RomRaider

I finally got a chance to mess with what I described here awhile back:
topic1126.html

Basically, I created a ROM hack that is active whenever the knock signal is set and keeps track of the load and rpm range in which the knock "event" occurs (increments a counter in that cell). The load and rpm ranges have good resolution (15 x 17), designed to follow the base timing axis value for the most part (the midpoint of each range corresponds to an axis value in the base timing table). Basically, it is similar in function to the FLKC table, except that it counts the knock events regardless of whether or not it results in a change in IAM, FLKC or FBKC (which means it is active at any RPM and any load).

To use it, it would work as follows:
1. Patch the ROM with hack and flash to ECU with Ecuflash.
2. Drive
3. Run the knock record app to check the table (you can also clear the table to start fresh).
4. go to #2

There are problems with releasing it to the public as it stands now as it is not very user friendly. This is due to some issues in the way the 16-bit ECU handles RAM that I need to address for real-time tuning anyway, so likely I'll work on it and then release it later as a test which will also advance the RT progress.

The screenshot shown is after a 30 minute drive (some city/some highway). Usually the IAM settles at 15 with this current tune I'm working on.

nice work !

Can't wait for the 32bit stuff

Sick. I remember thinking about that a long ass time ago, nice work making yet another cool idea happen....

Holy crap, you're really making things happen ... thanks again for all the work. Looking forward to this one!

This would be awesome.

Wow, shockingly similar to what I have seen on my 2.0LGT. Lots of knock at 1600-2800rpm / <1.15 load. I wonder if this is normal for most subaru engines.

I'll continue to post updated screenshots of the table as I drive it. My issue is that my IAM is changing from this low load knock within the minimum load/rpm ranges for rough correction.

Are you doing that just to make sure you produce recordable knock events?

No, I know it works. I'm using it to tune with to see how useful it would be as well as out of curiosity to see what knock events (probably false knock) occur at low rpm and high rpm outside of the normal knock control ranges. I have the initial IAM at 8 with this tune.

again, great work.
i see that the new site is still being worked on to get all of the faq and article info up... once those things are complete it'd be great to have a thread or link that gathers all of the various tools and programs that have been developed to list them all in one place, sort of a 'top hits list' for open source tuning/logging.

Very nice!

Here's an updated SS with some timing pulled in the problem regions (primarily low load, 2000-3200 RPM). IAM is 16.

It is recording quite a few knock events below loads of 0.3 g/rev, which I'm guessing is just noise.

Yea, definitely just lugging noise down there.

Are you planning to integrate this into Learning View eventually?

Since this needs a ROM hack, I will likely integrate this into the ram writer for real-time tuning. Maybe as a early test for managing ROM patches.

Looking good !