mchuang

iTrader: (11)

Look where it is saying low octane ignition, high octane ignition, and knock sensor filter. Do it at your own risk. This is on my car so I will risk as I please lol. But I think the factory ecu is playing a lil to safe so Imma test it out. Have access to a good tuner locally to help me out. Just be careful if you decide to play with it. The only reason I get the assumption that it is very conservative is because the timing maps and fuel maps are very conservative also some people here who are tuned by tuners on here are seeing like 10 - 15 knock sum, the most I have seen is 2 and it is not even consistent. I make a run then 30 min later make a run and the first run shows no knock sum, then second run shows 1 measly knock sum for 200 rpm after 4600 and pulls timing 1 degree. I am like wtf same gas, same motor, same map and last pull no timing pull. I have been reading around on it and found that the stock knock sum is too sensitive, I think more sensitive than dsm back in the day.

vboy425

iTrader: (19)

not always .96 is rich. i have tune plenty of evo with the SAFC. sometime .96 volotage is about 11.2 AFR and some time .92 is 11.1 AFR. ffor my evo. it show .94-.96 on my zeitronix wideband show 11.4-11.1 AFR. Remember the o2 voltage is just a narrowband, it's not very accurate.

jcsbanks

My narrowband O2 shows typically 0.87 to 0.89 when I'm in the low-mid 11s:1 on my wideband. I would never rely on narrowband for tuning.

jcsbanks

It appears that the ECU may be able to retard 1 to 3 degrees across the rev range simply to determine if what it is hearing is noise. This sounds like a good strategy to me for finding phantom knock. My ECU ignores lift off det, but otherwise seems to respond appropriately. I don't think it is too conservative.

I would think that the ECU uses adaptive methods for determining knock, probably has some DSP in there, the anti-noise adjustements as mentioned (8 of them). It probably uses gated acquisition in a time window per cylinder. And it has probably been developed in a multi-million dollar lab.

I found the J&S Safeguard limited because it picked up phantom knock at low RPM and ignored high RPM det. It needed an adjustable threshold by RPM. The gold standard has to be det cans, but a knocklink where you are looking at the level of background noise and look for spikes on top is also quite good IMHO.

jcsbanks

Girlie, I would interpret your logs as follows, at 5,6,7000 RPM you have 9,9,15 degrees at similar boost to me on similar octane, running similar(ish) AFR ALSO on a IX. Mine runs the rollowing timing at 20 PSI, 11.5:1 AFR, 98 RON:

4000 8
5000 8
6000 10 (knock count of 1 from 5600-6600)
7000 16

So you are within a degree of what I'm running, just that I have my map probably more retarded (and get the odd 1 knock count) and your ECU is doing it for you with higher knock counts.

Just my 2c. Let me know if you agree/disagree. I'm starting to think that the IX needs less timing than the VIII.

What I'm not sure on is if the knock counts should reduce when the ECU corrects the timing. I am suspecting not, that the knock count/sum is a sort of semi-processed marker that modifies timing. Would be interesting to see what the car would do with low octane fuel at stock boost compared with high octane fuel - does it always keep the knock count/sum high on the low octane fuel?

andenbre

iTrader: (4)

my stock map knocks 1-2 counts at 19 psi from 3800-5800 on 93 R+M/2 in 90 degree heat

Girlie

Here is a third gear pull I did on that same night. It was done about 10 minutes before the other one. (car was shutoff in between). Log looks pretty good across the board, little knock on top that pulled a degree of timing around 6500.



RPM Timing Adv Knock Sum
2500 17 1
2531.25 17 1
2593.75 16 1
2656.25 15 1
2687.5 15 0
2750 14 0
2781.25 13 0
2843.75 13 0
2906.25 10 0
2968.75 8 0
3031.25 8 0
3062.5 8 0
3156.25 8 0
3218.75 8 0
3312.5 8 0
3375 8 0
3468.75 8 0
3593.75 8 0
3656.25 8 0
3781.25 8 0
3875 8 1
4000 8 0
4093.75 8 0
4187.5 8 0
4312.5 9 0
4406.25 8 0
4562.5 8 0
4625 8 0
4750 9 0
4843.75 10 0
4968.75 10 0
5062.5 10 0
5156.25 10 0
5281.25 11 0
5375 11 0
5468.75 11 0
5593.75 11 0
5656.25 11 0
5781.25 12 0
5875 12 0
5968.75 12 0
6031.25 12 0
6125 12 0
6218.75 13 0
6312.5 13 0
6406.25 13 0
6500 13 0
6562.5 12 6
6656.25 12 6
6750 13 6
6812.5 14 6
6875 14 6
7000 15 6
7031.25 15 5
7125 16 5
7187.5 16 5
7281.25 16 5
7343.75 16 5

