Its a long term value that is updated over time. IE: you have to do quite a bit of part throttle knocking for this trim to even be updated. After a few seconds of knocking the value will be updated. Speaking in generalizations it takes about quarter tank of gas for this value to start to come close to its final setting.

By the same token, it can take a few hundred miles to get full octane value back after you've filled up with good gas.

 

I see. thanks for the clarification. so it acts in a similar way as long term fuel trim, small increments over time

 

So is this similar to the IAM value on Subaru ECU's?

 

what does IAM stand for

 

I actually logged this exact condition today.. It went from 255-215, but only took about 10 minutes to return to 255..

I think the further you drift from 255, the longer it takes.

 

MalibuJack, can you post the log or sections/interpretation of it - ie knock sums to make it decrement the value, how much it decremented, what effect this had on further timing, was it appearing to keep the knock sum below a certain value by regulating octane value etc?

 

I'll have to post the log of it later.. Plus I can give you a representation of it later.. It definitely worked as described though, there was a bit of knock repeatedly under light load part throttle driving, which caused it to shift from 255-215, then there was a period of light load cruise driving where it remained at 215, then I had accellerated under light load with no knock (maybe 1 count or so) and as I did the octane number returned to 255...

 

Thanks, look forward to the details

 

Well you guys are (not being mean) freaking out about 10 counts of knock like its going to blow your motor. When I bought my car it was knocking 30 counts at WOT, lots of part throttle knock, and lots of phantom (not real) knock. The octane value was near zero. After fixing a problem it would take near a full tank of fuel to get my timing back. A quick trick was just to cycle battery power. Maybe evo's update faster, but the theory is the same which is important to understand.

Also I see a LOT of people getting part throttle knock and dismissing it as not real. While it isn't very dangerous I'm betting a lot of this knock is real and being blamed on noisey downpipes and such. You'd be amazed at what just a degree or 2 of timing can do. I used to have a lot of part throttle knock which was fixed by tuning the timing down low. Then I had problems when transistion into boost so I had to fix those timing cells. Then I had problems with knock at 6000rpm's and had to fix those timing cells. I too also have a noisey downpipe.

 

Actually I dont freak out over any knock count under 20 or so.. However I have been able to determine the difference between real knock and phantom knock because I listen to my engine with det cans..

The further from the 255 you get, the longer it takes to return.. I'm sure most are already aware that resetting the ECU resets the octane number..

I can definitely tell when my downpipe is causing knock though, since I put an insulator on there and for the duration it was kept from making contact, that particular knock went away..

I think your missing the point of keeping a low knock count, technology for tuning and ECU's has improved.. We are just taking advantage of these advances. Heck, back in the carb days nobody cared about knock at all unless you audibly heard it which was rare due to loud exhausts..

You also took my point out of context, which was that I was able to see how the ECU reacted to some sustained level of knock by altering the octane number.. But also notice that the sustained level of knock lowered the octane level, the more it does, the lower the octane number and more deeply your in the low octane map.. This is undesirable and is avoided by attempting to keep the lowest knock count possible.

 

The ecu's powering the origional USDM based 4g63's and the current Evo's aren't that much different. The exact same code can actually be found in some parts of the bins.

I was just trying to state that real part throttle knock can and does occur. I see a lot of people dismissing this as false knock and getting an itch to disable the knock sensor or alter the filters.

The only point I'd like to stress is that its very easy to do a few little things to make the engine bay "quite" to let the knock sensing system work correctly. Relying on it is very important.

 

Oh sure, I definitely see knock at part throttle, tip in, decel, etc.. Every car I've worked on has it to some extent.. But one thing I won't do is alter filters unless its absolutely necessary, some engines have unusual harmonics and will produce phantom knock in different places than the original equipped engine, or more noise than the original engine, etc.. those are the things you might want to adjust your filters, ONLY if that is your only option. There are just so many ways to do this that its semantics at this point.

I have directly compared the two roms now that I can.. There is some similarities, and some bits of shared code, but the knock detection portions of the code are definitely different..

 

Ignition Advance Multiplier

 

FWIW, I started tuning aftermarket ECU's ~9 years ago when I purchased my Electromotive Tech II. I think my experiences have been very similar to yours. At part throttle acceleration and the transition from off boost to boost is where I saw most of the knock, which always made me concentrate on this area the most. I really had to yank the timing hard coming on to boost or the engine would pick up knock(audible knock). I have posted a map below from my old setup on pump gas. BTW, I typically ramp in 3-4deg of timing linearly from peak torque to readline, but this map does not show this.

Also, you can see that I was very aggressive with timing up top and that is LARGELY dictated by intercooler efficiency. On the same setup, I had to run 3-4 deg less timing with a smaller intercooler. I guess my point is this, put as large and efficient an intercooler as you can in place of the stock unit. It is decent as far as OEM intercoolers go, but you can do much better. It will definitely allow you to run more timing and more boost on pump gas.



BTW, I have posted a pic of my buddies setup (MP FAB) using the 24"x12"x3.5" Garrett GT core, which is the same as I will be using in the near future.