car scrapes the ground, it will flash.

I can lug my car up a hill in 3rd and I never see the light. Only when in boost and past 3k RPMs like in the video.

robertpaige wrote:Rob, if you take a look at my video, what do you think is happening?

I'm going to replace my knock sensor. I see the knock light flickering in free rev sometimes here and there. Also when my car scrapes the ground, it will flash.

I can lug my car up a hill in 3rd and I never see the light. Only when in boost and past 3k RPMs like in the video.

Hard to tell from the video. It's like putting a single point on a graph and asking what the trend is. At the very least it sounds like you have mechanical induced false knock. Try to figure that out and see if it goes away. Also you can turn down the boost and see if it gets better.

I too have false knock. Can't figure out where it's coming from.

My light does flash in free rev. I'm going to replace the sensor, and check the wiring as well. I don't believe I am seeing real knock, the car has more than rich AFR's and does not seem to pull any timing. The the fact it flashed when I scraped on something the other night leads me to believe that it is vibration based. Which makes me feel a bit better.

Robert,

The stock knock sensor grounds through the engine block. I would recommend checking and maybe adding a secondary ground from the block (not intake manifold) near the knock sensor back to the battery's negative terminal. Even if you don't keep this, I'd be curious to see what the results would be.

I think there may be an extra bolt hole on the bell housing, or you could add another wire at the stock ground point where the starter bolt is.

rob wrote:Hi all,

Here is my first attempt at what will eventually end up on my FAQ for how my knock light works. Let me know if you see anything missing, incorrect, or just plain confusing.

-Rob

Rob,

Here are my edits/suggestions. Feel free to use all, some, or none of them as see fit


Robtune Knock Light behavior

My custom knock light code is triggered off of a specific software flag in the Subaru knock detection algorithm. This flag is triggered only when the Engine Control Unit (ECU) has determined it believes that actual engine knock has occurred. At this point the ECU has a few options on how to handle the knock. It could do nothing, change the Long Term Knock Correction (LTKC), drop the Ignition Advance Multiplier (IAM), and/or immediately retard ignition timing. It is also important to note that a side effect of my knock code is that when there is a real Check Engine Light (CEL) error, the CEL will flash very fast to the point it almost looks dim. If you see this, you need to review your CEL codes and fix the identified issues. With the CEL flashing like this you will get no knock indication.

Now the question is, what to do if you see knock activity. It is always a good safe practice to immediately let off the gas when you see knock activity and (removed the word plain) imperative to prevent engine damage if you see more than a couple flickers at a time.

These are the logical steps that I suggest for tracking down the source of the knock light activity: (I moved this sentance out of the previous paragraph)

1) It is important to understand that occasional and brief knock activity should be considered normal. The ECU will continually try to reach maximum ignition timing. It will either max out all possible positive adjustment or hit some areas of knock that will affect the LTKC map (localized correction) or the IAM (global correction). This will be more obvious with lower octane or a motor with higher knock tendencies.

2) To determine if you are having real knock issues, you have to judge whether you are not having "occasional and brief" knock activity. Obviously this isn't always easy to figure out. I suggest looking for patterns of when you see the knock activity (the occasional part) and how long the knock activity lasts (the brief part). If you can reproduce the knocking at the same load/RPM consistently over a couple tanks of gas then I would not count it as "occasional". If you get more than a few knock light blips in a row, then I would not count them as "brief".

3) If you think you have real knock issue (removed comma & then) the first step is to separate false knock from genuine engine knock by characterizing (removed the word under) which conditions you are seeing the activity. That will lead you to understand the best course for correction. Now we need to talk about false knock. My knock code is triggered after all the Subaru hardware and software filtering, but even with that (removed comma) false knock gets through. False knock is most typically due to mechanical vibration, a faulty or loose knock sensor, or bad sensor wiring (shielding or grounding). If you see knocking in low load situations (free rev, or low load) (removed comma & then) you can almost guarantee it is false knock. Also, if you see consistent knocking at the same RPM even after continuously turning the boost eventually down to WG level, that is also a good indication of false knock. I have also recommended running the current tank empty and putting in 1/4 tank of race gas. If the knock activity continues (again after systematically turning down the boost) then you are seeing false knock.

4) After you have eliminated false knock (removed comma & then) I would say there is a real knock issue to track down as soon as possible. There are many many things that can cause real knock issues. Although (removed comma) this is not an exhaustive list (removed comma) here are the most common ones I have seen:

• Lean Air Fuel Ratios (AFR’s): There are also many many reasons why this can happen. In terms of troubleshooting a knock issue (removed comma) identifying a lean condition is the first step. You will need a good Wide Band (WB) oxygen (O2) setup with logging capability.
• Mass Air Flow (MAF) sensor issues: The ECU depends on the MAF sensor to determine engine load. If the value seen by the ECU is incorrect, then it could be applying ignition advance values from the wrong part of the map.
• High Compression Ratio (CR): My tune is based on a factory 8:1 CR. Anything higher either from the wrong motor (late model ej20G motors) or based on component choices during an engine build and all bets are off.
• Excess Carbon Deposits: Another issue is excessive carbon deposits in the combustion chamber which can lead to hot spots where the air-fuel mixture can ignite prematurely.
• Fuel issues: My tune is based on 92 octane. There is enough adjustability to run on 91, but again below that and all bets are off. It is also possible to get low quality fuel; in this case (removed comma) knock light activity should change with the change in fuel type or source.
• Ignition issues: Incorrect or failing plugs and a failing ignition system can also cause knock issues.
• Poor intercooling: Basic Ver1/2 WRX intercoolering should be enough.
• The tune: Given the number of hours run on my current tune by both myself and (removed my) hundreds of customers (removed comma) I feel ok saying the ignition maps should be considered the lowest likelihood for causing knock issues. It is also possible for something to go wrong with burning the chip, either human error or some kind of data corruption.

That's great. I think the addition of the knock light FAQ is really a good thing.

Update: Had the fuel cut issue again (at 18psi), crusing in 3rd at around 3000ish rpms, light turned yellow so I hammered down the gas, got about to 4500 and the car cut out super hard. Then did another pull and the car was fine... I then lowered the boost, looks like 12-13psi is where i'm at in 4th gear now. The knock light has definitely calmed down, but still present.

Ugh

Legacy777 wrote: Rob,

Here are my edits/suggestions. Feel free to use all, some, or none of them as see fit

Thanks for the review and suggestions Josh. I hope you didn't run out of red font For my first attempt I was aiming to cover all of the technical details and make sure it made sense. I didn't see any edits there or have any questions so will assume it is clear.

robertpaige wrote:Update: Had the fuel cut issue again (at 18psi), crusing in 3rd at around 3000ish rpms, light turned yellow so I hammered down the gas, got about to 4500 and the car cut out super hard. Then did another pull and the car was fine... I then lowered the boost, looks like 12-13psi is where i'm at in 4th gear now. The knock light has definitely calmed down, but still present.

Ugh

Robert,

Do you remember who you bought the ecu 2nd hand from? I can look up to see if they kept the boost cut.

-Rob