Here is the one I posted earlier. Again 3rd gear, started the pull later in rpm band due to traffic

RPM TimingAdv KnockSum
4344 9 2
4406 8 2
4438 8 2
4531 8 1
4625 7 2
4688 8 2
4781 8 2
4875 9 2
4938 9 3
5031 9 3
5094 9 3
5188 9 2
5250 8 9
5313 8 9
5406 8 9
5469 8 9
5563 8 9
5625 8 9
5688 9 9
5781 9 8
5844 9 8
5938 9 8
5969 9 8
6063 9 8
6125 9 8
6188 9 8
6250 10 7
6313 10 7
6406 10 7
6469 10 7
6500 11 7
6594 11 7
6656 12 7
6719 12 7
6781 13 6
6844 14 6
6906 14 6
6969 15 6
7000 15 6
7063 15 6
7125 16 6
7188 16 6

JCSbanks. I have seen a couple different IX timing maps, and all of those show more timing being applied vs a VIII map for similar mods.

A side note here. When I was tuned it was spiking 23psi, tapering to 19-20 psi. From my Autometer gauge, I am spiking 20psi tapering to 18psi now. Could the lower boost I am now running be some of my problem? Maybe putting it in a lower load cell that has more timing or a fuel cell with leaner AFR? I know the rule of thumb is too turn boost down when it knocks and go from there, however in my case could it be the opposite?

Girlie

nj1266

iTrader: (6)

Another way you can do this, which is equally dangerous, is to isolate the knock sensor on the block by using dynamat behind where it screws in. Some people do this on the 02-03 SE-R Spec V. That car had a very sensitive knock sensor and when the car is modded it picks up all kind of noise that is not really knock. Some people even use a dummy knock sensor on the firewall and run w/o a knock sensor on the block. It is very dangerous to do on, but it is an option.

nj1266

iTrader: (6)

I agree with you, but that is the only thing that the OP has at the moment. So the O2 voltage can be used like old school DSMers used to do to tune their cars before widebands became cheap and available.

nj1266

iTrader: (6)

Could someone tell me the relationship between knock counts and timing pulled, ie, how many X knock counts must the ECU read/register before it pulls Y degree of timing?

jcsbanks

See my log here: http://john824.fotopic.net/p31381938.html click "full size" knock sum spike of 4 at 6400 with 2 degrees being pulled. I've pulled a degree out again at 6500 RPM, as I had 0 or 1 knock count when I ran 2 degrees less here.

I have previously had higher knock counts from running lower boost - just the part throttle maps can be quite aggressive.

nj1266

iTrader: (6)

RPM Tim Knock
2500 17 1
2531 17 1
2593 16 1
2656 15 1
2687 15 0
2750 14 0
2781 13 0
2843 13 0
2906 10 0
2968 8 0
3031 8 0
3062 8 0
3156 8 0
3218 8 0
3312 8 0
3375 8 0
3468 8 0
3593 8 0
3656 8 0
3781 8 0
3875 8 1
4000 8 0
4093 8 0
4187 8 0
4312 9 0
4406 8 0
4562 8 0
4625 8 0
4750 9 0
4843 10 0
4968 10 0
5062 10 0
5156 10 0
5281 11 0
5375 11 0
5468 11 0
5593 11 0
5656 11 0
5781 12 0
5875 12 0
5968 12 0
6031 12 0
6125 12 0
6218 13 0
6312 13 0
6406 13 0
6500 13 0
6562 12 6
6656 12 6
6750 13 6
6812 14 6
6875 14 6
7000 15 6
7031 15 5
7125 16 5
7187 16 5
7281 16 5
7343 16 5

Edited Girlie's log so it can be viewed better

Now please explain this to me on this part of the log:

6406 13 0
6500 13 0
6562 12 6
6656 12 6
6750 13 6
6812 14 6
6875 14 6
7000 15 6

The knock sum went from 0 to 6 and the ECU only pulled 1 degree of timing. Then the knock sum continued at 6 and the ECU advanced the timing from 13 to 14 and then to 15. Does that make sense? If the knock goes up shouldn't the ECU pull timing?

jcsbanks

The timing will be increasing with RPM and decreasing load. The knock sum effect is superimposed over that?

Girlie

Thanks!

That is what I cannot figure out. Also check out the other log I posted. With 2 counts of knock it pulls a degree of timing? Is it RPM dependent on how much timing is pulled via knock counts?

Also this part:

4093 8 0
4187 8 0
4312 9 0
4406 8 0
4562 8 0

It advances at 4300 then goes down at 4400? No knock present.

mchuang

iTrader: (11)

Yes it is. If you look at the knock sensor filter table it has a certain threshold for each rpm range and load point.