I think I remember seeing somewhere ID posted a photo of 4 injectors matched against eachother.. Couldn't we just use the largest variance as the value to input into the table?

Excuse my ignorance, and thank you guys for taking the time to explain!

 

Jeff, add this to you definition and then try lowering the Closed loop Idle Integral Gain value to help reduce the random swing.

for 9653:

<scaling name="I-Gain" units="%" toexpr="x/327.73" frexpr="x*327.73" format="%.3f" min="0" max="0.78" inc="0.00305" storagetype="uint8" endian="big"/>



<table name="Closed-Loop Integral Gain Settings" address="36B8" category="Closed-Loop Control" type="2D" scaling="I-Gain">
<table name="Conditions" type="Static Y Axis" elements="4">
<data>Closed-Loop I-Gain After Fuel-Cut, Decrease Fuel IPW (Lean)</data>
<data>Closed-Loop I-Gain After Fuel-Cut, Increase Fuel IPW (Rich)</data>
<data>Closed-Loop I-Gain Idle, Decrease Fuel IPW (Lean)</data>
<data>Closed-Loop I-Gain Idle, Increase Fuel IPW (Rich)</data>
</table>
</table>

Edit:
9653 has quite high stock gain values here:

0.051 to lean the mix
0.066 to enrich the mix

try something like:

0.021 to go lean
and
0.055 to go rich. Maybe a little lower.

 

Further comment on the latency change requirement:

I applied revised correction values for FIC1100 to my evo9 the other day and yes I did have to bump the latency UP two steps to get it all back square again.
To clarify this aspect of the tune, the IPW Small Pulse table had stock injector values, so at about 0.6mS was adding 48uS.
My idle IPW was about 1.7mS, with a 14V latency of 1.104mS, ie 1.7mS Idle IPW minus 1.1mS Latency = 0.6mS.

But the FIC1100 Small Pulse Linearization had zero compensation requirements at 0.6mS, and so when that value was plugged into the table it left an idle IPW shortfall, which needs to be made up by a latency increase.
A two step increase in latency (when using 24uS Latency Base) is 48uS.

This does make me think all the more that we would do better if we changed the Latency Base to 15uS (for a 15uS step increase) as per EvoX and most atmo Mitsubishis that use Hi-Z injectors.
I think that will be my next step.


My "tuned" latencies now are so very close to FICs published data sheet, whereas they were about 60-100uS less before applying the Small Pulse Linearization data (from mrfred).

 

Excellent!! Will try this soon as I'm off work.

I haven't had to touch my latencies yet, mine are matched to ID's data... so perhaps I was masking that issue elsewhere.

I made a lot of changes at the same time, I've been fighting my car never being able to start in the cold haha so I've made some major changes elsewhere and will revist the latencies.

 

Would FIC 1450's benefit from this? They always seem to be a bit finicky for me.

 

Here is the idle with the updated values.. Doesn't appear to swing as much however it tends to stay more on the rich side. Will play more tomorrow.

 

I am sure the FIC1450s would benefit, but getting the data can be difficult - for most of us anyway.

 

Merlin, what address would we use for the 9417 rom?

 

for 9417: 36B8

 

I'll see if I can get the data from FIC. They're probably busy as hell though.

 

+1 Thanks guys!!!

 

for 88580714 the same as 88590015 ROM

 

Hi..

Something doesn't seem to be right with the code information for the the table as posted in the first thread.

This is how I see it in my browser, doesn't look anything like the table defenition I have seen posted on the forums before.

 

Can somebody please post the table def for Evo 9 88590015 ROM? I think the forum bug made it go away.

 

"quote" the post that has the definition that you want. it will appear correctly in the quote.