I was torn on where to post this because it's technically maintenance, but it's performance-part related and you need to have an AP3 or equivalent to take datalogs for comparison so I'll put it here. This is by no means the 'definitive' guide on troubleshooting, but it should hopefully assist novice owners in being able to isolate and solve issues on their own. By all means, if you have a suggestion feel free to post it and if you're going through the process and are hopelessly stuck reach out to the community and ask questions.
DANGER-Doing multiple pulls with extreme negative corrections can cause catastrophic failure of your engine, so proceed with caution. If you are experiencing excessive negative corrections repeatedly, consider stepping down your tune to something less-aggressive before proceeding. Just because the ECU can pull timing doesn't mean it's optimal for the engine, and pulled timing can be an indication of too aggressive a tune or too poor of fuel quality.
Read first-Datalogging 101
Note- All of this basically assumes that you're familiar with your vehicle, are aware of the fuel quality that you have (Read: You-re-using-an-inferior-fuel), and understand that some parts require a tune. If you find yourself having to run a lower tune than your setup or fuel should require then you need to get in contact with your tuner for an updated map or find better fuel. If you absolutely cannot get your car dialed-in then err on the side of safety with positive corrections-a fractional amount of power is not worth the cost of an engine overhaul or replacement.
IF AT ALL POSSIBLE, HAVE A BASELINE WITH YOUR VEHICLE! I cannot stress this enough. If you don't have one and you have issues down the modding road, you very well may be having to get one later. Buy your datalogging device and all of the parts you want to throw on it, but before installing any parts install your datalogging device and go perfect your procedure and get some clean information of which you can later compare against. You will not always refer back to it, but it's the best-possible scenario on which to build against as your car evolves (i.e.-your first baseline is for comparison against your first modifications, your first modifications become your baseline for your second, et cetera).
Preface-Before going through these steps, it is imperative that you are able to collect clean datalogs-to include your baseline. If you cannot collect clean data then you could be chasing your tail trying to fix a problem with your car that is showing up in your data when it's actually your datalogging road or procedures. I've done this and it sucks, especially when your datalogging road is 45 minutes away. I shoot for having three comparable datalogs, which may not be possible depending on the issue you're having or trying to address, and if they're not comparable then I try and evaluate why and fix that before going to collect the data again. I will be assuming that the reason you're here is due to random or not previously known negative corrections within your logs and have written it accordingly.
Q: You just told me that I may not be able to collect clean data, but that I can't evaluate unless I have clean data so what gives?
A: Through a systematic approach you will be able to know exactly why (or at least narrow it down to a relatively few possibilities of why) you're having inconsistency in order to address it.
I've broken this down into two types of troubleshooting, Specific Part and General. Specific part would either be after having installed only one part, or the procedure to rule out individual parts during General troubleshooting.
Specific Part Troubleshooting
1. Did you have clean logs prior to installing a part? If yes, proceed to step 2. If no, fix your datalogging procedure.
2. Are your logs consistent? If yes, proceed to step 4. If no, verify the installation of the part. Loosen, but do not remove, all mount hardware, and then re-torque the item and any related hardware and proceed to step 3.
3. Datalog to see if the fault still exists. If yes, proceed to step 4. If no, you're done.
4. Replace aftermarket part with OEM part.
5. Datalog to see if the fault still exists. If yes, re-install your aftermarket part because something external is interfering with your data. If no, you're done and have a faulty or inconsistent aftermarket part.
That basically covers those that have taken consistent logs from the beginning. If you have something external interfering with your data, think really hard about what all work you did on that exact part and what components are related. Maybe you installed a DP and inadvertently shifted your CBE to where it's contacting a frame. Maybe you installed a FMIC and didn't notice that one of your clamps was touching something else (though that would technically have been found in step 2). Maybe you needed gas before doing your 'after' datalogs and got a bad batch of fuel. While the possibilities are relatively endless, you will find that, more often than not, it is something directly related to the work you did. If you installed a part that requires a tune (FMIC, DP, etc), then make sure you're running the applicable OTS tune.
Generalized Troubleshooting
Now, the fun and exciting stuff. Despite my advice, there are some people that are like me and will stockpile a bunch of parts for a larger install. While it makes a lot of things simpler mechanically (like doing your downpipe while your charge-pipes are already removed and out of the way), it introduces a ton of error possibility. On top of your mechanical installation, we're also going to be introducing tune-related possibilities. Because there are so many different possibilities of installs, review the list and see where you think you fall in to start.
There are no current established trends in the community at the time of this post, but maybe you're reading this a year from that date and there are. If there are and you have a part that's known to raise eyebrows, definitely start there. The key to this is going to be systematically reducing the possibilities, so you may be done in two trips to your datalogging road or you may be going there every day for a week. For the majority of this we are going to assume that someone went from clean stage 0 logs to a full stage 3 install with tune, though I've added a couple issues that might not be applicable to everyone but will be applicable to some.
