Sale ends todayGet 30% off any course (excluding packages)

Ends in --- --- ---

Another injector deadtime/scaling thread

General Tuning Discussion

Forum Posts



Tech Articles

Discuss all things tuning in this section. News, products, problems and results. 

= Resolved threads


I'm absolutely racking my brain trying to figure out injector characterization. I'm using a Cobb Accessport with Access Tuner Race to tune my 2004 Subaru STi. First off, I attempted to scale these 1400cc EV14's with a starting base fuel pressure of 55psi, which might be causing my problems. I have my characterization parameters from my injector manufacturer. Scaling with the Cobb AP using a 4 digit scalar number. Am I to adjust my latency to hit a desired AFR at idle or do I scale the 4 digit number? Adjusting the scaler allows me to hit my desired AFR at idle, but Im wondering if I'm going about this the wrong way. I was told to enter my latencies from the manufacturer and never touch this but to adjust the scaler for idle afr. Im not using factory O2 feedback, so I'm relying solely on my WB readings. I suppose I'm confused on which injector table to actually mess with. I did some reading that told me the latency is responsible for idle afr and the scaler for cruising. Help!

Scaling aftermarket injectors for an OE ECU requires you to adjust both the scaling and latency values. Here is an earlier thread that explains the basis behind scaling injectors:


Thanks Andre. I followed cobb's method and just have had some troubles. The tune is in forced open loop for the time being. Here's my question though. I can adjust my scaler and achieve a desired AFR at idle but i can also do the same from adjusting the latency. Which way is correct?! With a base fuel pressure of 55psi, does my latency need to be adjusted around this? I'm just so damn confused. I'm not using fuel trims to make calculations for fueling, since I'm forcing open loop.

Latency has a much larger effect at very small pulse widths as the latency ends up being a much larger percentage of the pulse width when the pulse width is short. What this means is that at WOT or even cruise, the effect of the latency on the AFR is limited. At idle however the effect of adjusting the latency is much larger.

The way I scale the injectors is start with the scaler and get that to the point where the closed loop trims under a normal 3000 rpm cruise are close to zero. Once this is achieved you will finds the closed loop trims at idle are probably still quite a way out. You can then adjust the latency to correct this. Of course once you adjust the latency you will need to go back and check the cruise trims too. The whole process is iterative and takes a reasonable amount of time to achieve the right results.

Sorry to bring back an old thread to life, but I thought this is better then opening an new one with the same topic basically.

So my issue in short is that I can't get my injectors dialed in for about 2 months now. Long story short, here are the details of the situation.

We are talking about an EDM 07 STi with a heavily modified EJ257 engine, with 10.2 forged pistons, Cosworth cams, oversized valves, ported heads, Litchfield LM450 twinscroll turbo, FMIC with rotated intake manifold, DW 850 top feed injectors, I have the stock intake on temporarily, ...

Using RomRaider/ECUFLash/ECUEdit (A8DK100Z ECU) and so far I have the scalar at 820 cc/min (althought the flow sheet says they are ~890 at 100% duty) and I think I could get the latency tweaked, so that idle trims end up at avg -2% total and light load cruse trims are again around 0%.

However if I do a WOT pull (not even to redline, but only to ~6k RPM) the AFR measured by my Innovate MTX-L drops to 9.2-9.4 but the Final Fueling Base commands a 11.5-ish AFR.

I have the per injector compensations all zerod out, and as I mentioned above I have the OE intake (although without the bottom part of the housing, so the filter is directly exposed to the engine bay) on the car to be able to dial in the injectors, so I hope the OE MAF scaling at the high load range should not cause ~10% error in AFR. Not to mention that if I modify the scalar with 10% then I will end up above 900cc/min which would be very much off sync from what I red on the NASIOC/ROMRAIRDER forums as ballpark scalar numbers for these injectors (820-850).

So I am not sure if this might be a mechanical problem, like a boost leak, or there are some other compensation tables/logic beyond the tip-in enrichment ones that are not defined yet in the opensource/ecuedit definitions and are not included in the Final Fueling Base monitor when logging, or something similar.

The car otherwise pulls good, and runs smoothly.

Any insights, experience around FFB, or suggestions are appreciated.


It's always a good idea to find out where your MAF scaling is at prior to making an injector change. Despite what you'd like to expect, we often find that there is a significant error between the commanded and measured AFR under WOT. From what you've explained it sounds like your injector scaling and latency is probably close enough. You can then adjust the WOT error in the MAF scaling.

