question about evoscan knock sum!
#91
Evolved Member
The issue here is that the stock ECU can and will pull timing with 3 counts of knock and at 6 counts will start to interpolate between the high and low maps. you can find instances where a 3 count of knock won't pull timing but this is rare.
I'm not trying to start an argument, but you ask if I2r99gst can read English. Have you actually looked at the grammar and spelling in your post?
MB
#92
Evolved Member
iTrader: (2)
Originally Posted by evoRS06
can you read english?
Originally Posted by evoRS06
if your tunning yourself tune for scarace knock
Originally Posted by evoRS06
the combustiion engine
Originally Posted by evoRS06
l2r99gst maybe your right
And thank you...I am right.
Also, since I can't really read English, maybe you can explain why you like writing contradictions to yourself...maybe a split personality?:
Originally Posted by evoRS06
to the original poster its the graph your wanting to tune to, it WILL NOT BE SMOOTH if your knocking and holding counts
Maybe you woke up on the wrong side of the bed. Why are you acting like an *** when people are trying to help you understand the workings of the stock ECU and your misundstanding of knock?
Also, to answer your own contradicting quotes above, a graph (assuming that you are meaning a dyno graph) can be smooth or rough with the presence of knock. It matters if the increase in knock from one point in time to the next is a rapid increase or not. This is what causes the rough graph (because of the pulled timing), not solely the fact that you are knocking.
Eric
Last edited by l2r99gst; Jun 26, 2007 at 02:06 PM.
#94
Evolving Member
iTrader: (1)
Join Date: Jul 2006
Location: Oak Harbor, WA
Posts: 152
Likes: 0
Received 0 Likes
on
0 Posts
i dont see a contradiction there. I said with more knock, as in a quanity, and that it wont be smooth if your holding the counts. I just got off the phone with lucas english and he seems to be on the same page; but i guess he doesn't know about the evo stock ecu too, huh? he's been tuning the 4G63 for 10 years and just ran 11.5 on an '04 SSL. basics are you dont want knock, but if you get it in a map that creates more power then so be it. Im not saying go run 6 to 8 counts constant and expect more power or your motor to last, but if you hit 6 at 3k or 3.5k and it tappers down to 0 then the in-between isn't knock, its the ecus way of bringing the timing back up to the target degree on the map. Look man im not trying to bash you and if its coming off that way im sorry, im just trying to share what i know.
#96
Evolved Member
iTrader: (2)
If you hit 6 counts of knock at 3k or 3.5k, then the ECU will pull 2 degrees of timing and because of the decay function it will pull timing after that, until the knock count reaches 0, where, as you stated, there really isn't any knock. The knock count will decay back to 0, as long as no more knock occurs.
That is the whole point why you should tune just to the point of knock and no more than a count or two, preferrably 0, if you know the point at which you will begin to knock. Timing being pulled reduces power and the knock decay algorithm in the stock ECU will continue pulling timing during that decay time. This decay will take longer during a higher knock count to get back to 0 and the intended timing in your ignition map.
In that same scenario that you stated where 6 counts of knock appeared at 3k RPM and tapered to 0...if you simply removed 1 or two degrees at 3k, then you may have 0 knock the whole run, get more timing, as the map wants, and make more power. I don't where you think less timing is giving you more power.
Also, I know you're not trying to bash me and are just sharing what you have experienced or heard, but in my years with DSMs and now Evos and learning a lot about the stock ECU from the DSMLink creators and now through the disassembly of the Evo ECU, I'm simply sharing what I know, too.
Eric
Last edited by l2r99gst; Jun 26, 2007 at 05:04 PM.
#97
EvoM Guru
iTrader: (6)
EvoRS06 - this is the answer:
- TUNE FOR ZERO KNOCK -
why?
1) When the ECU detects knock it pulls timing, so you might as well tune for it in the first place!!!
