So as far as I understand it (and the search button has shown me) - With larger injectors and a higher-flow pump I need to input correct latency values and then a flow scaling number that the car likes.

So I put in the Latency values supplied by Deatschwerks (the 650's do vary from the stockers in the 06 STi), and then tried various Injector Flow Scaling values untill it idled nicely and wasn't spitting gas out of the muffler. The AFR at idle was hovering right around 14.7 with the AF Correction #1 in low percentages (3-5%).

While it seemed okay like this AF Correction %3 was something like 50 Percent ! when I'd get on the gas lightly to accelerate (staying out of boost) it would go to 39-40ish but I stayed out of boost to be safe.

My questions to the guru's :
Question 1 : What is the relation between AF Correction 1 (which is always reccomended to log) and AF correction 3 (just randomly selected and it showed a high value spooking me.)

Question 2 : MAF scaling - would it be possible to input the correct latency values, enter 650 as an Injector Flow Scaling value, and then just tweak the MAF values to work with the new injectors ( this had been suggested as an alternative). If I go the MAF route, what side effects might I experience if any ?

Altering the enitre MAF table to correct for fuel injector change isn't the best way to do it. Engine load is calculated off of the MAF input and may be skewed if you get too crazy. You should be able to get close with just the fuel injector trim and the latentcey values.

You will find that things may look normal at idle with a set of latencey and scale numbers and then go wacky at other load/rpms. If so then take from one and correct with the other. Get it as close as possible. Then you can fine tune a couple of cells in the MAF table.

So if I set the injector value to 650cc with correct latency and then just re-scale the MAF, then that could affect the car's feeling - even if it is not knocking and maintains steady AFRs - it may feel different in areas where it is incorrectly calculating engine load ?

I imagine this would have an amplified effect during quick MAF voltage changes - like when you floor it from steady state cruising ? Could that be a fair assessment ?

With my Deatchworks 850cc injectors i had to run a good deal more latency than i calculated from their information provided.

If i set the injector scaling so that it would idle it would be super rich at WOT. So i took the advice i had seen earlier to use an injector setting according to what the injector is and then adjust the idling fueling by modding the latency. Working much better so far.

Be careful with latency data. The enginuity defs were changed.

I have DW 650 side feed injectors in my LGT right now. Post up your latency values and I'll post mine up later tonight when I get access to enginuity. Then we can compare.

My AFR Learning B, C, and D are very stable and vary from +1 to -1 so I think it works. Also, I didn't touch my MAF scale when I installed the injectors.

Here they are -
6V - 3.00
9V - 1.79
12V - 1.37
16V - 1.00

What did you use for an Injector Flow Scaling Value ? was it 650 ?
When I tried 650 it was spewing clouds of unburnt fuel from the muffler !

What are the different AFR learning values - A, B, C, D - how do they differ ?

My injectors were used. I had them cleaned and balanced at DW. Unfortunately, DW only balances within 3%. If you buy them brand new, they pick 4 that are within 0.5% of each other.

Anyway, 2 of mine were around 660 and 2 were around 640 as reported to me by DW.

I started with 640 but then after driving around for a while, I calculated that the correct value was closer to 650 based upon my AFR corrections at high speed cruise.

Are you using the December 2007 enginuity defs?

AFR Learning A, B, C, and D refer to the 4 AFR learning values for the different MAF ranges in closed loop. Value A is basically idle. Value D is 40+ g/s. I can't remember B and C but if you open up enginuity, you'll see it.

No - new Defs for DEC 2007 ? I guess I need them !
I'll DL them and put them in place of my old ones tonight.
Will it make a huge difference ( i noticed the bold :) ) ?

The injectors were new and very close to each other...

Thanks for the help everyone !

Yes, it is very important because I am using the new defs.

In order to compare latency values, we need to use the same defs.

Cool. Done and done. I'll put the new defs in tonight.

Here are the values that I am using.

6V 5.16ms
9V 1.98ms
11.5V 1.21ms
14.0V 0.88ms
16.0V 0.74ms

The latency for the 14.0V was calculated based upon my logs. All other latency values are guesses.

My flow scaling is 648 cc/min.

The above are based upon 2005 LGT enginuity defs from December 2007.