Thanks Andre.

Unfortunately the stock injectors are not available as I bought the car with 1000cc injectors, so I will probably do a boost leak test to be on the safe side, but I am more comfortable with the situation with your feedback and will probably do one more little change in the scaling and swap to the APS intake and scale that properly.

Be mindful that as you change the MAF scaling, the load calculation changes, and that can affect things like spark timing. If you reduce the airflow in he MAF table, load calculated goes down, and timing will likely advance.

For that reason I often leave the stock scaling with the stock intake on Subarus. Yeah it's not the most accurate due to production tolerances and vehicle to vehicle differences, but it keeps you have from having to deal with all the side effects of changing the MAF scale. It means you adjust the target AFR table as needed to the AFR you want and accept that offset. You just bias the table lean or rich as meeded. But for a non stock intake I go through the steps of making the Target and measured AFR align.

Alright, so at the moment I am everything but happy.

I was playing a little more with the injector setup and I was quite happy with the result (apart from some anomalies here and there in the log curve of STFT).

So I decided to change the intake and install my 70mm APS CAI. I changed the MAF scaling based on the difference in cross sectional area and even a little more at the bottom end of the curve and started the engine.

Fuel trims went through the roof right away and maxed out (this means 25+% positive correction). The ECU had a hard time to keep the engine running.

I just could not believe what I see, but anyway as a temporary solution I increased the MAF scaling at the lower section of the curve and retried. After engine start it still was compensating +20ish%.

First I thought this proves my earlier thoughts about a leak, but I did a quick test and I did not observe any hissing (except through the 3rd port of the EBCS). So I think I am good there.

Then I installed a set of OE injectors with OE settings, bot nothing significant difference in fuel trims. I then changed back to the factory intake (with MAF scaling obviously) and now I have a negative 20ish% correction.

In my last despondence, I uploaded an absolute factory map and the same negative corrections remained present.

Any thoughts? I am really out of ideas now.


so you pressurized the intake system to check for leaks? Like you hooked it up to an air compressor?

Yes. That is exactly what i was doing, but i was doing it only gently.

Alright, so the mystery is resolved (and I thought I share this with you guys so everyone can learn from it).

It is in my mind a while ago that this might be nothing else but a grounding issue (and I was right), but I wasn't sure how, as I checked the applicable wiring diagram and the MAF sensor is basically a direct wired sensor to the ECU and it has it's ground from the ECU, common with other sensors in the engine bay.

Yesterday night we had a little argument with my darling and I could not sleep. So I decided to get up around midnight and started (actually continued) to research more on this until I ran into this thread on NASIOC (https://forums.nasioc.com/forums/showthread.php?t=2403736).

Long story short, the APS CAI is an all aluminum piece of art and the Denso MAF sensor has a metal base plate that is connected to the ground pin of the sensor. Now the base plate makes contact with the CAI by nature as it is attached, and through the mounting screws too so it gets double grounded through the APS unit (not like the OE filter housing that is all plastic) and these cars being 11ish years old, I guess not all grounds in the car are equal any more and this makes a fake voltage reading of the sensor. (I was actually surprised at the first time I tried the APS CAI, that the MAFv was around 0.7V instead of the 1.4V that I saw with the OE filter housing at idle, because that is way off compared to the difference in the cross sectional areas of the two intakes, but I did not have an idea of what could have caused that.)

The way I tested this first is with the OE filter housing, engine running, and watching real-time fuel corrections on my laptop, I momentarily touched a test cable to the mounting bolt of the MAF sensor and a chassis ground and voila, the corrections shifted from the ~+/-3% range to +25% right away. When removed the test cable, it dropped back to normal. With this result I was already sure that I found the issue.

Knowing that I insulated the MAF sensor from the APS CAI (at the moment with temporary taping and stuff like that), flashed the new scaling and the correction was ~-10% (which makes sense as the scaling I uploaded was a little too high compared to the area differences). So in two more flashes I could get the idle corrections to ~+/-3%, and that makes me freaking happy at the moment.

I still need to do the permanent insulation of the MAF sensor, and then I can go and do the whole MAF scaling.

Thanks for everyone here for their ideas and credits go to the NASIOC guys.

That was a good read Bela, and thanks for taking the time to post your successful results, and good work finding that problem! OEM grounding schemes can be really frustrating at times!

Great information! Thanks Bela!

We usually reply within 12hrs (often sooner)

Need Help?

Need help choosing a course?

Experiencing website difficulties?

Or need to contact us for any other reason?