2) When you get say 5 counts of knock @ 3500/250 you are more likely to knock at 4000/250, and 4500/250 - it's what I call "a knock avalanche" - knock upsteam in the map induces knock downstream that wouldn't normally be there.
- TUNE FOR ZERO KNOCK -
why?
1) When the ECU detects knock it pulls timing, so you might as well tune for it in the first place!!!
2) When you get say 5 counts of knock @ 3500/250 you are more likely to knock at 4000/250, and 4500/250 - it's what I call "a knock avalanche" - knock upsteam in the map induces knock downstream that wouldn't normally be there.
Last edited by tephra; Jun 26, 2007 at 05:21 PM. Reason: remove personal attack
#98
Evolved Member
iTrader: (9)
Sometimes knock happens without respect to tune (I get knock in first gear at mid throttle before boost every time), engine temp and heat soak also makes a decent tune knock, but its hard to make power with significant knock when timing goes off track and fuel maps switch so tuune befor knock threshold
#99
Evolved Member
iTrader: (22)
lets not forget logging software, can users really believe in all of them? Im sure some say there is some sort of knock when actually the OEM software says otherwise. Software algorithms are not all the same, just like adaware, spyware and other software on your home computer says,no virsus found or spyware or whatever. You still could have a virus even though they say otherwise.
So some of the confusiong not is coming from how the heck the the OEM software does its magic, but sometimes its the logging software that is confusing you!!
So some of the confusiong not is coming from how the heck the the OEM software does its magic, but sometimes its the logging software that is confusing you!!
The basic algorithm for those programs says "look for this type of behaviour and if you see it knock it over". New definitions just extend the types of behaviour to look for...they don't change the basic algo of the program but now "poof" the malware can now be found. Update those and they should find the malware and if they don't then get your money back .
That has nothing to do with the values that the logging programs that the majority of this forum use...which wherever possible explicitly ask the ECU for whatever values it is using for its various comparisons. You could have made this basic argument for Calculated load...which was a best guess effort outside the ECU to approximate what load cell the ECU was in but not knocksum.
To be fair people are always afraid of something new...flashback a little over a year on this forum and other tuners were questioning the validity of certain values returned from these programs as well...my how their tunes (pardon the pun) have changed .
Last edited by codgi; Jun 27, 2007 at 12:54 AM.
#100
Account Disabled
The OEM ROM is not transfereing memory dump to the datalog software. Its reading in real time the interpretation from the OEM. Therefore, the datalog software can and will miss this "real time data" on occasion. The sampling rate is not infinite, there is room to "skip a beat" it would be better if you can download froma memory dump, but we cant.
This is completely untrue. Don't start with this kind of junk. The logging software (EvoScan, Mitsulogger, LogWorks) all read the knock sum value from the ECU. There is no modification of that data.
Its quite easy to prove this for yourself by watching the timing and the knock sum together. If what you propose were true, then the logging program could show knock counts without the ECU pulling any timing. This never happens for significant knock sums (4+). I have never seen an instance where EvoScan showed a knock sum of 2 (or more) where the ECU did not pull back timing.
And if you actually could read the programming languages for the software and the ROM, you could see for yourself that the software is simply requesting the knock sum value from the ECU.
Its quite easy to prove this for yourself by watching the timing and the knock sum together. If what you propose were true, then the logging program could show knock counts without the ECU pulling any timing. This never happens for significant knock sums (4+). I have never seen an instance where EvoScan showed a knock sum of 2 (or more) where the ECU did not pull back timing.
And if you actually could read the programming languages for the software and the ROM, you could see for yourself that the software is simply requesting the knock sum value from the ECU.
#101
Account Disabled
iTrader: (465)
It is at a tuner's discretion as to what level is safe for the how the car is intended to be used.
One thing being overlooked is as of right now at this moment, the temperature, humidity and conditions will differ than any other time so a tuner may decide to leave some minute sums or make other decisions based on forward thinking data and not just what is being read from the ECU on that very moment in time.
