Also compare these two load-based timing maps. The top one is the off-the-shelf stage 1 timing map for a 2005 Subaru Legacy 2.5 GT (turbo). The bottom is the stage 2.



You can see as you look vertically down the load column that the stage 2 is "flatter." Remember that the basic load calculation on a Subaru is

MAF grams/sec (taken from the MAF voltage and the airflow lookup table) * 60 seconds / rpm = grams/rev

so at 5000rpm if I have say 3.8 MAF volts, then look that up to be 200 grams/sec, I would calculate my load to be:

200 grams/sec * 60 sec = 12000 grams (seconds cancel out)
12000 grams / 5000 rpm = 2.4 grams/rev

The base timing is going to be interpolated presumably between the closest four-cell matrix. Then the Subaru ECU adds a long term octane learning adjustment on top of that ("dynamic advance"), plus any immediate response to knock sensor noise. There are also some other obscure corrections that may be added in. Other manufacturers with MAF systems have a similar overall process to determine load (measure airflow, use a bunch of lookup tables and constants, divide some value by rpm) but the actual numeric calculation can vary widely. Evos and Rx-8's for example are a lot more involved in the way they numerically calculate load.

And FYI here's another factory timing map (boost-based) off a turbo piston engine. This is from the 4.3L GMC Syclone:



Y axis is rpm, X axis is manifold absolute pressure where any value over 100kPA is boost. 2.11 degrees are added on top of these values when the vehicle enters power enrichment mode (open loop, WOT).

 

man that's way more timing than we run NA....

 

yes i do notice that you have retarted the timing a little bit more on the S4 N/A rotors with respect to the T II rotors, arghx i'll be using your timing table but do you think that it would be a good idea to start off with this table and look at the values of the Knock sensor on the Rtek and if the engine is not knocking then start advancing the timing little by little until it starts knocking a little and then retard it by 2 degrees until the knock is gone? also should i only change the values on the leading and leave alone the split table you also gave me?

 

^ I think you'll have to get input from someone who has more experience with interpreting the knock reading on the rtek.

On a dizzy setup there's only so much advance you can really run because of the limitations of the design (centifugal advance relative to the base timing). The Rx-8's (10:1 compression) main timing is this:





WOT you are going to be around the 1.00 load column

 

true.

we used to run an FC in ITS/PS1, and with 9.7 rotors and a haltech,
it made best power @26 BTDC L.

which is similar enough to the rx8