RomRaider

I've been scratching around the H6 roms for the VVL thresholds used for the switch from low to high lift, and I believe I've found the area in the rom, and a whole range of related parameters, including RPM, ECT, oil temp, engine load and some I have not yet deciphered.

The discussion started here: viewtopic.php?f=8&t=8328 , but I would like to rather bring it into its own thread (will ask Dale to move those posts here).

I had a look at a 2.0 legacy NA rom as well, and have a fairly good idea where the parameters are in there.

Since I don't have an NA Subaru to test / log on, if anybody is keen on testing this, let me know.

-------------------------------- posts copied from other thread -------------

Been doing some research into this, looks like the switch over points are used to affect OSV duty cycle, initiating either low or high mode.

The VVL_Life_Mode parameter will read 1 at low mode, and 3 at high mode (and I've picked up a numerous routines where this is checked, and some variables that feed these numbers back to the VVL_Mode parameter).

~18% OSV duty cycle at idle, rising to 100% on initial 'high mode' choice, and then lowering to ~80%.

See this: https://dl.dropbox.com/u/3099895/VVL%20 ... P0631C.pdf

I think I may have found the area dealing with the conditions for high and low lift, fairly complex set of variables that go into it, including RPM, accel pedal, engine load.

Does a switch over point of 3800 ~ 4000 RPM sound too high?

From Jeff's post on this thread: http://forums.nasioc.com/forums/showthr ... 107&page=2 seems like I may have found the RPM thresholds.

Which interestingly, are set at 3800 and 4000 - the best point that the above thread found in terms of torque vs. power.

Correct. These are actually in the 'Rev Limit' of some ROM's and are heuristic in that they activate over 4000rpm and then turn off below 3800rpm(so once ON, keep it above 3801rpm and it stays on) which is why the smaller ignition timing tables cut off around 4000rpm.

There should also be a coolant temp activation as well as a Req Torque switch point as well.

ECT is in there as well, as well as vehicle speed.

Now I just need a guinea pig to test these out

Here is a log with lift mode and OCV logged (when I was looking at stuttering problems)

It seems to switch to "3" sometimes under lower rpm - 2500 ish - but normally only kicks in at 3800

That is incredibly useful...

Looking at the logs, I believe it confirms that I've found a whole set of parameters related to the VVL switching thresholds. When load, accelerator pedal are relatively stable, you will see that the switch from low to high occurs at 3800 ~ 4000. However, when load increases rapidly, it can happen earlier, or when the accelerator pedal is cracked open quickly, it can switch over earlier. To confirm this behaviour, there are a number of tables that relate the following variables:

RPM to accel pedal - at low rpms, 100% change in accelerator pedal position (I believe) will cause switch, tapering to 5% at 3600 rpms
Engine load delta (negative and positive) to a value as yet undetermined
Accel pedal angle (0 to 100) to a value as yet undetermined

There are also single, float parameters involving:

IAT
ECT
RPM
Vehicle Speed

And some others, as yet undetermined.

No problem - let me know if you want anymore data and I will try and get out and do some more logs

As I suspected from tuning them.

Nissan and Honda's VVL system has similar triggers

Well - don't know if you would be comfortable doing so (I'm sure Fused is, as he requested I look into this, but he's not be on lately), but would you be interested in testing some Alpha rom and logger definitions around this, which would help confirm this find?

I'm quite amped about this, it's a brand new set of variables related to an area that's been inaccessible before, but I gather can result in some interesting gains with respect to the NA engines - which have limited scope for tuning in the first place.

So what would be the approach, in general? The most obvious point I gather would be to extend or delay the initial switch over point to one that yields the best torque at the earliest RPM.

would be up for it - but need to sort out my avcs first (one bank lags the other/sometimes fails to advance fully) and I don't get a chance to work on or drive my car a great deal.

Increasing the torque at lower rpm would pretty much be the holy grail in terms of what I feel is missing from the h6's performance

Tune the car with VVL off completely to say 4800rpm
Then tune car with VVL ON from a lower point(3000rpm, too low and you pull oil away from crank bearings)
Map out where the torque overlaps and then activate around 200rpm earlier to allow the engine time to switch etc. Obviously AVCS affects VVL as well so it is an interesting 'mix' to achieve a powerful results

You just pointed to another RPM variable usage for me - there are 3 single float RPM parameters:

3800
4000

These are the switch to high / low mode triggers under steady state conditions, with hysteresis - switch on closer to 4000 rpm, and switch off at 3800
and

3000

Following the logic, below this point, VVL_mode is always set to 1. Your point above

Oil temperature also has thresholds.