[winsome]

Dear all

Can those of you who have tried the ap log of knock share their knowledge with me, i have tried stage 1 and 2 both of them have know over -1500, anyone else have the same reading? i am using shell v power ron98 in hong kong.

however friends who are using stage 2 with similar reading have tracked many time without problem, what is that suppose to mean?

Should I be worry at all? any know the usual stock car knock level?

Thanks

Winsome

 

[Jm-Imports]

i believe the original maps had a lot of knock..


thats why its essential you have custom tune..

speak to ben @ GTC

 

[David.Yu]

I think Ben and Thistle are still working through and determining just how much knock is acceptable.

The ECU is very quick at adapting and changing the timing if you get to a dangerous level and tuning it so there's no knock at all is probably robbing the tune of quite a lot of power and response.

 

[thistle]

Stock US car on 91 PON (~95 RON) runs about -1400, but this example was in Salt Lake City which is cooler than Hong Kong I believe.

For custom tunes I would prefer the knock to be about half that as it leaves a greater margin before the approx -2000 when the ECU switches to low octane maps.

The staged maps are being revised based on the latest findings, but it is an improvement rather than a safety issue.

 

[ARW]

Knock seams to be dependent on the actual car. Mine (standard apart from a GTC Y-pipe) was not happy with a stage 2 99Ron Map (running on Shell 99Ron V-power) and gave peak knock of -1700 plus so I ended going down the custom route which I have to say after a couple iteration of fine tuning from Ben & thistle (thanks again guys) I think is almost spot on now (need to do a couple more runs to check).

Out of interest I recently put the stock map back on to get a baseline and even that was still giving knock of -307, which I think is viewed as acceptable but just goes to show different cars behave slightly differently.

 

[Rich-GT]

Hi Guys

Have been doing quite a lot of logging and analysis over the last few days, not quite ready to share the info yet, hopefully in the next 24 hours or so when I have crunched it down and turned it into something presentable. At this point I will just share a few anecdotes.

All of the following is for a standard car, no Y pipe, and the biggest set of comparatave measurements I have is for a standing start through the gears. See the graphs in this thread for examples.

http://www.gtr.co.uk/forum/124538-update-your-accessport-knock-now-included-3.html#post1175868

The highest knock reading I have seen is -903 and that was with the Stage 1 map and Tesco 99.

Lowest Knock was with the standard map, again with T99, when there were no knock events. (Not tested with VPower, would expect the same)

Interestingly while warming the car up for this run I data logged and was supprised to see some knock during the warm up. I did the warm up in Auto and knock was occuring in the higher gears, just like when you get pinking in older cars when pulling away in too high a gear. The knock was mild but I just knew that I did not like the way the GT-R gets into such a high gear at such low revs.


See comments in the fuel thread.

http://www.gtr.co.uk/forum/124924-fuel-petrol-your-gtr.html#post1176441

But consistantly throughout the tests I have been getting more knock with Tesco 99 than with VPower 99. I know that this is at odds with thistles results so may just be the batch of T99 that I had? Am now back on VPower and getting the same results as before the change to T99.

Final anecdote its that when doing the run through the gears test I see more knock when doing it in Auto than manual. I assume this is because in Auto the car changes gear earlier? Perhaps thistle can explaion this? I am now doing all tests in auto, not just because of this but because they are more consistant, with no missed gear changes, also you can watch the road and not the rev counter.


Rich

 

[thistle]

I haven't tried V-power in the GTR though, so I should do so.

Regarding manual vs auto, there could be something in the different engine speeds, but overall the apparent trends with knock analysis can fool you since there is a lot of randomness/inconsistency to it. On the Evo for example, I spent about a few weeks logging and retarding individual cylinders with respect to knock to see if it was concentrated on one cylinder. I thought I had reached a conclusion, would test again and it would not be consistent.

Part throttle knock control is not really a problem, but a feature of many modern turbocharged engines. On the present staged maps though, there is generous timing in the part throttle areas to give response, turbo spool etc which may encourage a bit more knock control activity though.

 

[Rich-GT]

I suspect this chart oversimlyfies things, but it's either this or a whole mass of data.



Chart shows the maps in increaing order of performance, these being stock and the Stage 1 & 2 Cobb maps initially supplied with the AP. Then the GTC 99 RON map and finally 3 Custom RPM Boost based maps, the first two with the ignition retarded 4 & 2 degrees relative to the third map.

The number in the Tesco99 & VPower column is the number of knock events during the run. Typically 3rd gear is the worst, knock is very rarely seen in 1st.

I have then colour coded and "guessed" the missing slots in the matrix.

Green is no knock
Yellow is a single knock event
Orange is 2 or 3 knock events
Red is 4 or more knock events.

The following perhaps overcautious but I would suggest.
Green and yellow use with complete safety.
Orange use with caution, That last run at the drag strip.
Red do not use. Also in the red there is often a doubleknock events and a knock reading is present from 2nd gear on throughout the run.


So remember this is my standard car with my maps, results may be different with Y Pipe / full exhaust / other mods. Also just one batch of Tesco 99, although I will not be trying it again, but may try some BP102.


