RomRaider

Yes I update it before I posted about it.

Still interested in logs for any cell showing FLKC in LTV.

Will get more logs. Do I have to take LTV screenshot right after the logs? In the log I took today for SD card testing, I had some very minor FLKC (maybe 15 events for 13,500 records) in them, but since I didn't have laptop attached I didn't have access to LTV. So if I took LTV screenshot tomorrow morning for example - didn't drive in between - would LTV show stored values for FLKC or would they have been reset?

They may still be there, depends on if they were sensed and kept or cleared normally.

This type of diagnostics you'd need to log with a laptop so you can see what's going on and react to it.

I went for a drive today to do some logging. i did some pulls, but mostly a regular spirited drive. I have managed to create three logs, each time driving for a while then stopping and creating a screenshot of LTV view (thus interrupting the log). Then I would start another log, drove for a while, stopped the log, and took another LTV screenshot. Luckily for me - but contributing nothing to our quest to find FLKC numbers in LTV - I didn't have ANY knock events at all, and sure enough there was nothing in LTV either. I am posting one screenshot just to show that last row is still at -125.00 but all others were 0, and that was consistent. The AF learning is different so I know it logged something

I also created a new logcfg.txt for my SD card, so after I was done with logging to laptop as above, I switched to SD card logging (still on same drive). I created logcfg file so it starts logging when RPM>0, stops when RPM=0; and create a new log any time throttle opening goes over 95%, and again when it drops below that. I wanted to see if it would work. And it worked - mostly. It seemed to create a new file when I did a pull in 3rd gear, but it logged some seemingly random throttle opening percentages just after the pull - for maybe a dozen events - and then it recorded more 100% throttle opening events in the new file even though I don't think I lifted the foot of as it shows in the logs. Other than that, it worked!

Ironically, in the last log ("log0003.csv" file), after the pull and on the cool down it seemed to have created some knock. Of course that was to SD card, and since I only reviewed the file afterward (to see if logging worked, not for knock) once I went back to car and got LTV going again there was nothing in it (all FLKC values were 0). But I find the FLKC values in the log itself a bit weird as all the time logging to laptop just prior to that similar conditions created no knock. So I am not even sure the knock was "real".

I know we are trying to figure out the LTV table only, but I will attach two logs from SD card anyway - any idea what it looks like to you might help. One is the pull file (3-rd gear pull), and one is the file created after the pull with all the knock events of the entire hour's worth of drive - or more. If helpful I can post other logs created just prior on the laptop for comparable conditions, but didn't want to overwhelm the thread.

Will continue to log - I am interested in resolving the LTV's -125.00 values; and learning in the process.

Finally had time to log with laptop and provide more hopefully useful information. The attached files were made by me driving and my helper (thank you!) looking at the RR logger view while logging to file. When the FLKC would register something other than zero, he would stop logging to file, and switched over to LTV view and take a screenshot.So attached are three logs and three LTV screenshots (one after reach log file was stopped because FLKC value in logger window has changed). I didn't change the names as they contain time stamps so you can follow along - if that is not enough I can make it clearer.
I don't know if lack of any FLKC values in the Learning View are indicative of some specific Subaru knock strategy or that I am missing the fields all together in my Def file. I also noticed that my knock comes in -1.5 degrees FLKC, and it either goes to 0, or it trends towards zero in 0.25 degree steps. I don't have anywhere anything like 1.4 value that others refer to. So Subaru might have well changed how it handles knock prevention. Or I am looking in wrong spots... Since I am using "non-sanctioned" def file, I am still a bit apprehensive about the values I read. For example, the CL Fueling Target Compensation A and B load tables have an odd load and RPM ranges in the tables (see attached screenshot). So I am not sure if my values are correct or not in the Def file.

General observations and questions if you don't mind me asking here, given my ROM is completely stock, and there is only two even remotely relevant mods on car so far. One, I deleted the sound tube that took some air intake into a cabin to hear engine better; and two, I have delete axle back mufflers tubes:

1) A large majority of my knock comes between 3,700 and 4,000 RPM, with load between 0.8 and 1.2 g/rev. I have noticed this in the last week of logging my daily drives to an SD card, as well. There are very few of those events, all told, but when they happen they happen by and large in that load/RPM range.

