jcsbanks

Yes.

(((255-octane)*low)+(octane*high))/255

MalibuJack

iTrader: (5)

Yes.. the if your octane number was 128 (50% of 255) your values will be interpolated 50% between the two.. As far as I have seen, the algorithm is kept very simple..

galvitron

iTrader: (2)

Great. Thanks guys.

MalibuJack

iTrader: (5)

Ya know, this does actually open the possibility of a map-switching mod in the future..

jcsbanks

Map switching could actually be a relative doddle if we can crack the comms. I've not traced whether any spare ADC channel pins are brought out to the ECU plug. If they are then a potentiometer to scale between low and high maps would be easy enough. We just disable the inc/dec of the octane number and make it read the ADC instead.

I also thought it would be nice to modulate the boost level with the octane number.

AlwaysinBoost

iTrader: (17)

wow this is awesome... I've said it before and I'll keep on saying it; its an awesome time to be an EVO owner!

dan l

iTrader: (3)

****ing awsome news. Allthough logging evos I noticed how similliar they are to DSM's and had a slight head start on some of the people. I'd just like to take this chance to reiterate (and I do appreciate ALL the hard work and discoveries) the importance of the grams/rev value. Unfortunatley I don't have the time I need to help you guys out. I can only spend 5-10 mins on the board a day and just try to give tips where I think I can help.

P.S. Finally got my evo IX saturday!

P.P.S

Even though your octane value is increasing at low knock levels, you may notice that you can still loose a degree or 2 of timing due to the real time knock pull (.3 degrees per count or whatever it may be). Just so that this is understood as I see some confusion as to the octane value and knock sums with respect to real world timing.

The ecu may also pull timing when the baro is reading high evevation and/or the coolant temps are elevated.

MalibuJack

iTrader: (5)

It'll all be documented soon enough..

andenbre

iTrader: (4)

makes me wish I studied computer language, now I will just step back into the shadows and let you guys work your magic. Thanks again.

codgi

iTrader: (22)

x86 or bust guys! . Have'nt touched the stuff in a good bit though I suppose my assembly is very rusty right now. I'll help out from time to time where I can as well just really busy right now with work/move etc.

When your ready for code reviews/testing let me know...its what I do for a living

ST

iTrader: (8)

john - i'm curious if there is any specific conditions the octane rating increments? the reason i ask is because after getting my car to run again today (long story), i was able to log an event where some ks pushed down the octane, but subsequent low ks counts (many 0's actually) didn't change the octane rating at all. Could it be possible that high ks also incurs a "penatly" period before it starts incrementing again?

jcsbanks

It has a minimum coolant temperature and load areas that it will only make changes in.

I have the 10 bit load index RAM location that is used to look up the boost level for wastegate duty/boost limit. I can chop the bottom two bits of this to make a load8 with 1.25 steps of load. I just need to chuck it out the serial, but I can't work out how the routines deal with the request IDs. There is 16 times the memory area of the 1g DSM to look through

MalibuJack

iTrader: (5)

That would be a nice little mod to make eh?

C6C6CH3vo

iTrader: (9)

First off, thanks Mr. Banks and Jack for applying your talent to help my Evo get faster, and most importantly - doing so generously without greed.

Below is a peice of log showing a large amount of knock which was unusual because I usually get no more than 1 or 2 counts with three or four log ID's in a single pull. The only difference is it was recorded coming off a 4th to 3rd downshift and my clutch is toast.


By the looks of the knock numbers, and how linear they appear, would you guess tranny noise was picked up or that was actually knock? Once again I normally see no more than two occurances of knock with 1 -2 counts, for each logged pull.

Also, shouldn't the OCT number be less (251) with that amount?


Knock - Timing - OCT
5 - 14 - 255

4 - 14 - 255

12 - 10 - 255

12 - 8 - 255

11 - 6 - 254

11 - 7 - 254

10 - 7 - 254

10 - 7 - 254

9 - 9 - 254

8 - 10 - 253

7 - 10 - 253

7 - 11 - 253

7 - 12 - 253

7 - 12 - 253

7 - 13 - 252

7 - 13 - 252

7 - 14 - 252

7 - 14 - 252

6 - 15 - 252

6 - 16 - 252

6 - 16 - 251

6 - 16 - 251

6 - 17 - 251

6 - 17 - 251

6 - 17 - 251

jcsbanks

The linearity is normal - even a short knock spike will decay over a much longer period even after the knock has abated. This is sensible really.

The octane number moves slowly - it is supposed to be a long term trim. It is behaving as the code predicts. It doesn't reduce more quickly if the knock sum is higher.

So, there are timers in the code both for knock sum decreasing and octane number increasing and decreasing.