DMT,

that is amazing, I would have thought Mitsu would have been able to refine that intake better than that. You don't think it could have been an injector with a hair? Do you remember what rpm the difference became an issue?
If I understand you, you were able to balance the exhaust temperature with a senor in every port.

OK, I will agree a fueling error that big can mean power. Generally though you can run half a point in afr with not too much change in power. I'm repeating what big name commercial tuners say.

 

thats good info DTM, although I wonder how much injector differences come into play here..

presumably if you get 4 injectors that spray exactly the same then AFR's should be equal in the cyl?

 

Tephra I believe its more of the intake manifold design than a difference in the injectors. #2 is the leanest cylinder as that is what most people run there EGT's and the cylinder that most have damage on in a failure.

 

Without getting into too much, the AFR differences were greater in low load and low rpm. At one point the AFR had a differential of .8 to 1.7. Cylinder number three consistently became the culprit with the highest cylinder temperature.
We probed all four runners 1 1/2 inches from the port on the bottom side of the runner with EGT and 2 1/2 inches away were the individual WB outputs.
After adjusting individual cylinder fuel percentages in a linear fashion, we were able to come within .2 of a point on average during full throttle pulls. Under specified load (mustang dyno constant speed) the AFR became a bit more erratic between cylinders as we encountered slight rich misfires. Once we upgraded to an HKS DLI, even under constant load the AFR remained pretty consistent. Averaging now in the .3 range. At this point, please don't take it the wrong way, I will only give certain information about our findings. 1/2 a point of AFR differential betwwen cylinders can be ok, but you are talking about a forced induction (4cyl) application that generates a lot of heat and friction. Ask almost anyone here on EVOM to do a compression test on their STOCK blocks. You will find that 9.5 out of ten people have a 20psi drop in cylinder number three. It is inherent to the design of the engine and its associated components. We merely wanted to find out what it is actually processing during its cycle.

 

IMHO I do not think that the injectors played that much of a role in the difference between cylinders. Injector spray patterns and latency account for about 1/2 % + or - of the margin. VE accounts for the rest.

 

The order in which we concluded was the leanest to the richest:

2
4
1
3

So your post is extremely accurate.

 

thank you for giving up your hard earn observation.

am I misunderstanding?

I was thinking you were saying 3 was lean. That just goes to show how little I understand egt readings.

Once again, thank you.

 

so how much power are we talking about gaining if we get individual cylinder trims working on the stock ecu?

 

Max 9-11 HP as a peak comparison point. But it is more than just power that can be gained. Reliability, torque down low and the midrange, along with longevity, better average MPG and better drivability. Look at it like this, the engine is running for the most part with drag when one or two cylinders do not produce the same amount of twisting force on the crank per revolution as the others. Its like having the parking brake on while driving. You notice the gains right away in sudden throttle exchanges. Almost like how a stroker in an evo surprises you when you first drive it. The torque is very noticable.

 

Sounds like it might be a good idea to get injectors flow tested and blueprinted with a custom offset for each cylinder, especially #2. At a minimum, if you flow tested them you could put the highest flowing one on #2. This would be a mechanical solution to not having individual cylinder trims in the stock ECU, though to the extent that the differences between cylinders change with RPM/load, I guess you'd have to compromise and choose your priorities.

Also, the cylinder-to-cylinder variances potentially suggest a benefit for port injection on an alky kit. You could customize the nozzle sizes for each port to the inherent VE differences of the cylinders to level out the AFRs under high load (spray on) conditions.

In fact, what you could do is get the injectors blueprinted for low load/low RPM (spray off) situations and then use the port injection nozzle differences to even things out in high load/high RPM (spray on) situations.

Then again, a lot of trouble for very little gain...

 

so really the individual differences would probably be the same across all evo's.
so if we could setup something like another table that told the ECU how much extra or less to inject for a particular injector then we could iron out this problem a bit- ie:

1 -5%
2 +10%
3 -10%
4 -5%

 

that might not be a bad idea, however it was mentioned that the greatest difference was at relatively low load. That would suggest that at high load the manifold becomes less of a fault.
That would explain why spark plugs don't show the same results as DTM expressed.

DTM would know better than I, but I'd be inclined to make smaller adjustments.

 

Ideally a fuel modifier with 1000 rpm increments along with 10 load sites would do the trick. Just having a static modifier works too, just like Riches post of a mechanic fix via injector flow and positioning.
The stock intake manifold, is just a plain old restriction at a point under high load and rpm. That would explain the variance down low. We haven't tested , but could theorize that an aftermarket intake manifold would make the results linear and worse for the leaner cylinders. Just a thought, don't hold me too it.

 

OMG: I came back to visit this old thread and it became my dream! Tephra asked the winning question(my thought but not posted question of what would be gained on stock ecu with proper fuel trims on individual cylinders etc) I'm almost getting teary eyed... I'm just trying so hard to suck down so much info down at once at this point... I have one question though, if you got this trim down and added exhaust parts, would you notice any gains would the levels in trim change? You said "a fuel modifier with 1000 rpm increments along with 10 load sites would do the trick." What exactly did you mean?(what do you reccomend using?)