2) Coincidentally, that is also an area where Base Timing Primary Non-Cruise table seems to retard timing relative to neighbouring cells. I have attached that file too for your reference. I find that weird but I am new to this and someone might have done that for a good reason. I am just curious and would like to hear what do you think of it. Is this 2017 (or 2015+ STi) specific?

What can I do next to help figure this knock strategy/def file out?

Finally had some time again to look at the def and logger files, and I have some newbie questions. Apologies if this is considered basic knowledge - it is not to me, so I figure I'd ask. In my definition file I have a few tables defined not exactly twice, but once with Statis Y axis, and then with some other parameter as y axis. Example is like this:



In the tables when viewed in RR, I only see one value, and there is no reference to RPMs in my table. What is the purpose of defining table in these two different ways? There are other tables like that, but I didn't want to dump a whole bunch of code here.

Table "Rough Correction Learning Delay (Increasing)" also has two entries, and only difference that I can detect is in one storage type is defined as unit8, but in the identical one below it, same storage type is defined as unit16.

Reason I am asking this is a) I want to understand the logic behind it; and b) I want to clean it up if it is not needed for my specific car. I am not a pro tuner, and I am only ever going to play with my car, so want to know what is applicable to my car, and what is in there for general what-if scenarios.

My knock has not changed - still get minor -1.5 events here and there, but no changes to IAM, and nothing in Learning View tables. This is another reason why I am cleaning out the def file - I want to be able to understand if the numbers I am reading are actually correct. Nowhere in my knock-related tables do I have a value of -1.5 as corrective value. It refers to -1.4, and increments of 0.35 everywhere. So why is mine starting at -1.5 and reduces in 0.5 or 0.25 steps? I am really curios.

You are looking at the BASE definition, from which all ECU specific definitions are inherited.
To know which tables are actually used by your specific ROM you need to refer to the section of the definition file that's directly related to AE5W101V.
Those "duplicates" you refer to, one is used by your ROM and the other by some other ROM, but again, it is the BASE from which all ROMs read the tables from.
You can ignore the ones that aren't used by your ROM, the Editor will do this automatically and not even load them. So you are kinda wasting your time investigating them.
To see a list of tables loaded for your ROM check the Editor View menu and the Properties for your ROM.

I'm not sure that one FKLC event will register in LTV. I believe it has to be consistent and reoccurring to be considered as a learned value.
There is a thing know as preemptive knock correction. It's not defined for all ROMs. It reduces timing in advance of detecting a knock event to try an avoid it. Maybe this is what is showing in your log. If you don't have confidence in the LTV results, then the logs are your definitive proof of knock. Just be sure to log load and RPM to know which fuel or timing cell to edit.

Aaah - OK, that makes sense. Your comment about base and ECU-specific code forced me to look closer into definition file, and now I see there are two "sections" - AE5W101V, and "32BITBASE" so I suppose that is what you are referring to. It seems like AE5W101V section only lists the names of the tables, and addresses in the ROM where to find the values, and 32BITBASE acts as the "formatting" tool for those tables where names match... I know I am probably oversimplifying it - please correct me if that is wildly off-base.

I have read through the preemptive knock thread you referenced, and it is quite interesting. It got me thinking, so I tried and added an address for this "preemptive" value I found in that thread for AE5W100V ROM (which I modelled mine on anyway) - DCE80. I added this code in AE5W101V section of my def file:



And "formatted" the table by copying one of the existing ones and renaming it in "32BITBASE" section:



I then put my ROM image I have saved on the computer into RR using this newly modified def file, and I got the value from that address in the newly added table. But the value is again -1.4 degrees.

So no matter what I do, I can't find -1.5 degree value defined anywhere ... But I think this "preemptive" concept is definitely what I am seeing in my logs. I am just curious as to why I can't find the "source" of it anywhere in defined tables.

I am not yet confident enough in my RR understanding to be making actual changes to the ROM, and flashing them. So this is part of my learning process. It is not that I am not confident about LTV, but when I see inconsistencies I want to learn as much as possible about their causes

Do you have a correction value of 0.5 set somewhere? Three of these = 1.5. The logger may not be fast enough to catch three events in succession.

As far as I can tell, 0.5 value is only related to IAM. See image.

