Spark timing nightmares & questions

[ringram]

Hi guys,
Some questions for the experienced amongst us.
It seems no matter what timing Im running whether at a certain level or 10 deg retarded I can always get the odd little spike of KR.
Its not logged as the raggard saw tooth of actual sustained knock, more like the odd litte spike now and again, generally during throttle and mid to large airflow transitions. This is in OS3 on an OLSD tune.

My question is should I worry about this, or is the odd tick now and again indicative of "false" knock or just the general action of a busy engine during transitions? (ie) Should my logs always and never have any KR logged at all, not even 0.2 deg etc?

Im going to do some tuning on a dyno after my cam install in the mid term and I was also going to run my stock (puke) tune and see if it exhibits the same behaviour.

Just wondering what you guys see and experience.

[oztracktuning]

Some will say that some knock is OK, its not necessarily bad. Plus some of it may be false recording and not be spark related at all.

The proof of the pudding would be what the engine looks liek if a car at your level of knock had its heads off and someone had a look inside. Recently had the heads off were and it looked fine inside even though it had lots of advance at full throttle and some knock retard <2degrees mostly at under 2000rpm at high loads.

[ringram]

Thats exactly where most of mine is, on the motorway top gear under 2000rpm.
Ok interesting, thanks dude.

I wonder if anyone would care to comment?

[ToplessTexan]

I'll offer this observation which may have nothing to do with your KR. :)

In most thresholding based energy detection systems you cannot completely eliminate false detections w/o also losing true detections. Similarly, you cannot maximize detection sensitivity without generating false alarms. You usually have to spec both an adequate detection sensitivity and an acceptable false alarm rate. Generally you try to restrain occurance of such false alarms to regions where you can discriminate against the detection for other reasons (such as the detection not being physically meaningful given the context.)

I'm speculating here but it wouldn't surprise me at all if there were some false knock detections. (I don't mean pipes banging and the like.) It might or might not surprise me if these "detections" were able to be recognized as false and not reacted to as if they were true knock detections as that could require substantially more sophisticated software/firmware/hardware. If the reaction to the false detection (retarding of timing) is small enough and the work involved in characterizing it as false is hard enough, it might be a very good engineering decision to just call it a day. ;)

[VetPet]

Top gear at under 2,000 rpm I might suspect that you'd get a bit of light knock. What gear are you in? Is there any difference depending on the weather (dry vs humid)? Did you buy any fuel from someplace you normally don't? Will it still do it in the next lower gear? Do you actually hear any knock or is this strictly what you're seeing in scan mode? I'd say you may be trying to chase a ghost here and that's it's not worth losing sleep over.

:cheers:

[caver]

Delta knock as soon as you mash it and there is a large airflow change the ecu anticipates knock and reacts accordingly.

[GMPX]

My question is should I worry about this, or is the odd tick now and again indicative of "false" knock or just the general action of a busy engine during transitions? (ie) Should my logs always and never have any KR logged at all, not even 0.2 deg etc?

I have zero, but that is running 98Oct and total of about 26deg timing, not sure what your UK fuel is like, I assume it's probably pretty good because all the BMW's, Mercs etc sold here must run on 98Octane.

Anyway, one thing to do is make the knock detection more sensitive cylinder by cylinder {B6244}.
Drop the values for one cylinder down, maybe to 2 or 3 at the point this 0.2deg normally happens, this will make it VERY sensitive to the knock signals (or other noise). You might even figure out which cylinder might be causing the problem.
So once you figure out which cylinder it is you could de-sensitize it a little, if you are pulling 10 degrees out from that point and you are still seeing knock it doesn't sound like the real thing to me.
Some tuners would rightly say you need to listen to the engine to see if it is really knocking.

Cheers,
Ross

[dfe1]

Not trying to be a smart-ass, but a friend of mine has an expression that I think applies here-- picking the fly sh*t out of the pepper. Even though FlashScan is nearly perfect software, we don't live in a nearly perfect world. Small amounts of knock, whether real or phantom, aren't all that unusual-- especially at low rpm when heavy throttle is applied. In theory, as long as the amount of knock doesn't exceed the system's knock retard capacity, no harm is done. Obviously, the ideal situation is to have no knock at all, but an occasional small spike shouldn't be anything to worry about. Just my $.02.