RX-8 LS Swap - Tuning Help

Make: Mazda
Model: RX-8
Year: 2005
Engine: 5.3L LH6 (Gen 4)
Transmission: F-Body T56
Rear Gear Ratio: 4.10
PCM: E40
Operating System: 2005 Chevy 6.0 SSR Manual (1250154)
Station Compression ~10.1:1

Mods:
Summit Racing Stage 3 LS Camshaft (231/242 113+3, .600/.600 Lift)
Trick Flow? Double Platinum .660 LS Dual Valve Spring and Retainer Kits TFS-2500286P
Trunion rocker upgrade
Chevy Performance 12576341 42LB Fuel Injectors
Cylinder Heads 799 milled 0.010"
LS7 Lifters
Dorman LS2 Intake (copy of older Fast intake and flows like a TBSS)
Summit HDR 5/16 x .105 wall 7.375 pushrods
CX Racing 1-3/4" Long Tube Headers. Ceramic coated.
Monster Stage 2 Clutch and lightweight 18 lb flywheel
Dual 3" Exhaust with MagnaFlows
180 thermostat
Ported 87mm throttle body
Delphi AF10043 MAF Sensor
Fel-Pro 26191PT .048 Head Gasket
RX-8 Fuel Pressure is 59.5 PSI
Airaid custom cold air intake

Had a LS1 swapped 94 RX-7 back in the day and decided it was time for a project car again. Did this budget build and it's been running for 6 months without issues. Even has A/C!
Ive had trouble finding a local turner that would touch a swapped car. Attached is the current tune from someone local. PCM was from a 2005 Envoy but we flashed a base 2005 SSR LS2 file to start with since pinouts the same and allows for easy cruise control hookup.

Decided to purchase HP Tuners and AEM 30-0334 OBDII X-Series Wideband to learn myself.

Problems with current tune:
1. Cold startup is decent but could be better. It starts and dips down low once, but doesnt stall. Idle stabilizes quickly.
2. Hot starts suck and crank a long time. Hunts for idle very erratically 5+ times up/down before settling after 15-20 seconds. Goosing the throttle during those 20 seconds causes a weird sounding intake backfire.
3. Seem to get a lean intake backfire and stutter when applying more throttle in high load, low rpm situations in 5th or 6th gear on highway. The stutter goes away and car accelerated normally.
4. Insufficient O2 switching error code. O2s barely switch at idle. Common with longtubes but I found a thread on here that can supposedly help wake them up.
5. High negative LTFT trims when cruising on light throttle. -12 or higher.
6. Returns to idle slow under ~1500 rpm.
7. MPG seems worse than it should be. Only getting 14 on highway cruising 2000 rpm @ 74mph. Previous LS1 swapped FD RX-7 with similar specs was better.
8. Full throttle pulls can feel inconsistent depending on starting RPMs
9. Enable DFCO.

Would someone be able to take a look at the current tune and see if anything really sticks out that's incorrect? I've never tuned a car before.
Attached my low rpm high load stutter log file.

1. Fixed
2. Fixed
3. Fixed
6. Fixed
8. Fixed
9. Fixed

Removed one factory oxygen sensor in longtubes and installed the Bosch LSU 4.9. Turned off LTFTs, disabled MAF, and put car in open loop and went through 5 or so iterations fixing my Primary VE Table. Most cells show an average of -2% on rich side.
How does my VE Table look in the upper end?
From reading the VE Table seems to follow the torque curve, correct? Similar setups to mine peak torque around 5500 rpm and start dropping off. I haven't gotten many data points above 5750 rpm. Will try and get some more later this week.
There is a dip at ~2500 rpm that seems to coincide with a raspy exhaust note change between 2500-3100rpm.

Re-enabled the MAF and disabled speed density. Did not have any troubles tuning the MAF, but just like the VE.... I need to dial in above 5750 rpm. E40 PCM has a weird MAF bug/patch so I am not exactly sure how it behaves above 11,500Hz since the chart ends.

WOW! The car drives SOOOO much better now and feels more responsive. Cold and hot starts are like a stock engine. Only the slightest hint of low speed bucking. After everything else is dialed I'll remove the wideband, plugin in the factory O2 and see how everything behaves.

How can I tell the PCM to ignore oxygen sensors at idle? This cam has 10.5 degrees of overlap, causing the wideband to show 1.15 λ at idle. The gas smell at idle isnt bad when O2s arent trying to compensate for the extra oxygen. Still want to play with injection timing boundaries.
Attached is all the changes Ive made if anybody wants to compare.

You can't tell it to ignore the o2 sensors at idle. It's either they are working or you have them disabled completely.

Injection timing and what you want it to idle at should help. A higher idle like 850-900 would help but then you don't get that cammed sound.

