Nistune Forums

would like to get a little poll going on who actually finds the knock system is reliable to be taken seriously... I cant hear any audible knock in mine but yet it still jumps to the knock maps..

how come the poll does not work ? it says you must enter two then it says thats too many ???

Disclaimer: I'm NA, my nissan isn't my daily driver and I can get a used ka24de for $400 easily so I'm not as cautious as I should be.

I'm a amateur, but from everything I've read with Nissan tuners they either disable it, or they use the knock sensor with an external op-amp.

For me the stock knock detection gives way too many false positives, I simply ignore it and go by ear and have even disabled the "bit" in my nistune tune as well as turning off the feedback in the timing map.

PL has the same opinion and I've heard it from other tuners

The factory knock sensing is too sensitive and can end up landing you in the knock tables a lot when it shouldnt be

On the other side I found that Pete was picking up knocking using headphones but with Nistune pulling the knock count from the ECU it wasnt always reporting all the time. After that I've reduced the importance of reporting knock from the ECU to focus on other things like sorting out address files and finding more tables/how they work etc

yeah now I have that light on there.. I can see that I am in the knock tables alot.. so I think I will disable it too.. there is nothing audible anyway.. and most of the time on decel it goes off from a light backfire.. it isnt a bang back fire just a pop..

went to turn it off and it does not seem to be doing it.. is there any flags in the consult that need to be ticked ? no matter what I try it keeps jumping to the knock maps.

you've turned off the knock feedback flag? and the knock flags in the timing map?

yep.. all flags disabled and knock sensor unticked..

My main issue with the factory system is that it's inconsistent. It seems to work fine on some cars (I guess by "work fine" I mean that it doesn't false trigger) and then others of the same model are ultra sensitive.

I have a couple of regular customers for example - one has an S15 that was jumping to knock maps with everything totally standard. Even std boost level and std exhaust! Yet there's an S13 that I've done heaps of tuning on that still has the sensor active - and it's making over 250rwkw on a daily basis!

Another problem is the actual area where knock sensing is active in the maps. Most of them only sense knock up to 3200rpm - this means that there's no knock sensing in the on-boost area. What's the point???? The general concensus is that above this point there's too much mechanical noise to reliably pick knock from mechanical noise.

PL

Interesting subject.

From my understanding and I could be wrong, doesn't the knock limit tables tell ECU that if knock count excedes those limits, then engine is knocking and ECU switches to knock maps???

Primary, Secondary knock limit tables seem to be different on HCR32 ECU, BNR32 ECU, etc, so it sort of coincides with what your saying about some ECU's being more sensitive than others. It might be the settings on Primary, Secondary knock limit tables are set to be more sensitive???

Settings on HCR32 ECU, 0-15. On BNR32 ECU, 0-8. It goes up to 255.

Although, can't see S13, S15 Primary, Secondary knock limit tables to see how they are set.

I was actually referring to different cars of the same type giving distinctly different results. So same ECU. Inferring that the knock sensors themselves can't be trusted more than anything else. Whether this is to do with the technology of the sensors (they're just a piezoelectric element) or what I don't know.

I'd expect to get more engine noise (and hence more false knock indication) on high mileage engines and/or highly modified engines. But the best behaved engine I tune has the most k's on it - and makes the most power. And the one that false triggers all the time is an S15 with under 100,000 k's on it and was false triggering when it was totally std!

PL

PL wrote:I was actually referring to different cars of the same type giving distinctly different results. So same ECU. Inferring that the knock sensors themselves can't be trusted more than anything else. Whether this is to do with the technology of the sensors (they're just a piezoelectric element) or what I don't know.

I'd expect to get more engine noise (and hence more false knock indication) on high mileage engines and/or highly modified engines. But the best behaved engine I tune has the most k's on it - and makes the most power. And the one that false triggers all the time is an S15 with under 100,000 k's on it and was false triggering when it was totally std!

PL

Yeah i would agree its a bit random, on some cars (mostly gtr) i can increase the knock flags right up to 6500rpm and get pretty accurate map switching ie when my gizzmo k_mon picks up noise the ecu switches to knock maps. On others i found the ecu switches accross too early.

I wish i had a depeer understanding of a few values in the adress's in the software such as knock limit 1 and 2, knock thresh holds and det ign retard. I have played with the knock limits and have good results on gtr but im flying blind

Hi Trent,

Yeah, I agree. It'd be nice to have more information on what some of the values do, but the reality is that only Mr. JECS knows this until people (usually Matt) disassemble the code - and/or people bench/dyno test things out. It's usually a combination because even after going through the code it's not always obvious what's going on.

Unfortunately there are people out there working on this, but because of the time consuming nature of the work, when they do work things out they won't tell anybody else! Which is understandable I guess. But the whole Nissan tuning community would be a whole lot better off if there was more information sharing.

The other problem is that the code for different ECU's can vary significantly. So once you've worked one out it often doesn't follow that the others will be the same. It's nice when they are though!

PL

well get this.. I have connected a lamp tot he boost output of the ECU... then in the boost knock maps I have put 100% and in the main boost maps all zero's so this in effect should tell me when the ECU turns over to the knock maps.. assuming that the ecu runs fuel, timing and boost knock maps all at the same time.. however it is still switching them over.. I have disabled all the flags in the timing map and also the feedback and it still does it.. unless there is more to the boost knock maps than that...

Great idea - but I'm guessing there may be more to it than that. Graeme's S15 behaves perfectly when I disable the Knock flag.

PL

Well yeah I think too... cause when the light is on it is still going well.. when it is off so it makes me think it is switching to the knock boost map for a different reason.. My car really does not like the heat... in the mornings first second and third if I hit it break traction... and when it is hot I am lucky to get thirds.. not that I try.. but yeah it just seems way down on torque..