Is comprehensive aftermarket tuning for Corvette C7 Stingray really possible? The answer is... NOOOO
Why? The cause: direct injection and variable valve timing.
There is a reason why +General Motors needed to run TEN MILLIONS HOURS of computer analysis to get the factory tuning within the ballpark, allowing a direct injection fuel system to provide both emission compliance and durability. Why? KRIGING is the answer. What is Kriging? Kriging is another name for Gaussian process regression. This rather complex method of statistical computations is necessary due to the complexity of direct injection process. Unlike the conventional multiport fuel injection, requiring simpler polynomial interpolation, direct injection requires SPATIAL interpolation and super high computational ability of supercomputers, something that is available to a large manufacturer like General Motors but is only a pipe dream to aftermarket tuners.
Now, anyone with even some intelligence would be wondering all along how is this possible that while GM needed ten millions of processing time, aftermarket tuners can improve the factory tuning in a small fraction of time? Well, the truth is they cannot even come close, any of the currently available so called "tunes" for LT1 amount to hack jobs, addressing only a small portion of the overall tuning algorithm and likely endangering the longevity of the directly injected LT1.
What is the saving grace for these aftermarket tuners? Ignorance of the typical Corvette crowd is the key to keep the aftermarket tuners in business. These people have absolutely no clue to what is required to properly tune a directly injected engine. Instead, they rely on the reputation of the tuners that was developed during earlier days and convince themselves the reputation is equivalent to unquestioned loyalty. As soon as these engines start blowing holes in the pistons due to detonation due to overly lean condition, they may learn.
No doubt GM will be happy to accommodate their engine woes and replace trashed engines under factory warranty, haha...
It has nothing to do with direct injection and everything to do with an update to how volumetric efficiency, aircharge, and torque are calculated now. The multi-dimensional arrays are part of the basic air charge calculation, no necessarily the injector firing control. ...but this is still THE most important part of the "tuning" (really, calibration) process for modified engines.
ReplyDeleteOh really? Knock retardation is not an issue with DI tuning? One would think that lean condition would be a real concern and require quite a bit of work to recalibrate the torque tables but apparently for you, it is not an issue, LOL...
ReplyDeletePlenty of experience on various DI engines here (from various OEMs), so I'm pretty familiar with the physics and combustion issues at play. What do you consider "lean"? Have you seen the hook test across all delivered lambdas at the speeds/loads/cam angles in question? You might be shocked to see what AFR some of the other competitors DI engines run at, even under boosted conditions, without knock.
ReplyDeleteThe complexity of the aircharge model is a completely separate issue from knock control and delivery of the commanded fuel mass (by way of DI pulsewidth and event timing coordination). I'm agreeing with you that many "tooners" out there do not grasp the complexity involved with properly calibrating the new polynomial air path and torque models. But try to stay on task if you want to dig deeper.
Some people calibrate direct injection engines for a living, and some people write blogs I guess.
Lean like in constantly tripping knock sensor, clear enough?
ReplyDeleteLooking at the latest version of HPT, it seems that there were quite a few torque coefficients, most likely GM is doing virtual torque tables for these like it was done the VVE.
From what I see, there really isn't a way to calibrate the requested E92 torque tables correctly and there is a lot of problems with knock sensors and considerably more so with C7 than with 6.2 truck counterpart.
Interestingly enough, HPT does not have a help file for E92 like they have for Ford and Dodge. So yes, any tuning attempts right now are a real crapshoot, with tools willing to risk their new motors while so called tuners have no clue what they are doing. Now, do you actually have anything specifically related to this subject or just like to share your general experience? Blogs do not blow up motors, unlike incompetent tuners.
Lean is a fueling issue, knock is a spark issue.
ReplyDeleteThese are and should be two separate discussions. You can run rich and still get knock if the commanded spark is wrong. There are also tables that adjust the delivered spark as a direct reaction to the delivered AFR. Many "tooners" miss the fact that these exist, or worse yet, zero them out thinking that they don't matter. Bang...
Using HPT to view the calibration only shows the tip of the iceberg. HPT might have 300 variables available to see/adjust. In a modern DI/VVT control system, there are over 10,000 variables. It's very likely that some of what one doesn't see in an aftermarket tool has a big impact on how the engine will run.
LOL, great job playing the semantics game. Yes, of course, spark is the direct cause of detonation, however, lean mixture is always much more susceptible to being ignited causing detonation than rich AFR under WOT unless you know something that I do not know. Once again, GM spent a lot of time optimizing their torque tables to maintain a very fine balance, it should be pretty obvious that even though the factory calibration resembles the previous LS generation, it includes considerably less leeway, allowing for small and very localized adjustments and with the strong possibility that any gains within one region result in torque loss in another.
ReplyDeleteOverall, it seems that in spite of your nitpicking, you appear to agree with the point made here.