I wrapped up my high RPM MAF and VE tuning, then re-enabled the oxygen sensors, LTFT, etc.
Attached is the latest tune file and a short drive around log. The car feels really good driving around and nothing felt funny. However, I need more time to confirm everything.
I am confused as to why vcm scanner shows my injectors are maxing out at 100% duty cycle. These are 42LB LS3 injectors at 59.5psi.
Also, my VE table it pretty much maxed out at 5500 rpm and 100kpa. Why is this? Dynamic airflow is disabled and it uses the MAF above 4000 rpm anyways.

How does the tune and log look?

You do have the long term trims carrying over into PE so that is adding to how much fuel it's giving. Hard to say if that is real or not, not sure if you reset the LTFT and fuel trim learning when you turned everything back on.

The wideband channel isn't reported so I can't see how closely the commanded AFR matches the actual at full throttle.

I don't have a LS3 tune on me to see if the rest of the injector data is correct. The flow rate, min injector pulse, short pulse adder and the offset data should all be changed when swapping injectors.

Welp. I copied in the injector data over from a Z06 LS7 E38 PCM since they are almost same and both Gen 4.
Offsets of the e40 go from 0,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80 and the e38 from -80,-70,-60,-50,-40,-30,-20,-10,0,10,20,30,40,50,60,70,80. I copied the matching values from the e38 and interpolated the missing values.

Tune seems totally out of whack now. Guess I'll start fresh with my VE table and MAF curve.

The Flow Rate, Injector Offsets, and Short Pulse Adders are different between a stock GM tune and Deatchwerks.
Should I use the values from GM or https://www.deatschwerks.com/gm-calibration-data?

Injector duty cycle dropped after entering the correct LS3 injector table in all tables and fixing my VE table accordingly. I tested my fuel pressure by removing the schrader valve from fuel rail and attached a long hose and gauge so i could see it while driving. It DID drop a few PSI from 59.5psi down to ~53 psi at WOT and 6800 rpm. The injector duty cycle is still higher than expected and hits 90%. Even with fuel pressure dropping a few PSI the injector duty cycle should never hit 80%

Attached is my current tune and log while 4400+ rpm was 10-12% rich. Injector duty cycle hit 93%. I still don't understand why its so high.
Tomorrow I'll attach a hose to the fuel rail and jump the fuel pump from fuse box. See how many long is takes to pump out 1 gallon and estimate its true lph.

I tested my fuel flow and it's roughly 1.1 gallons per minute out of the fuel rail. So roughly 240lph. Finalized my VE and MAF tables and injector duty cycle is hitting 83% now. Much better numbers but still seem a tad high for 42lb on a 5.3L. At this point I am all out of ideas.

Finally got my wideband bung installed around my X-pipe and logged everything. My VE and MAF tables were -1% to -2% percent rich everywhere in final tune when tuning them independently in open loop. When I turned on closed loop and both dynamic airflow and MAF.... I get many -7 to -12% rich STFT+LTFT in light load areas. With larger number higher up in the rpms.

Why is this?! What is the correct course of action with this new data?

Scanner log on dropbox since its large.
https://www.dropbox.com/s/gsfjecyw6c...nally.hpl?dl=0

Cant look at your log right now (on my phone).. Watch your O2 sensor volts. Someone correct me if I'm wrong, I'm still learning also.... but they should be reading around 450mv at 1 lambda.. Higher numbers is a rich condition.

I have a similar 5.3 build.. slightly milder cam 228/230 @ .050 and DW injectors as well.. I got the 42lb for flex fuel. Thanks a bunch for posting that link!

DW sent me some crap in email I cant make heads or tails of... Hopefully that link you posted will be easier to plug into my tune.

If you havent already, Maf and SD tunes hafta be redone after putting that data in.

Also, what are you Idling at in normal mode, Maf, and SD?

I can't go any lower than 1000rpm in Maf, and 1200 in SD if I dont want it to die at a stoplight. 😅

Something appears to be wrong with bank 2.. Both my O2 sensors are reading like that.
In normal mode, you'll see negative trims on bank 2. Shouldn't see much correction on bank 1... Might be a vacuum leak, or exhaust leak somewhere.
Logging Misfires on all 8 cylinders might also help figure out the issue.

I have a Magnaflow dual 3" muffler also... wondering if maybe it is a backpressure issue... Interested to see what the fix is.. Anyone have any ideas?

Everything else looked good to me..

For some reason bank 2 always runs 1-3% richer than bank 1 on average. Couldn't find any exhaust leaks. Checking injector o-ring seals this weekend.

I'm learning we may not be able to trust the wideband sensors.. Both my narrowbands are in agreement, but while they read rich, the WB is reading lean.

Gonna do some partial throttle tuning the next few days with LTFTs, until I can trust the WB. I emailed Matt at AEM about the WB sensor, maybe he can shed some light on it. Recalibrating my WB just made it worse. 😅

Did you ever get your banks balanced out?