Rich

 

[thistle]

It is a sensible analysis.

My engine runs Tesco 99 and runs the GTC Stage 2 99RON RPM map in your yellow or orange category, so I've left it like that, although more timing is not necessarily better. I've not flashed it now for a few weeks, which is a record for me being a tinkerer. It is colder up here, and I very rarely have the car sitting in queues which would heatsoak it.

 

[Rager54]

Guys sorry for my ignorance but what is 'knock'? I get my GTR next Saturday (at last) and when run-in would like to get the COBB access port and not sure whether to try stage 1 or 2. My car will have Y-pipe fitted from new. If you experience knock does it cause any damage to the engine?

 

[thistle]

Knock is the spontaneous ignition of unburnt air/fuel mixture in the cylinder before the flame front has smoothly reached that point. If severe, it produces cylinder pressures up to ten times normal levels which will damage the engine.

However, modern turbocharged engines on full throttle invariably run on the limit of knock where knock sensors constantly tune the ignition timing to stay at that point. This is because at the knock threshold you have the best emissions/economy/torque.

On the GTR, the knock control is excellent and preventing it getting out of hand.

I would run stage 1 with your Y-pipe.

 

[Mark B]

Can I ask how people are getting such high knock counts, mentioned above? Surely the factory knock control should be retarding timing enough to prevent it, or is it that the boost and fuelling after modification, removes scope for retarding enough to combat it?

 

[thistle]

What we're logging IS the factory knock control at work, retarding timing.

 

[Amos]

Surely there should be no Knock in these cars, I thought the KNOCK word was BAD....!!!

 

[Rager54]

I have to say I agree with Amos all this knock stuff makes me nervous!!!!!

 

[Rich-GT]

I have to say I agree with Amos all this knock stuff makes me nervous!!!!!

Guys Guys Guys... Most modern cars have knock detection built into the ECU, the level of knock that is detected and then adjusted for is very low. All cars will rely on it to some degree to adjust for lower octane fuel / adverse load conditions etc that are outside the standard parammeters catered for in the maps.

Just be thankful that with the Cobb AP you can monitor a range of parammeters including the knock events, which are then immediatly adjusted for. Without it you would be completely oblivious to what was going on.


Rich

 

[Rager54]

Thanks Rich useful info...... I am still keen to get the Cobb access port just..... finding out as much as I can before I do

 

[thistle]

Would it be better that I'd never found the knock logging and left it as it was before where no one knew? Or do as someone perhaps sensibly suggested and make it a tuner only logged item otherwise we'd have a flood of anxious customers?

Or would it be better that it is found, documented, open and discussed? If you believe this, please reconsider what I've written about what the standard car does and how this system is working to prevent damaging knock.

I have never had a knock related engine failure on a engine running functional knock control like this system. I HAVE had knock related damage on a standard Subaru engine running R35 GTR type outputs mapped for no knock with testing in many conditions to provide a margin for safety. Obviously the margin wasn't enough as a single knock at high speed where it previously hadn't knocked caused damage. This engine was NOT running knock control.

I'd rather have a system like this that works all the time you've got any significant engine load, and behaves appropriately up to the rev limiter. A small bit of random activity shows you it is working, but if you chase all of this down to remove it in all circumstances you end up with quite a flat car, and even standard cars don't run like this. You can remove it one day, just enough so you don't make the car flat, and then in another circumstance it can return. Even leaving 2 degrees margin from knock isn't enough to eradicate it in all circumstances, knock threshold can move more than 2 degrees if you have a bit of oil mist in your intake at the wrong time, or the atmospheric conditions change.

Additionally, the ECU system has noise thresholds that are based on probability. Many systems take out a degree because there is a rapid noise transient on throttle movement that is enough to breach the knock threshold at that point. It is a compromise between sensitivity and accuracy. The system is listening during each combustion event and a noisy combustion that you might just hear in det cans if you had twenty of them has the ECU reacting. It is controlling it before it gets out of hand. Nissan also have a system where the ignition timing ramps up behind the map on throttle openings to actually reduce knock, it is quite neat.

I'm presently trying to disassemble the ignition routines. They are huge and complex. It does seem that Nissan is using a model of calculated combustion times rather than a lookup of ignition timing in degrees. The rabbit hole goes deeper.

 

[Rager54]

Agreed on all Thistle.... I have to say I have found the forum very open and useful... I think there are just a few of us baffled by the technical speak and are just on a very fast learning curve. I have just been on the phone to Ben and I have to say you all seem like proper guys who know there stuff (and then some looking at your mails!!!).
I guess I have never paid 60k for a car and considered tuning it out of the box before whilst it still has 3 years warranty! Just need to ask lots of questions prior to taking the plunge.
I have to say you have been very helpful and I feel reassured by your swift and thorough responses!

 

[[email protected] M/S]

Thistle,

It does seem that Nissan is using a model of calculated combustion times rather than a lookup of ignition timing in degrees.

I think Nissan try to work with predictive detonation control more than reactive. , the 350Z ECU uses a figure that is the calculated time to peak cylinder pressure rather than a ignition timing angle. The system are so complicated nowdays , so much working in the back ground that you cant see while mapping