Red intake ECUs and the new Primary Open Loop Fueling tables

[callisto]

In red intake vehicles, Subaru improved the OL/CL logic, and as part of that the Primary Open Loop Fueling tables that were shipped in black intakes (K00G, OFTv2/v4, Wayne, etc) have been redefined in red intakes (S10C, U01A, etc). This is my best effort at documenting what changed and how an experienced tuner could work with those tables. I'm not an experienced tuner, so corrections and clarifications are extremely welcome.

The main K00G table, "Primary Open Loop Fueling", is now a pair of tables in S10C, that I will be calling "(Base)" and "(Normal)" in v2 of my S10C XML. The ECU logic uses a "Timing Ratio" to interpolate between the two lookup results from the two fueling tables to get a Timing result. (The red intake logic also does this when calculating the Base Timing A/B results now, rather than the old single Base Timing A/B tables.)

For example, and these are completely made-up numbers, you're at Load 0.50 and RPM 3200 and your tables return Base 10.00 and Normal8.00 for that lookup, and you're operating at 95% Timing Ratio; then the result would be (5% of Base + 95% of Normal) = (Base + Ratio*(Normal-Base)) = 8.10 AFR. You can inspect the ratio directly with a logger; in S10C, it's a float at FFF8B1F8, and in K00G, at FFF8B1E8.

There are considerable differences between the Base and Normal tables in S10C; the stock Normal table goes leaner ~400 rpm earlier across most of the table past 2400 rpm, and compared to K00G has fully 14.7 AFR'd the entire block of values at load 0.85 and above for rpm 2200 and below, as well as dialed in more gradual shifts in some areas.

The secondary K00G table, "Primary Open Loop Fueling Additive", has also been split into "(Base)" and "(Normal)" tables, same as above. This is the same as above, except that the interpolation is done using the IAM ratio rather than the Timing Ratio; so, for example, if you're running California 91 gas at an IAM of ~0.94, then the final Additive result would be (6% of Base + 94% of Normal) = etc, same as above.

The additive tables in S10C have been revised and split from K00G as well; however, both Base and Normal have the same values. They've been split up so that you can adjust the fueling further based on IAM, but Subaru isn't taking advantage of that in the stock tune.

The K00G value, "Minimum Active Primary Open Loop Enrichment", has been increased from K00G: 14.00 to S10C: 14.41. I don't yet understand from a tuning perspective why they felt this was necessary, but duly noted.

The K00G table, "OL Fuelling Accelerator Compensation", has been replaced with a much more complex table in red intake ECU, "Primary Open Loop Fueling Minimum". In K00G, this table was just a simple "Accelerator angle" to "Extra AFR" table; in S10C, once the accelerator is at angle 0.85 or above, it looks up Load/RPM in the Minimum table and uses that result instead. All of the table's values in S10C stock are 0.11-0.16, but it's not a smooth gradient; less fueling in the dip, more around it, etc.

The K00G tables "Front Oxygen Sensor Scaling" have been revised, both in axis and values, from the narrowband definition in K00G to the wideband definition in S10C. For tuning a red intake, no need to rescale it; for tuning a black intake, here's a free OEM rescale courtesy of Subaru.


So, that's what I've worked out so far about Primary Open Loop Fueling. Hope it's of use to anyone.


Bonus related CL/OL content that doesn't have its own post:


Nearly all of the K-series CL/OL tables are gone. The code isn't even remotely the same anymore, and a lot of the undocumented constants supporting the few remaining tables are different behind the scenes. Consider very carefully whether to use any CL/OL values from a red intake ECU in a black intake tune; however, here are some key points of interest to be aware of, whether tuning red intake or black intake:

CL to OL Transition with Delay (BPW) has increased from 36ms to 100ms.

CL to OL Transition with Delay (Injector) has been heavily revised, including both the RPM values and the injector durations.

CL Fueling Target Compensation B (Load) has been heavily revised; A remains unchanged, and in my S10C xml, C is just a copy of A.

CL Fueling Target Compensation A/B (ECT) have been heavily revised; everything -10C and above is zeroed out now.

end.

[tomm.brz]

thanks for your findings

Anyway you could also look at Q00G and S20G tunes, they are in blacl manifold cars but have inherited the closed loop strategy of red intake cars

S20g is similar to U001A, maybe a bit richer and less timing because in europe fuel is much shittier

fun thing is every restyling OFT tune is based on K00G, as it is more predictable and simple and defined and very stable
In fact usually when retuning red intake ecu, tuners tend to make identical the various primary fueling table, as to have consistency

[callisto]

Originally Posted by tomm.brz

you could also look at Q00G and S20G tunes, they are in black manifold cars but have inherited the closed loop strategy of red intake cars

S20g is similar to U01A, maybe a bit richer and less timing because in europe fuel is much shittier

Thanks! I have Q00G on my list and I've started into S20G. I noticed the port injection ratio table in S20G is considerably more complicated than any other, as far as I've seen so far. I didn't realize they were compatible with black intakes, so that's super interesting to consider.

Is European 95 fuel about the same as California 91, or is it better/worse?

Originally Posted by tomm.brz

fun thing is every restyling OFT tune is based on K00G, as it is more predictable and simple and defined and very stable
In fact usually when retuning red intake ecu, tuners tend to make identical the various primary fueling table, as to have consistency

Yeah, all the OTS tunes derive from K00G. I'm not quite tuning, though; I'm on a quest to understand how the stock ROMs behave, especially for red intakes. My current project is working on datalogging the entire long-term fuel trim table (all 35 entries!) so that it's possible to see unevenness (and to verify that learning has stabilized).

[tomm.brz]

Actually also K00C could be force flashed to black intake restyling cars without problems

yes 95 ron is similar to us 91, even slightly worse