|
|
|
|
|
|
BMW Garage | BMW Meets | Register | Today's Posts | Search |
|
BMW 3-Series (E90 E92) Forum
>
Monitoring ignition advance for all cylinders on piggybacks
|
|
08-03-2011, 03:46 PM | #90 |
Lieutenant Colonel
65
Rep 1,708
Posts |
The last graph, timing event happened right when you stabbed the throttle... cool.
For the procede operation, if we have to use 1 cylinder for reference... how about giving us the option to choose the cylinder that we log and autotune. Of course we need the ability to log all cylinders at once to diagnose which is the weakest cylinder. Shiv, would monitoring all cylinders be a processing issue? Dzenno, do you see a trend with certain cylinder corrections more prevalent... or fairly even? |
Appreciate
0
|
08-03-2011, 03:49 PM | #91 | |
Banned
273
Rep 5,876
Posts |
Quote:
|
|
Appreciate
0
|
08-03-2011, 04:02 PM | #92 | |
Major General
178
Rep 6,631
Posts |
Quote:
__________________
E90 LCI N54 6AT
|
|
Appreciate
0
|
08-03-2011, 04:18 PM | #93 |
Banned
273
Rep 5,876
Posts |
Applying ignition "retard" globally through CPS is fine and does work as designed/intended as we know and have seen with the Procede and on the JB4/CPS module...the issue is when you base other algorithms and allude to adding ignition advance using CPS offsetting based solely on cyl 1's timing that freaks me out...especially if a piggy takes into account only ignition timing dropouts and not the ideal timing curve for given CPS offset...AFAIK this isn't the case with the procede anymore with their autotuning but I haven't had a chance to see it/log it so no idea..
|
Appreciate
0
|
08-03-2011, 05:03 PM | #94 |
Lieutenant Colonel
65
Rep 1,708
Posts |
AutoTuning
My understanding of autotuning looking at the opened maps and speaking with Shiv long time ago is that it targets a certain ignition advance at rpm/boost. So:
- There's an advance target that's related to aggression (actual I assume) - IC is adjusted based on DME adv target and potentially knock events Let's say you have X as the low side IC... maybe 30% and Y as the high side... maybe 70% When you reach X, then boost will start to increase... once you reach Y, then boost will decrease... and on you go. Much of above is conjecture on my part. This is without DIC... would love to know more about Shiv's DIC. |
Appreciate
0
|
08-03-2011, 06:13 PM | #95 | |
Mad Linux Guru On The Loose
1130
Rep 5,400
Posts
Drives: 2008 335i Sedan, 2023 M3
Join Date: Aug 2009
Location: Chicago, IL
|
Quote:
Individual ignition correction has been done for several reasons: fuel efficiency, power and primarily for safety. Engine manufacturers long ago have known that there could be a significant variation in the operating environment between cylinders, especially in longitudinally mounted engines where the front of the engine sees cooler temperatures than the rear of the engine...
__________________
|
|
Appreciate
0
|
08-03-2011, 06:49 PM | #96 | |
Major General
178
Rep 6,631
Posts |
Quote:
__________________
E90 LCI N54 6AT
|
|
Appreciate
0
|
08-03-2011, 06:49 PM | #97 | |
Mad Linux Guru On The Loose
1130
Rep 5,400
Posts
Drives: 2008 335i Sedan, 2023 M3
Join Date: Aug 2009
Location: Chicago, IL
|
I was just reading up on Controller Area Networks (CAN) on wikipedia and I wanted to throw this into the mix:
Quote:
So, using the CAN bus to obtain information of how the engine is running is not optimal simply because you are limited by the amount of data the CAN bus is capable of supporting AND the speed at which controllers are dumping data into it. All sensors (correct me if I am wrong) are analog. So they provide a constant signal feed to their host controller on the CAN bus. However, the controller itself has to digitize that signal and there is inevitably going to be loss of data. Theoretically speaking... Futhermore, a controller may not have the ability to broadcast constantly on the CAN bus all the time for various reasons: reduction of traffic collision on the bus and simply lack of processing power. That being said, I certainly hope that the DME dumps enough data, at high enough frequency on the CAN bus for anyone to actually run a real-time tuning solution. It LOOKS like CAN bus data is good for logging, troubleshooting and generally learning how things are working. I won't be surprised if the DME actually talks to sensors and makes adjustments much faster than it actually communicates on the CAN bus. ...just food for thought...
__________________
|
|
Appreciate
0
|
08-03-2011, 07:18 PM | #98 |
Captain
15
Rep 603
Posts |
This may be a stupid question....but what feedback is the DME using to alter the timing of one of the cylinders over the other. There isn't an individual knock sensor for each cylinder, correct? So if there is one knock sensor each cylinder would have to knock at a different and distinguishable frequency.
|
Appreciate
0
|
08-03-2011, 07:26 PM | #99 |
Mad Linux Guru On The Loose
1130
Rep 5,400
Posts
Drives: 2008 335i Sedan, 2023 M3
Join Date: Aug 2009
Location: Chicago, IL
|
No, there is not a dedicated knock sensor for each cylinder. Even if there was, there is still no need for different distinguishable frequency for each. The DME will *see* knock events occurring on different pins on its harness. After that it will know where it is occurring.
__________________
|
Appreciate
0
|
08-03-2011, 07:30 PM | #100 | |
Captain
15
Rep 603
Posts |
Quote:
|
|
Appreciate
0
|
09-02-2011, 12:30 AM | #103 |
First Lieutenant
8
Rep 395
Posts |
That is really interesting, as we all try to fine tune basing our parameters on Cyl 1 while in fact there might be even worse numbers on the other cylinders.
What i dont get is, ur FBO+RBs which means you would need much less boost to reach max load whilst in your log you are holding an average of 14 psi tapering off at redline + the car is pulling timing. By the looks of it timing pulls or retard is a normal thing with these cars so might as well run the piggybacks and knock it to death..lol Cobb lowers timing but i dont think lowering even more is wise, too much boost and no advance = not much power and pointless if you ask me. Anyway will keep reading and learning.. Thanks |
Appreciate
0
|
Bookmarks |
|
|