Knock during throttle stabs

[NJ_Phil]

Stock car and tune never knocked.
Put in Kooks long tubes, noticed slight knock.
Then installed 3bar map, ID1000s, seems to knock a little worse.

Noticed the collector flange was pretty close to the bell and I'm wondering if it's hitting when the motor torques.

Car runs great otherwise. MAF mode.

Any comments appreciated.

Thanks

[IARLLC]

The KR is very minor and I think it is fuel related. It looks to me like it goes lean for a moment when you crack the throttle. I am not sure if you reset LTFT before taking this log because while most of your table shows slightly rich (good, -1 to -4) the sensors show lean during quick throttle application. My thought is that your VE table might be slightly lean. I am guessing that you already made quite a few adjustments for the 3-bar, injectors, and headers but it might pay to do another run or two in VE and richen up by 1% or so.

As far as possible contact between the flange and bell you could check for "witness marks"...if none yet you could put a little ink or paint on the edge of the flange and try again to be sure. But honestly I think this is pretty minor KR.

Just my two cents.

[NJ_Phil]

Thanks. I'll richen it and pull some timing to see if it still knocks, pointing to the headers or something rattling around. I also have a brass alky filter that might be banging against the hood hinge.
Looking at it again seems like the EQ error oscillates +-5% just idling under CL. Haven't looked at many tunes but this seems pretty unstable.
Injector Duty cycle seems really low at idle (.4%) and saw pulse widths around .5 ms. Again not sure if that's normal, but sure the short pulse adder is setup properly

[5FDP]

Burst knock can also be the cause of knock during quick throttle movements, it's a predicted knock table in the computer that yanks timing away.

I'd reduce that by half. It's the base vs cylair delta table.

[schpenxel]

I really don't see any knock right when you have quick throttle changes? A few times you basically floor it and let off and it isn't showing anything.

I do see some under "TOTAL knock retard" which is likely just burst KR doing it's thing

One thing that I very much disagree with is that when people select "knock retard" in V3+ it pops up trying to use the generic "spark retard" sensor. THESE ARE NOT THE SAME THING!

Total knock retard = spark retard [sensor]

Plain knock retard doesn't have a generic sensor equivalent, even though HPT will try to get you to use spark retard [sensor] anyways

So--short version, my opinion, don't use generic sensors for spark retard in V3+

I'd love to see this changed.

[NJ_Phil]

Thanks for the tip on selecting the correct sensor. Although the easiest thing is to de-sensitive "something" LOL, I think the knock is real, as small as it is, and not going to adjust any tune settings....at least for now.
After looking at a few other logs I notice the knock on tip in/stab is related to the WB AFR value. During CL, the WB AFR oscillates +-3% lean to rich from commanded AFR every 500 milliseconds. The oscillation period matches the "switching" of the narrow band 02 sensors. I get the knock only if I tip in during a +3% lean WB cycle. Thought the AFR swings were the new injectors not stable under 1ms PW but the stock injectors did the same thing.

[schpenxel]

If you want to post another log or two I can check it out too--just by what I saw in that earlier one I wouldn't change anything knock sensor wise.

You could turn burst KR down probably if you haven't already--the stock settings are a bit much IMO

[NJ_Phil]

Log of a 2nd gear pull is attached. WB tracking of commanded AFR is much different today, CL EQ error is a lot tighter now for some reason, but I like it.

There is some slight knock in CL (< 2*)but more concerning knock around 4000 RPM so I got off it, just worried what's gonna happen when I feed it more air later on.
'
The only changes to the base tune was adding the 3bar map sensor and ID injector data.

To run richer in CL, should I just drop the base Stoich value set in the AFR table for 0.0 (100 gas)? Just want a quick way to richen it up for a test.


EDIT: Definatley leaned out under PE. I'll take care of that 1st thing.

[schpenxel]

Log of a 2nd gear pull is attached. WB tracking of commanded AFR is much different today, CL EQ error is a lot tighter now for some reason, but I like it.

There is some slight knock in CL (< 2*)but more concerning knock around 4000 RPM so I got off it, just worried what's gonna happen when I feed it more air later on.
'
The only changes to the base tune was adding the 3bar map sensor and ID injector data.

To run richer in CL, should I just drop the base Stoich value set in the AFR table for 0.0 (100 gas)? Just want a quick way to richen it up for a test.


EDIT: Definatley leaned out under PE. I'll take care of that 1st thing.

I see what you mean now--went a little lean @ WOT that one time. Still not bad though.

For just an overall quick change you could use lower stoich value or you could change the whole MAF/VE table(s) (whichever you're using or both) by say an increase of 5% to get a bit of a "global" change.

What wideband are you using? AEM 30-0333?

[NJ_Phil]

Yes, the 30-0333. Can't believe how nice it responds. Would have done anything for these kind of tools 20 years ago.

Just flattened my PE EQ ratio to 1.15, stock GM was all over the place. Then did some quick MAF tuning and increased airflow by 6% for that range and now EQ error is super close during PE.
I stabbed the throttle quicker on this last pull so the knock came back upon PE entry but lost the log when I changed the channels before saving it.

100% sure the knock up top was lean knock and I'll get rid of that easy enough. I'll post another log during CL and transients, just want to get it safe enough to pull 7,000, safe = 0 Knock.