1. Mechanically inspect all of your work. This will eliminate a lot of possibilities because you're probably going to find a clamp touching something, a loose bolt, an unplugged sensor, etc. Start from where air first enters the system at the airbox and work your way to the exhaust tip inspecting every mechanical connection, sensor connection, mount bolt (just use your fingers, you don't need to actually put a wrench on anything just yet), etc. Take your time because if you blow through this step you may spend many more hours later to come to the solution you should have arrived at here. Collect any OBDII faults and see if there is anything that could affect further work. Anything you find or are suspect of fix, then move to step 2.
2. Datalog. Is the fault still there? If yes, move to step 3. If no, you're done.
3. KAM/ECU reset, verify -1 OAR prior to and proceed to step 4.
4. Datalog. Is the fault still there? If yes, move to step 5. If no, you're done.
5. If using a custom-tune or not, load/reload the applicable OTS tune and move to step 6.
6. Datalog. Is the fault still there? If yes, move to step 6a. If no, contact your tuner for an updated map or if no custom tuner you're done.
6a. Load a lower-than-optimal tune and re-datalog. If the fault is still there then contact your tuner for an updated map or find better fuel. If not then move on to step 7.
7. Part/area trends. If there's a part that's known as being suspect or of having issues, start with that part/area and complete the Specific Part troubleshooting on that one component and if you reach step 5 of that then proceed to step 8.
8. If you're now here, you've found that the trending part isn't the issue and have it reinstalled. One part at a time, complete the Specific Part troubleshooting on each aftermarket part. There is no specific order to do this in, but I would go from the easiest (RMM, imho) to the most time-consuming (FMIC, imho) or difficult (DP, imho). What we want to do is systematically rule-out each component/system. Theoretically, you should eventually see a resolution. They key is to always re-install your aftermarket part once it's verified to not be the cause because we ultimately want to have all or as many as possible of the aftermarket parts installed. If you've worked through every possible part and still have a fault, move on to step 9.
9. If you're now here, it's pretty much time to go back to stock and get an appointment to have your car checked out because something is mechanically wrong. Post up on the forum, ask questions, etc., before you do because there's always the possibility that someone else ran into 'that one thing' and can offer advice or a specific solution.
So, there's that. A hell of a lot of words, but a very simple (though possibly time-consuming) overall procedure once you think about it as a whole. I should also point out the possibility of OEM-part failures. There is always a possibility that you have a OEM failure (or most likely simply a less-than accurate reading/measurement) that may not ever be thought about or remembered. Always keep it in the back of your mind that, while minimal, it is always a possibility. Maybe you damaged a wire or kinked a line or something during a maintenance procedure. If you've gotten to the point where you've done tons of work and absolutely aren't seeing an improvement the possibility is real that there's something related to the system that's failing or been damaged.
DANGER-Doing multiple pulls with extreme negative corrections can cause catastrophic failure of your engine, so proceed with caution. If you are experiencing excessive negative corrections repeatedly, consider stepping down your tune to something less-aggressive before proceeding. Just because the ECU can pull timing doesn't mean it's optimal for the engine, and pulled timing can be an indication of too aggressive a tune or too poor of fuel quality.
Read first-Datalogging 101
Note- All of this basically assumes that you're familiar with your vehicle, are aware of the fuel quality that you have (Read: You-re-using-an-inferior-fuel), and understand that some parts require a tune. If you find yourself having to run a lower tune than your setup or fuel should require then you need to get in contact with your tuner for an updated map or find better fuel. If you absolutely cannot get your car dialed-in then err on the side of safety with positive corrections-a fractional amount of power is not worth the cost of an engine overhaul or replacement.
IF AT ALL POSSIBLE, HAVE A BASELINE WITH YOUR VEHICLE! I cannot stress this enough. If you don't have one and you have issues down the modding road, you very well may be having to get one later. Buy your datalogging device and all of the parts you want to throw on it, but before installing any parts install your datalogging device and go perfect your procedure and get some clean information of which you can later compare against. You will not always refer back to it, but it's the best-possible scenario on which to build against as your car evolves (i.e.-your first baseline is for comparison against your first modifications, your first modifications become your baseline for your second, et cetera).
Preface-Before going through these steps, it is imperative that you are able to collect clean datalogs-to include your baseline. If you cannot collect clean data then you could be chasing your tail trying to fix a problem with your car that is showing up in your data when it's actually your datalogging road or procedures. I've done this and it sucks, especially when your datalogging road is 45 minutes away. I shoot for having three comparable datalogs, which may not be possible depending on the issue you're having or trying to address, and if they're not comparable then I try and evaluate why and fix that before going to collect the data again. I will be assuming that the reason you're here is due to random or not previously known negative corrections within your logs and have written it accordingly.
Q: You just told me that I may not be able to collect clean data, but that I can't evaluate unless I have clean data so what gives?
A: Through a systematic approach you will be able to know exactly why (or at least narrow it down to a relatively few possibilities of why) you're having inconsistency in order to address it.
I've broken this down into two types of troubleshooting, Specific Part and General. Specific part would either be after having installed only one part, or the procedure to rule out individual parts during General troubleshooting.
