My issue is that when I get the phantom knock its definitely over 6 counts, sometimes 20*+ which I know is very bogus knock. So if the knock map isn't going to combat such a thing from pulling my timing, is there any way around it? I wish you could 'intersept' the knock value somehow and hold it 'hostage' (lol) so the ECU doesn't do anything unless certain criteria are met (certain RPM,load,TPS ranges or whatever). Then if you could datalog the address location where that value is being held "hostage" so you can still log the knock sum value even though the ECU technically isn't seeing it. This way you could log the 'actual' knock vs. the knock that the ECU is actually seeing. This is just babble so forgive my airyness.

 

I am also running into this phantom knock issues in two major areas: 70-100 load 3-4k rpm and 5-5.5k in the 200-220 load when trying to tune on the dyno. Both aren't knocks since we've verify with a spark plug scope that there was no detonation on the plugs while timing was anyway from 1-4 degrees so it was running on extremely low timing. Octane was always on 100.

I know this would taken care of the 3k-4k rpm range but has anyone experienced what I am in the upper range? How safe would it be to implement this in that range?

BTW this is a 35r on a built bottom end.

 

all i get is 4 sometimes 5 knock if i start the pull from 2500-3000 between 80 and 100 load,the thing is if i start the pull it just kept on knocking all the way...if i do a pull from 3500-4000 rpm i dont get no knock at all all the way to redline

 

any update on the patch?

 

Tephra said he is gonna port the patch to 94170015 sometime in the next week


"I personally think the octanenumber thing is bs anyways.

I have never seen it drop below 98% which means that its stil using high-octane by 98%...

anyways not so much of an issue

I will try and get a 94170015 out next week so more ppl can test "


I think i need to donate again

 

Cool!

I don't know about that though, I had my octane drop to freaking 20%.

I could hardly drive the car...lol

 

Question about the patch. If you build a KS of like 8 during the 2500-3500 rpm range and the patch mul's it by the .25 or .5 from the table, but then you reach 4500 rpm and the 3d table is 1.00 will it use the built up KS of 8 or only any extra KS that adds after the 8? Because that will make this patch not work very good. If i still have the 8 KS and its pulling 2-3* then its the same deal with out the patch.

 

All the patch does is take the current KS reading and multiple it by whatever % is in the map for the given RPM/LOAD.

There is no 'memory' just a snapshot of whats happening at that point in time

 

Tephra, is this a "proof of concept" patch or do you envision expanding on it? The reason I'm asking is because of a few applications where the engine is artificially noisy by 3-4 knock sums due to motor work. I'd like to be able to ignore 4 knock sums (with say a 500rpm interval from 1000-8000rpm's or so). Then above this 4 knock sums pull timing as normal. I'm afraid of putting a "blanket" percentage on amount of knock sensor suppression because when a part fails, and the car trys to spit out 35 knock sums, I want it pulling a crap ton of timing. In my example 35 -( 4 knock sums ignored)= 31 knock sums worth of timing pull.

Just a thought, I know your a busy guy.

 

I don't understand, why does everyone say there's no room for memory on the 96940011? I've added every upgrade I can find with no problems at all. What's this about?

 

You probably just added exisiting maps already located in the ECU.

Adding features such as Tephra's patches (beside the knock light, valet, ntls) is just not possible, since these are added to the ROM.

 

So any testers have any feedback on how it is working so far? has it fixed a lot of the phantom knock issues?

 

Now that I've thought about it more, what might be simplest is this: Have a 1D table that has a subtractor value. Create the code so that when the ECU produces "knock count" it subracts this "subtractor" value from it and then feeds it to the other routines that use the knock count. You can log the original knock count and feed the modified knock count to the other routines. This way I could set it to say 6 and then anything up to 6 counts the ECU will think is zero. So if I got 12 counts it would think it is 6 counts and adjust my fuel/timing accordingly. I would think this would be an easy first step while more complicated code is created.

lol, posted that without seeing page 3. Either way, this looks cool too. Look forward to it.

 

I kind of like this simple idea.

 

I think this would not work well as for me at least, when i hit 20+ Knock counts at 3500 i want the ecu to see it as zero, but when i get 3 knock counts at 6500 i want the ecu to see 3 counts.