"Super Knock" and you, a scratch n sniff discussion

Discussion in 'Mazdaspeed 3/6 ECU Tuning' started by Enki, May 8, 2017.

Watchers:
13 users.
  1. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    This is more of a hybrid hardware and software/firmware sort of thread so I wasn't really sure where to put it, but I guess here is as good as any so fucking deal with it.

    Hokay, so I've been mulling over this SAE whitepaper @rfinkle2 graciously allowed me to read yesterday and a couple things occurred to me; before I get into that, a bit about the whitepaper:

    It's from 2014 and entitled "Investigation on Pre-ignition and Super-Knock in Highly Boosted Gasoline Direct Injection Engines" and has several pretty charts and lots of sciency bullshit 99% of the people that read this won't give a sloppy toss about. The gist is that there's a form of knock that's exclusive to DI engines that's pretty much an order of magnitude (10x) worse than regular knock; the kind of random, untraceable knock that causes a rod to Kool-Aid Man out the side of the block. The paper covers the test conditions (low rpm, moderate load, slighly under stoich AFRs) and finds that the so called "super knock" is very random (read: unpredictable) but provides cylinder pressure spikes that happen sooner and with significantly more force than "regular" knock.

    So, back to my thoughts, and I'll keep this as short and sweet as possible for the sake of discussion:

    1. This is probably what's going on with completely stock cars that like to get off the highway then shit a rod.
    2. We know this sort of blowup doesn't happen with (properly) tuned cars or cars running corn mixes.

    Point 2 leads me to think about the primary difference (that could be a factor in detonation, at any rate) between stock/OTS and a full tune, which is fueling. The stock tune is full on retard rich, which washes down the cylinder walls and likely coats everything in the cylinder with fuel; it's this excess fuel, I think, that is contributing to ye olde ZZB.

    That said, why would a tune and/or corn mixes prevent this? Two reasons:
    1. Less overall fuel sprayed compared to stock in tuned cars
    2. Extra knock resistance from alcohols

    Go go Gadget discussion.
     
    • Like Like x 2
  2. ConeKiller

    ConeKiller Motorhead Greenie Member

    Joined:
    Feb 10, 2016
    Posts:
    186
    Trophy Points:
    63
    Location:
    babytown, IL
    Ratings:
    +291 / -1
    Is this super knock different than LSPI?
     
    ConeKiller, via a Nexus device, May 8, 2017
    #2
  3. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    Looks like it's pretty close to the same thing based on description alone.
     
  4. neganox

    neganox Feline Führer Moderator Platinum Member

    Joined:
    Feb 9, 2016
    Posts:
    1,100
    Trophy Points:
    113
    Location:
    St Johns, FL
    Ratings:
    +2,090 / -5
    Super knock because mazda wants to run all the timing at random as fuck times to make things efficient somehow. After I got a custom tune the only knock I ever saw was slight spool knock or if my methanol wasn't flowing properly. I've never seen those random 6+ KR readings since.
     
  5. ConeKiller

    ConeKiller Motorhead Greenie Member

    Joined:
    Feb 10, 2016
    Posts:
    186
    Trophy Points:
    63
    Location:
    babytown, IL
    Ratings:
    +291 / -1
    Is it bmep dependent? Citation of paper?
     
    ConeKiller, via a Nexus device, May 8, 2017
    #5
  6. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    This show what you're after?

    super knock.PNG

    The paper is pretty old and doesn't state what the dependencies are, only what they tested at (and yes it was under load/boost).
     
  7. ConeKiller

    ConeKiller Motorhead Greenie Member

    Joined:
    Feb 10, 2016
    Posts:
    186
    Trophy Points:
    63
    Location:
    babytown, IL
    Ratings:
    +291 / -1
    @xfeejayx @phate
    If super-duper knock = d detoknock in the graph it looks some kind of pre-ignition to me.
     
