Bryant89

Hey guys, I was running a set of unknown brand/size injectors and I swapped them out for a set of FIC 950.

I'm now trying to scale the injectors properly using Merlin's Injector Setup Method #2.

I've logged LTFTs at idle and cruise and these are my best values:

Scaling: 713 (previous injectors tuned by tuner at 850 scaling)
Latencies:
4.69v 3.312
7.98v 2.184
10.03v 1.464
12.01v 1.128
13.99v 0.960
18.60v 0.408

Clarification 1:
The voltages in FIC's datasheet do not match Mtisu's stock voltage figures, but I used FIC's voltages (8, 10, 12, 14v) instead. Is this correct? I had adjusted the latencies in % until I achieved good idle LTFTs.
I also note that my scaling of 713 is much lower than the generally accepted number of 812 (should I be concerned?)

For the next step, I changed my fuel map AFRs to 11.5 for cell load 120% and above for all RPMs.

Merlin's guide says:
Run the car into some low level of boost ie above load 120. Keep the boost levels to less than 12 psi preferably. Have EvoScan or Mitsulogger logging Airflow Hz. Hint, turn off everything else except rpm and load. Compare and record the air-fuel ratio reading to the set 11.5 value. If the error is less 0.2 then maybe you wouldn't bother with any further adjustments and just note the offset and allow for it. If more than 0.2 and you want to dial it out, proceed on.

Clarification 2:
Does this mean my AFR should remain at 11.5 throughout the RPM range?

When I attempt to WOT, I do achieve 11.5 AFR from 3k to 5k RPM but the AFR begins to taper off to about 12 from 5k RPM onwards. Am I doing this correctly? Or should I look into MAF scaling next?

Thanks in advance for the advice!

OzAirtrek

I don't have the answers but very keen to also know. I'm planning to use this same method to scale my 650cc. I also want to know if it is standard to rescale the voltages.

merlin.oz

Dialling in unknown injectors can be a pain but the basic procedure is fairly straight forward.
An unfortunate complication is that the injector scaling definition calculation that you will typically find in EcuFlash definitions is incorrect.
On this forum you will see injector scaling calculation of :

toexpr="29241/x" frexpr="29241/x"

which result in an injector size less than the real injector size.
The correct calculation is as follows:

toexpr="34140/x" frexpr="34140/x"

This scaling equation is derived from Mitsubishis injector sizing and is correct.
Armed with this you can set the injector size to what you think it actually is and proceed from there.

To tune fuel correctly your definition need to show the table I call MAF AFR MATCHING.
This is the second of the two 21 element MAF tables and it is a MAF fuelling correction using a 128 offset.
So a table value of hex 80 or 128 in units results in zero correction. Values larger than 128 add fuel, smaller values reduce fuel.
I have this table setup to read correction in percentage which makes fine tuning of the MAF correction very straightforward.
I generally set this MAF correction table to a flat 100% ie no corrections when starting with unknown injectors and when the intake setup has been altered in any way from stock.
Which pretty much means I would normally zero the correction and start from scratch.

So, to the tuning, set the MAF AFR MATCHING to 100%, load the expected injector size, load the injector latency with your best guess or factory data if available and load the fuel map with a flat 11.3 (gasoline tuning) AFR and start logging. You might as well ignore fuel trims and latency at this point unless latency is so far out the car wont achieve closed loop operation at normal operating temperatures.

Drive the car into boost, say 5-10 psi will do. Observe or log actual AFR. Best procedure is to log the widEband through the ECU so you can have together on the same log
FUEL MAP AFR
WIDEBAND AFR
AIRFLOW HZ

Now note the AFR percentage error in the log between actual and mapped AFR at specific airflowHz.
If the error is much more than 7%, change the injector size so you get a better match.
Raising the injector size results in less fuel an a leaner AFR. Reducing the injector size gives a richer AFR.
Repeat this step until you get the best AFR match to the fuel map AFR.

When you are satisfied you have the best injector size in the tune, now repeat this step and start fine tuning the MAF AFR MATCHING table.
Typically this table will have some addition and subtraction.
When you have low boost tuned reasonably well it is time to get serious with fuel trims and latency.

Most evos typically will want the MAF correction to have some subtraction below 100Hz, maybe up to 10%.
This is where having the true latency data is really useful as for example Injector Dynamics data can be relied on.

Short term fuel trim tuning is done with latency as one input and MAF HZ correction below 100Hz as the other.
If I know what the latency should be then I use that and tune the MAFHZ correction until I get a good low tune trim, something less than 3%.
So we are looking to get Long Term Fuel Trim - LOW to about +-3% or better.
When you have achieved this you can move on to LTFT-MID and LTFT-HIGH.
I will typically set the mid to high airflow Hz to 185Hz.
This means LTFT-MID is operating from 75Hz to 185Hz and LTFT-HI is operating above 185Hz until the ECU drops out of closed loop.
Log the car while driving in the LTFT-MID airflowHz window. This will be something like 60-70 kph in 4th gear on flat level roads with minimal interruptions.
It will take some minutes in the window to get the trim to start reacting.
Likewise for LTFT-HI you will need to drive at 14.7AFR above 185Hz, 80-90kph typically.
When you have useful trim data in the logs, adjust the MAF AFR MATCHING table.
For example, if LTFT-MID was adding 5% fuel to achieve 14.7AFR then raise the MAF correction by 5% from 100-180Hz
Fine tune the LTFT-HI section, noting from the logs when the ECU drops out of closed loop.
This might be as high as 300Hz on a well tuned setup. My evo 9 happily maintains 14.7 AFR to 1.5 psi of boost.
Exactly where it will cease closed loop operation is dependent on the closed loop load tables.

When you have got all this running nice tight fuel trims the car should be running fairly sweetly you can finish off tuning the high boost area.

OzAirtrek

Wow thanks for the detailed post Merlin!

I just put in some 650cc injectors and am still getting the scaling as close as possible.

I had not heard about altering the maf scaling like that. In my log i was noticing the fluctuation in AFR above 120 load (minor). I will move onto the maf correction next!

AlexEvo26

Thx for this very clear inputs, even after previous tuning exprience this is very, very useful.

I would also and this post :
https://www.evolutionm.net/forums/ec...ion-table.html

maybe you are not very concerned with 950cc inj, but I suggest to have a look into this and maybe get FIC info.
you should start by setting this first before applying the method above


@ Merlin : I understand you suggest to modify the « injector scaling » expr in the XML definition file. This would allow to use a scaling value closer to the « real » value (eg 560 vas 513 for the stock injectors).