Weather changes, AIT changes, the conditions the car is operated in may be different than the condition in which the car is tuned in.
All this needs to be taken into account when determining where to leave a map at the end of a tune.
One thing being overlooked is as of right now at this moment, the temperature, humidity and conditions will differ than any other time so a tuner may decide to leave some minute sums or make other decisions based on forward thinking data and not just what is being read from the ECU on that very moment in time.
Weather changes, AIT changes, the conditions the car is operated in may be different than the condition in which the car is tuned in.
All this needs to be taken into account when determining where to leave a map at the end of a tune.
#102
Decay interval knocksum > sample interval
In this case can you explain how knock sum data is ever missed?
"The OEM ROM is not transfereing memory dump to the datalog software."
Perhaps you can explain the following code then that shows that it is?
ROM:000217BA mov.l @(h'DC,pc), r0 ; [00021898] = REQ_MUT_0
ROM:000217BC mov.l @(r0,r1), r10
ROM:000217BE mov.b @r10, r8
(r0=MUT table base, r1=request ID offset, r8=byte to the serial port)
That MUT table contains the addresses of real RAM variables. The logging is easily fast enough. Some values are clipped or heavily quantized, but practically I don't see the limitations, and if I did I would be addressing them by rewriting them.
In this case can you explain how knock sum data is ever missed?
"The OEM ROM is not transfereing memory dump to the datalog software."
Perhaps you can explain the following code then that shows that it is?
ROM:000217BA mov.l @(h'DC,pc), r0 ; [00021898] = REQ_MUT_0
ROM:000217BC mov.l @(r0,r1), r10
ROM:000217BE mov.b @r10, r8
(r0=MUT table base, r1=request ID offset, r8=byte to the serial port)
That MUT table contains the addresses of real RAM variables. The logging is easily fast enough. Some values are clipped or heavily quantized, but practically I don't see the limitations, and if I did I would be addressing them by rewriting them.
Last edited by jcsbanks; Jun 28, 2007 at 11:17 AM.
#103
Evolved Member
iTrader: (33)
Join Date: Aug 2006
Location: St. Louis
Posts: 1,182
Likes: 0
Received 0 Likes
on
0 Posts
I was running 93 octane BP fuel and running 24 psi (stock tune)....had a few 3's and some 2's. Right after that we flashed in the new tune from TTP. Knock went down to a few 2's some 1's and the rest was 0's. Pulls were made in 3rd gear starting a 3K all the way to 7K. I think a few hits of 3's wont hurt anything.
#104
Account Disabled
iTrader: (465)
Decay interval knocksum > sample interval
In this case can you explain how knock sum data is ever missed?
"The OEM ROM is not transfereing memory dump to the datalog software."
Perhaps you can explain the following code then that shows that it is?
ROM:000217BA mov.l @(h'DC,pc), r0 ; [00021898] = REQ_MUT_0
ROM:000217BC mov.l @(r0,r1), r10
ROM:000217BE mov.b @r10, r8
(r0=MUT table base, r1=request ID offset, r8=byte to the serial port)
That MUT table contains the addresses of real RAM variables. The logging is easily fast enough. Some values are clipped or heavily quantized, but practically I don't see the limitations, and I did I would be addressing them by rewriting them.
In this case can you explain how knock sum data is ever missed?
"The OEM ROM is not transfereing memory dump to the datalog software."
Perhaps you can explain the following code then that shows that it is?
ROM:000217BA mov.l @(h'DC,pc), r0 ; [00021898] = REQ_MUT_0
ROM:000217BC mov.l @(r0,r1), r10
ROM:000217BE mov.b @r10, r8
(r0=MUT table base, r1=request ID offset, r8=byte to the serial port)
That MUT table contains the addresses of real RAM variables. The logging is easily fast enough. Some values are clipped or heavily quantized, but practically I don't see the limitations, and I did I would be addressing them by rewriting them.