I was going through the logger and my def file to try to understand what parameters, and values those parameters, am I supposed to see in my log, and in my tables. For example, I would expect to see "Boost error" parameter in my logger window (to chose, and pick it for logging if I wanted to) but it is not in the list of the parameters I can choose from when I fire up the logger. It is supposed to be E35, according to the logger file:

I would then expect a line item in logger window that says "Boost Error", but it is not there.
So, similar with those knock parameters - I am really only trying to set up a "rulebook" for my car, and my def file - what am I supposed to see, how can I verify the values, and then perhaps think about changing them knowing what to expect. So apologies for being a bit slow... this is lots to process
By the way, thank you for being patient and explaining what are probably simple beginners questions.

RR Logger should show the parameter if the ECU ID matches.
Are you sure the Logger is set to use the same def file as you are looking into?

The latest version of Logger def (v346) shows this for E35 Boost Error*:

There should be ~50 extended parameters where the name ends in * or 1-byte or 2-byte or 4-byte.

never mind - user error ... when I first started using logging, we were back at logger 235 version or something. I created my "profile" from then available, standard, parameters and I didn't realize that unless I am hooked up to the car and looking, I will not see extended parameters in RR logger window. Sorry for that - I should have spent more time investigating before asking!

Long story short - what I thought was proper parameters to log turned out to be "P*" parameters - rather than extended "E*" parameters. Sinc my first logging was before the logger was updated for my ROM, and I created a log profile those parameters in my profile never got updated with new, extended ones when I later started using logger file that dschultz updated for my ROM at later versions.

Now that I actually logged the car with extended parameters, all my -1.5 deg FLKC values, and 0.25 degree decay of those initial -1.5 deg have started showing up as -1.4 deg, and have a decay of 0.35 deg. As I would expect them to be. So lesson learned. I will spend some time logging and see what I can see in LTV tables again, now that I use proper values in the logger. Will post back with updates in a little while.

And I am sure I will have more silly newbie questions...

Hello again -

I have been logging a bit with SD card - simple, while going to and from work, etc. while accepting some limitations - over the last little while. I have also tried to read more on logging and romraider use, but as always I always seem to find an answer once I already ask a silly question .

I have some knocking logged every now and then (see my latest log attached) but it seems to be either during the gear change or only at lower RPMs so I tend to believe it is false knock. I run 94 octane gas, and in general car runs great. As you can see in the log, I get some FLKC here and there, and some KnockSum, but IAM stays at 1. Having seen the log file after drive today, I went back to the car and tried to pull LVT from it (yes, I realize the data might have erased from ECU in the meantime) just in case. Despite the logged values, my LVT is all 0's for knock and my highest RPM range is still suspiciously populated with -125.00 values. No matter what I do the knock-related values don't change at all. Other learning parameters (AF correction for example) does change, and it does stay available to LTV even after car restarts, so I don't think it is a matter or car being restarted.

To avoid guessing, my FLKC logged parameter is E95 parameter, the FKC is E94 and IAM is E96 in the v345 logger file. Is there anythign else I can do to get us closer to establishing LTV values working for my ROM?

Hello

life got in way of posting before, but I figured I'd update the thread. After a while, I pulled up the LTV view again a couple of days ago and lo and behold, my -125.00 values changed! Well, some did... See attached screenshots; each one is after a very short drive (5 km or so) back to back. I still have two -125.00 values but those are in load ranges I don't think I have been into for much at all - given winter driving and daily grind rather than racing. I still find it odd that Subaru set -125.00 as starting value, but that is what seems like a logical explanation why they are there.

I have also found that my A/F learning table has more columns that is needed (notice 1st, 2nd and 3rd columns for Air/Fuel Ranges) in same attached image. I am guessing that can be changed by editing the definition file and updating the sizey of the corresponding table but it is already asking for only sizey=3, so I am not sure why there are 4 columns in the LTV. Am I the only one with 2017 sti, who has that type of table content? If so, what am I doing wrong?



Car still runs great, even though I still get some FLKC of up to -1.4 in the load/RPM range as caught on the second screenshot - so I am inclined to think this is more my commute-style driving than anything really wrong with the car. I will continue to monitor it and see how it goes when we get out of the cold winter. This is the same area in the map where base timing is already set for -41 degree or so.