    ConeKiller, via a Nexus device, May 8, 2017
    #7
  8. xfeejayx

    xfeejayx Greenie Member

    Joined:
    Feb 22, 2016
    Posts:
    49
    Trophy Points:
    18
    Location:
    Chicagoland
    Ratings:
    +50 / -0
    Remember me saying "rate of pressure rise"?

    BTW, early ignition timing isn't the cause of this. That "ECU is advancing timing until it sees knock" story, from my experiences, is likely wrong. This is pre-ignition, likely caused by shit injectors, seals, shit on pistons, carbon land above the ring, .etc.
     
    • Like Like x 1
  9. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    From what I was reading, it's way, way worse than pre-ignition and is specific to boosted DI vehicles.
     
  10. ConeKiller

    ConeKiller Motorhead Greenie Member

    Joined:
    Feb 10, 2016
    Posts:
    186
    Trophy Points:
    63
    Location:
    babytown, IL
    Ratings:
    +291 / -1
    can you give us a citation please
     
    ConeKiller, via a Nexus device, May 8, 2017
    #10
  11. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    The document isn't mine to upload; abstract per Google search for "super knock:"

     
  12. xfeejayx

    xfeejayx Greenie Member

    Joined:
    Feb 22, 2016
    Posts:
    49
    Trophy Points:
    18
    Location:
    Chicagoland
    Ratings:
    +50 / -0
    It's all pre-ignition. Pre-ignition simply means that it is igniting before spark is making it ignite. Look at the graph you posted, it's pre-igniting.

    Conekiller is asking for a citation, not the whole paper.
     
    • Like Like x 2
  13. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    Well, this is the one from the paper:
    CITATION: Wang, Z., Liu, H., Song, T., Xu, Y. et al., "Investigation on Pre-ignition and Super-Knock in Highly Boosted
    Gasoline Direct Injection Engines," SAE Technical Paper 2014-01-1212, 2014, doi:10.4271/2014-01-1212.
     
    • Like Like x 3
  14. xfeejayx

    xfeejayx Greenie Member

    Joined:
    Feb 22, 2016
    Posts:
    49
    Trophy Points:
    18
    Location:
    Chicagoland
    Ratings:
    +50 / -0
    Read the paper, then remembered Lex has already covered this. I agree with most of what Lex is saying.
    http://www.stratifiedauto.com/blog/...the-mazdaspeed-disi-and-ford-ecoboost-motors/

    As far as why it's prevented by tuning...well, I don't know that tuning prevents it, but there seems to be correlation. Simply running alcohol to slow down the burn, and reducing low speed load limits may be enough to completely prevent this.

    My input: It's caused largely by the same conditions that have always caused pre-ignition>>detonation in reciprocating engines. it's just made worse because DI engines have a bunch of fuel in various states of liquid/vapor/mix being shot around onto different potential hot spots.

    TL;DR for paper: Superknock is just the result of really early pre-ignition, partly during compression stroke; this causes some absolutely ridiculous rates of pressure rise. Early enough that there is actually heat release before the injection event is supposed to happen. It can be controlled by various methods, including injection strategy.
     
    • Like Like x 2
  15. ConeKiller

    ConeKiller Motorhead Greenie Member

    Joined:
    Feb 10, 2016
    Posts:
    186
    Trophy Points:
    63
    Location:
    babytown, IL
    Ratings:
    +291 / -1
    Funny, I thought about the exact same article when this thread came up.

    Good summary.


    I couldn't easily get access to this specific paper, but these authors seem to be really deeply investigating/somewhat salami slicing their work with "super-knock". I found a freely distributed article from the authors which talks about the same phenomenon, here

    Regarding the relationship between pre-ignition and "super-knock":
    [​IMG]

    Interestingly the authors differentiate between knock and detonation (news to me), and it seems that this "super-knock" primarly relates the magnitude of the increase cylinder pressures (detonation v deflagration)[which is what feejay mentioned previously]:
    [​IMG]

    I think the best takeaway from the article is this table,which addresses some easy ways to prevent pre-ignition generally, but also "super-knock" specifically. Much of this echos what's presented in the stratified article.