Specific Part Troubleshooting
1. Did you have clean logs prior to installing a part? If yes, proceed to step 2. If no, fix your datalogging procedure.
2. Are your logs consistent? If yes, proceed to step 4. If no, verify the installation of the part. Loosen, but do not remove, all mount hardware, and then re-torque the item and any related hardware and proceed to step 3.
3. Datalog to see if the fault still exists. If yes, proceed to step 4. If no, you're done.
4. Replace aftermarket part with OEM part.
5. Datalog to see if the fault still exists. If yes, re-install your aftermarket part because something external is interfering with your data. If no, you're done and have a faulty or inconsistent aftermarket part.
That basically covers those that have taken consistent logs from the beginning. If you have something external interfering with your data, think really hard about what all work you did on that exact part and what components are related. Maybe you installed a DP and inadvertently shifted your CBE to where it's contacting a frame. Maybe you installed a FMIC and didn't notice that one of your clamps was touching something else (though that would technically have been found in step 2). Maybe you needed gas before doing your 'after' datalogs and got a bad batch of fuel. While the possibilities are relatively endless, you will find that, more often than not, it is something directly related to the work you did. If you installed a part that requires a tune (FMIC, DP, etc), then make sure you're running the applicable OTS tune.
Generalized Troubleshooting
Now, the fun and exciting stuff. Despite my advice, there are some people that are like me and will stockpile a bunch of parts for a larger install. While it makes a lot of things simpler mechanically (like doing your downpipe while your charge-pipes are already removed and out of the way), it introduces a ton of error possibility. On top of your mechanical installation, we're also going to be introducing tune-related possibilities. Because there are so many different possibilities of installs, review the list and see where you think you fall in to start.
There are no current established trends in the community at the time of this post, but maybe you're reading this a year from that date and there are. If there are and you have a part that's known to raise eyebrows, definitely start there. The key to this is going to be systematically reducing the possibilities, so you may be done in two trips to your datalogging road or you may be going there every day for a week. For the majority of this we are going to assume that someone went from clean stage 0 logs to a full stage 3 install with tune, though I've added a couple issues that might not be applicable to everyone but will be applicable to some.
1. Mechanically inspect all of your work. This will eliminate a lot of possibilities because you're probably going to find a clamp touching something, a loose bolt, an unplugged sensor, etc. Start from where air first enters the system at the airbox and work your way to the exhaust tip inspecting every mechanical connection, sensor connection, mount bolt (just use your fingers, you don't need to actually put a wrench on anything just yet), etc. Take your time because if you blow through this step you may spend many more hours later to come to the solution you should have arrived at here. Collect any OBDII faults and see if there is anything that could affect further work. Anything you find or are suspect of fix, then move to step 2.
2. Datalog. Is the fault still there? If yes, move to step 3. If no, you're done.
3. KAM/ECU reset, verify -1 OAR prior to and proceed to step 4.
4. Datalog. Is the fault still there? If yes, move to step 5. If no, you're done.
5. If using a custom-tune or not, load/reload the applicable OTS tune and move to step 6.
6. Datalog. Is the fault still there? If yes, move to step 6a. If no, contact your tuner for an updated map or if no custom tuner you're done.
6a. Load a lower-than-optimal tune and re-datalog. If the fault is still there then contact your tuner for an updated map or find better fuel. If not then move on to step 7.
7. Part/area trends. If there's a part that's known as being suspect or of having issues, start with that part/area and complete the Specific Part troubleshooting on that one component and if you reach step 5 of that then proceed to step 8.
8. If you're now here, you've found that the trending part isn't the issue and have it reinstalled. One part at a time, complete the Specific Part troubleshooting on each aftermarket part. There is no specific order to do this in, but I would go from the easiest (RMM, imho) to the most time-consuming (FMIC, imho) or difficult (DP, imho). What we want to do is systematically rule-out each component/system. Theoretically, you should eventually see a resolution. They key is to always re-install your aftermarket part once it's verified to not be the cause because we ultimately want to have all or as many as possible of the aftermarket parts installed. If you've worked through every possible part and still have a fault, move on to step 9.
9. If you're now here, it's pretty much time to go back to stock and get an appointment to have your car checked out because something is mechanically wrong. Post up on the forum, ask questions, etc., before you do because there's always the possibility that someone else ran into 'that one thing' and can offer advice or a specific solution.
So, there's that. A hell of a lot of words, but a very simple (though possibly time-consuming) overall procedure once you think about it as a whole. I should also point out the possibility of OEM-part failures. There is always a possibility that you have a OEM failure (or most likely simply a less-than accurate reading/measurement) that may not ever be thought about or remembered. Always keep it in the back of your mind that, while minimal, it is always a possibility. Maybe you damaged a wire or kinked a line or something during a maintenance procedure. If you've gotten to the point where you've done tons of work and absolutely aren't seeing an improvement the possibility is real that there's something related to the system that's failing or been damaged.