kf6ytc

iTrader: (10)

When loging in ecu flash, how do you determan that your ecu is pulling timing based on knock count?

bhcevo

To answer your question you will need access to the ignition timing maps that were running on your ecu at the time of the log. Ecuflash will get you this info.
Assuming you have these map(s) heres how.

If you got knockcount > 0 the question is...how much timing was being pulled, if any? There's no absolutely sure fire way to figure it out, because we don't yet know how to calculate exact load. Having said that, here's what I do to get a good guestimate.

Let's say you know you were at rpm 5000 with a timing of 7 and a knocksum bigger than zero. How much timing was pulled? To answer this questio you need to know what the timing should have been in the absence of knock. Ideally, to get this information you figure out your load, you look at your high octane ignition timing map and you find the timing value for that load at that rpm. Let's say you do this and the value at that cell reads 10. There you go. You had 3 degrees of timing pulled.

The problem is you do not know your load precisely or at all. So you don't know if you were at load 240, or 220 or 200 (typically). At 240 load and 5000 rpm the timing value is 8, which means you had 1 degree pulled. At 220 load and 5000 rpm the ignition map reads 10, which means you had 3 degrees pulled. At 200 load and 5000 rpm the ignition map reads 12, which means you had 5 degrees pulled. Which one?

The first thing I do is calculate load using the approximate formula: load = AirFlow*852/rpm. You can then assume this load is correct and go ahead with the analysis I offered above. This works out pretty well for me much of the time, but I don't rely on this analysis only.

What I also do is look at the logged timing curve as a function of time. It should slowly increase with rpms during the pull. A dip in the curve that coincides with a knocksum > 0 is an unambiguous indication of pulled timing. To figure out how much timing was pulled you can use the timing numbers before and after the knock event to pinpoint pretty well where you were on the ignition timing map. To do this assume that a knocksum of zero means no timing was pulled.

Sometimes what you have a timing curve that slowly increases, no dips, but timing was pulled everywhere so the above analysis is no help. Usually this coincides with knocksum > 0 during the whole pull. Look at your ignition timing map, allow the approximate formula to get you within the right 3 load columns, then use logic appropiately. Haha.

Also keep in mind that the ECU interpolates (linearly?) between load cells. So it is
quite possible to have a timing of 12 at rpm 5000 with no pulled timing and knocksum = 0, even though you don't have an cell on your ignition map that reads 12 (at rpm 5000 and high load). All this means is that your load was such that you were exactly halfway between the cell that read 14 and the one that read 10.

Another useful assumption is that AirFlow/RPM is proportional to load, even if you don't know the constant exactly. So if AirFlow/RPM stayed pretty constant during a pull then you know you never left the same load column, even if you don't know which column it was. This is most likely correct to first order, but it could be wrong.

It helps to think like an engineer or scientist when doing this sort of thing. Lots of different data, no exact analysis possible....use logic and the scientific method to figure out the most probably answer.

Finally, I am defining pulled timing as any deviation from the high octane ignition timing map.

vboy425

iTrader: (19)

this is an really easy answer, log your timing and see if it dips. for example

let say this a log see how the timing as it climb up and start to go back down?? the turbo hit full spool at about 3500-3800rpm base on load condition, at that momment your timing should drop down to like 4 or 5 degree of timing and form there it should steady climb up

rpm timing
3000 5
3500 6
4000 7
4500 5
5000 5
5500 6
6000 7
6500 14
7000 18

bhcevo

How dare you answer his question in the most straightforward way!!!

I must be the only one that subscribes to the maxim never use one word where ten will do.

vboy425

iTrader: (19)

LOL, sorry dude

honesty your post is good i understand it totally, but for someone who have never tune anything before going to be confuse.

bhcevo

Oh yeah, keep in mind that if you got knock during the whole pull then you might have no dip because timing was pulled everywhere.

I mention this because I had a bad knock problem and saw this more than once.

vboy425

iTrader: (19)

WTF ?? you're not making any sense. if it's knocking it's pulling timing and if it's pulling timing it's diping.

kf6ytc

iTrader: (10)

Awesome guys!! Thanks alot! I'm going to keep asking these basic questions as I get them, Not only will it help me, but it will help build a very nice database for the many new people partaking in this new method of tuning.


TJ

bhcevo

Hehe.

What I meant is that imagine it knocks during every single point you logged at WOT. This means that your timing log looks like

2
3
5
7
10

Now you put in 100 octane and redo the same pull exactly (lol) and get
with 0 knocksum

6
8
10
12
16

The first timing log doesn't "dip" if you plot it, yet timing was pulled everywhere.

When my car was knocking real bad I would see this sort of thing.

vboy425

iTrader: (19)

okay now i see your logic. Unless you have some ****ty *** logger that log like 4 frames per sceond. what are the chances of that happend, beside you of course. I have Pocketlogger from pcokerlogger.com and evo scan. I haven't once see that happend. But good logic/point though.


Also sometime you have to make a decision if it's really knocking or not. at this point it's come down to experience.

bhcevo

Is there a way to get evoscan to log faster?

I have ***ty short logs because I'm too much of a chicken to go WOT for more than a few seconds. But maybe I'm logging slow for some reason on top of that.

Now I'm sad.

MalibuJack

iTrader: (5)

Turn on Turbo scan mode in the dropdown.. It should log quite a bit faster..

Though the speed of your laptop, the cable your using, and whether or not you minimize the screen while your logging will have an affect on how fast it logs, I'm getting at least 10-30 samples a second, roughly 5-10 complete lines a second, which is more than enough for me to rough in my data..

nj1266

iTrader: (6)

You should log from 2500 rpm to 7300 rpm at WOT in 3rd or 4th gear to get accurate data, unless you are tuning for part throttle. Your data is more than likely inaccurate. I prefer no data for inaccurate data.

bhcevo

YOU DARE CRITICIZE MY DATA!!!!

Prepare for death......death by hubba bubbaaaaaa!!!!!!


<cue the monkeys>

bhcevo

Thanks Malibu. I'm getting about 5 lines per second or so. Hmm...seems kinda on the low end.

This just occurred to me. If I only logged the handful of parameters I cared about I'd log faster no? In this case I would uncheck lots of the boxes in the evoscan big window...