I'm MAF only for now, 2bar SD after I learn a little more.

Thanks for the advice Carson.

[schpenxel]

What's the polling interval set at? It doesn't look as fast as mine--try the fastest setting if it's not there already. I suspect it's capable of faster than you're seeing right now... (though at some point it goes from fast to borderline jumpy, so may or may not help)

[NJ_Phil]

Crap.. Nice find. Had it set to 100ms, lowered it to 10ms just now. Got 3 pulls in before cops showed up but got WOT dialed in good enough. Motor seems to like running at 11.2 with the stock spark tables which I think is way to freakin' agressive but calling it good enough. Log and latest tune are attached but I just increased AF another 3% over 6150hz since it's still leaning out slightly on top. Knock in PE is gone (almost) as long as I don't stab the throttle going from CL to PE. That annoying tip in knock is still there but nothing showing on the Burst knock channel so not sure if messing with any burst related setting will help. Also, not sure how cylinder knock for 8 cylinders is determined since there's only 2 knock sensors. Maybe the ECM is using the crank angle to determine a "window" to relate a knock signal to a particular piston or something like that. Cylinders A and E seem like the troublemakers, not sure which is which though.

Still don't have the balls to take it to 7000, tires just want to break loose. I think it's ready for a test and tune day at Island Dragway to see what it does before the blower and cam go in.

[NJ_Phil]

Not to beat this knock on throttle stabs topic to death, but here's the noobs theory on what's going on:

Burst Knock retard has nothing to do with the knock sensors and is the ECM anticipating possible knock caused by rapid changes in cylander air mass.
If Cylair increases by the number of g/sec in the Base cs.Cylair table, over some period of time or crank pulses, the ECM will enter Burst Retard anticipting knock.
The spark will be reduced by the value in the table, and stay reduced for the number of "Hold Time" cranks, then decay to 0 retard over the number of "Decay Time" cranks, with all kinds of multipliers to play with these setting based on RPM, ECT and IAT.

During throttle stabs like shown below, the MAF spike increased DynCyl air and triggered the Burst knock motor protection. My table pulls 9.8* @ .036 g/sec delta which I'm guessing is what happened.

knock.png


End result....It's not a burst in knock, it's protection for a burst in cylinder airmass and you want it there so don't tune it out

[schpenxel]

Yep, burst KR isn't real knock, just a predictive pulling of timing to avoid real knock. A lot of people bring it down a good bit from the factory values.

When you get into higher airflow / FI setups you'll pretty much have to reduce burst KR one way or another--the airflow increases on those happen so fast burst KR goes crazy if left with factory or close to factory settings. I saw one yesterday that was bouncing between 14* and -2* timing because of burst KR. Needless to say that's not ideal

[10_SS]

BTW if you enter pure SD mode (tuning VE tables only, disabled MAF) then Burst Knock will basically not work right at all, especially with boost, you will have to disable it. It seems to only work good with MAF mode. Not sure what year/car you have but the 2010 Camaro E38 SD mode burst knock sucks.

[NJ_Phil]

BTW if you enter pure SD mode (tuning VE tables only, disabled MAF) then Burst Knock will basically not work right at all, especially with boost, you will have to disable it. It seems to only work good with MAF mode. Not sure what year/car you have but the 2010 Camaro E38 SD mode burst knock sucks.

Car is a 2012 C6 LS3 vette with E38. Putting a centrifugal and cam in it over the winter and going to try a mixed mode MAF/SD tune first. I would imagine airflow spikes are huge with your PD whipple and it takes time for fuel to catch up with the air. If I get into that situation where it actually causes true knock, I might try to figure a way to limit the ramp rate of the throttle plate opening, looks like 80-100ms should do it or find a way to back the timing off especially after a shift. Better than having a knock to reduce it. Still have a lot to learn

[10_SS]

Car is a 2012 C6 LS3 vette with E38. Putting a centrifugal and cam in it over the winter and going to try a mixed mode MAF/SD tune first. I would imagine airflow spikes are huge with your PD whipple and it takes time for fuel to catch up with the air. If I get into that situation where it actually causes true knock, I might try to figure a way to limit the ramp rate of the throttle plate opening, looks like 80-100ms should do it or find a way to back the timing off especially after a shift. Better than having a knock to reduce it. Still have a lot to learn

not sure if you were reading some past posts about this, but yes I had to reduce opening rate from 100 to 5. Works great, and dont really notice the delay.

[NJ_Phil]

Didn't know the opening rate could be changed but just saw the "Area Rate Limits" table, so thanks for the heads up.
Looks like a good plan for stomp would be to lower the burst rate till "real" knock shows up, then slow the open rate on the throttle body till knock goes away

[10_SS]

I dont think you need to reduce it all that much like I did, just stop the blade from slapping wide open. You wont see much change until you get it down to 5-10 as that was originally explained by HPT since I lowered it to 25 from 100 and saw no difference. I originally had 3, then 5. Havent tried 8 yet. OEM is 100.

It's under Engine/Airflow/Electronic Throttle/Area Rate Limits/Opening Rate. ECM 33020.

Log Accel Pedal Position and PE TPS and you can play with it with engine not running, key on.

Plus, this is something you can do temporarily while you play with other settings, transients, bypass valve setup, etc and gradually bring it back in to faster speeds.