    [​IMG]

    Either reduce the temperature of the intake charge (more efficient flow or intercooling) or reduce the cylinder pressures(boost).
    Looking at the axes of the graph, to stay in the safe range: don't run a shit ton of boost/high load, particularly at low RPM's (aka, don't go WOT in 6th) and watch your BATS and/or run meth & eth to help absorb some of the latent heat of the intake charge and combustion (eth & water have a higher thermal capacity than 93).

    @phate; in particular, this is one of the reasons I'm interested in water injection w/low rpm boost.
     
    • Like Like x 3
  16. xfeejayx

    xfeejayx Greenie Member

    Joined:
    Feb 22, 2016
    Posts:
    49
    Trophy Points:
    18
    Location:
    Chicagoland
    Ratings:
    +50 / -0
    Remind me next time you're by, I'll pull it up for you. (speaking of which, tonight?)

    Deto vs knock = cause vs effect. Detonation is the kaboom, knock is the reaction from your block/piston/rings/bearings/window maker. When we have hardcore discussions, we actually do differentiate between the two because there can be different cures/bandaids for them. Or maybe you can handle some detonation just fine until it induces knock. Almost all production engines, as you already know, measure the effect...mostly because real time cylinder pressure monitoring is expensive, and the sensors/amps are not very durable.

    That's a good, accurate summary of the mechanical cause of SK. The avoidance is the same as it has always been for detonation on gasoline engines, it just matters more now since there are a bunch of pockets of fuel sitting around with various local AFRs.
     
    • Like Like x 2
  17. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    Gonna necro THE FUCK out of this thread because of an older podcast a buddy asked me to listen to:



    Quick search didn't show anything on this video and specific information, so here it is; I've taken notes on the video that was linked, and they are as follows:


    LSPI is low speed pre ignition (we know this)

    The LSPI flamefront starts at the rings and works to the middle of the cylinder.

    The oil and specifically mixing of oil and fuel is a major factor.

    Side firing DI (like ours) has this the worst, since the cylinders wash down more.

    The base oil mixing with fuel is not the issue, it's the oil additives that result in potentially low octane products that cause the LSPI events.

    Oils with high calcium and/or sodium are likely to induce LSPI; the values are not linear as far as knock induction goes, but higher concentrations will be more effective at producing knock and LSPI.

    Molybdenum and ZDDP can help suppress the chemical change that is responsible for the knock, while magnesium tends to not sway one way or the other.

    Fuels that have a cooler distillation curve are less likely to encounter LSPI.
    The distillation curve shows what temperature is needed for a % of fuel to evaporate out; lower is better and higher blends of ethanol are better (not in the oil at higher temps).

    DEXOS 1 Gen 2 oils are formulated around these findings and will have lower calcium/sodium with more moly and zinc.

    The exact chemical process that causes this is still under investigation.
     
    • Like Like x 9
  18. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    Haaaaaaaay just stumbled across this, good watch and great channel:

     
    • Like Like x 6
  19. Awafrican

    Awafrican Moderator Gold Member

    Joined:
    Feb 11, 2016
    Posts:
    3,586
    Trophy Points:
    113
    Location:
    Canada
    Ratings:
    +2,851 / -11
    Awafrican, via a mobile device, May 26, 2019
    #19
  20. Enki

    Enki Motorhead Platinum Member

    Joined:
    Feb 10, 2016
    Posts:
    2,244
    Trophy Points:
    113
    Location:
    Tucson
    Ratings:
    +3,236 / -3
    You know how when you pull your dipstick after even a short drive on fresh T6 how it smells like fuel? Yeah. That's really, REALLY bad.
     
    • Like Like x 2
Loading...

Share This Page

Users Viewing Thread (Users: 0, Guests: 0)