RomRaider

Just thought I'd say hello and share my build progress and plans and get advice from the experienced. I'm just starting to learn tuning myself and really just reading..and re reading things; but in many ways because of my setup, I feel what I'm reading doesn't apply, or does so differently even though some core concepts remain. I've downloaded Airboys spreadsheet and the MAF scaling tool, read the various tuning guides I've found and watched random videos I've found. I re did my gear ratios in the Gear Determination Thresholds, adjusted for my wheel/tire size. Here's my setup: All Work done by myself in my driveway. Car is a USDM 98 Impreza RS Coupe. Some knowledge checks and questions below.


Engine:
Ej207 V7 swap (~30k miles)
VF30 (+Turbo Blanket/+Heat Sheild)
Intake AVCS wired, JDM ECU
Full SD Tune (In progress)
IAG AOS
Killer B Oil Pan/Pickup/Baffle/Fill Tube Relocation
EVAP Deleted
TB Coolant Bypass
Thermostat Removed
GD Koyo Radiator/Slim Fan

Intake:
Perrin Turbo Inlet
8mm Phenolic Intake Spacers
Reversed Manifold
FMIC/Custom Piping (2.25" Hotside/2.5" Coldside)/DEI Gold Reflect Tape
Blow Thru MAF/IAT Just before TB/after FMIC/Tial BOV (really just using the IAT)
GS 3 Port EBCS (31hz)-Mounted to Turbo for response time
GS Divorced DP (De-Catted), AEM UEGO WBO2 in rear most bung

Exhaust:
Kinugawa EL Header/Up Pipe (Titanium Wrapped)
Dual 2.5" Exhaust x2 H&S Scoobysports

Fuel
AEM 320lph/E85 Fuel Pump
Aeromotive FPR (45psi)
Radium Fuel Rails
Radium Fuel Pulse Dampeners
ID1000 Top Feed
-AN6 Lines
Parallel Delivery
GD Fuel Tank Swap

Without getting into the body, brakes, suspension, thats the stuff probably tuning related. My challenges and advantages are probably already apparent. First, because I've changed my intake, fueling, etc all at once, scales where are all useless. On the positive side, having a blow thru MAF and using SD brings back some advantage. In full disclosure, I am using a trusted tuner, at the moment I am logging and tuning VE with him. What I'm trying to do is understand what and why hes doing it (without annoying him with a million questions), and be able to understand and eventually take over or make adjustments as they arise. Because I did not make a base map from scratch, I probably cant post my map or logs for insite?

Can anyone tell me if my basic understandings are on track so far?

STFT (AF Correction)/LTFT(AF Learning) (added together should net -10% diff to target)
+ Too Lean (adding Fuel to hit target)
- Too Rich (removing fuel to hit target)

AFR
-14.7=Lean (Too Much Air) (increases fuel delivery)
+14.7=Rich (Too Much Fuel) (reduces fuel delivery)

Closed Loop (Uses Front 02 Sensor to maintain AFR target)
11:1-11.5:1 CL AFR ideal for FMIC setup
Open Loop (uses MAF/MAP and ECU data...user-set table)

WGDC
Higher=Closed WG (Builds Boost)
Lower=Open WG (Reduces Boost)

TD Proportaional=Immediate Correction
TD Intergral Positive=Undershooting Boost
TD Integral Negative=Overshooting Boost

CL/OL Fueling system status.
8 = Closed Loop
7 = Open Loop due insufficient coolant temp
10 = Open Loop
14 = Open Loop due to system failure

VE
+adds fuel
-removes fuel

Here's my current questions:

-Is there a trick to matching up lines in the log to account for delay?
-What is the process of tuning SD and VE tables, what are you looking at/for and in turn, adjusting? Simply if AFR is reading high, drop VE until its on target?
-Once VE is tuned, is it on to tuning boost, AVCS and then timing? Any specific order?
-I can only assume there is no need to tune the MAF scale?

If anyone wants to walk me though the their thought/check process when tuning, it might help me.

VE like mass air flow, accounts for mass of air, or engine load, both of which are very tied together
thinking that adjusting these to account for fueling issues is technically true, but what you're really doing is effecting engine load, which is a major component to fueling.

if your injector calibrations are off, then your engine load is off.

engine load error is directly proportional to fueling error, under correct calibration conditions, which is where adjusting MAF or VE for fueling becomes effective

when on all pressure based axis tables, engine load may not be necessary to be accurate because very little compensations or tables are engine load based, but it's a VERY important component of stock roms or anything using engine load as an axis for fueling/timing or other things. Which is where on the stock ROM, doing an intake swap without tuning it is kinda a big deal. Both engine load and fueling will be off, I've seen people adjust open loop to get fueling the way they want but this means that engine load is all off. timing/fueling tables become custom and so do the rest of the tables, which makes it incomparable to accurately set up ROMs.

so to reiterate:
VE+ more engine load
VE- less engine load
engine load is used to calculate injector pulse width
again, regardless if it's carberry, the group N rom, or any stock ROM, they all use engine load to calculate injector pulse width

The arrangement of pipes for your intercooler is an interesting setup. hot side goes under the driver's side manifold and cold side runs up the passenger side, which is not very common and kinda neat.

Spacers are only good for moving the manifold up, while they might reduce direct heat transfer from the metal to metal contact of the manifold and heads, it does nothing for heat convection. Imagine the engine as an oven. as it's running, it generates heat. the throttle being closed is like the oven door being closed.

not running a thermostat isn't really the best idea, you might look into reading about what happens when you never reach "operating temps".

Since you're getting it done with a tuner, unless he okays you posting it, then no, you can't share it.
you can, and someone like me might come along and snag it, but it will get taken down and a neat red message placed by a moderator to indicate the error in posting the ROM.

otherwise most of your assumptions are correct. turbodynamics, etc.

The spacers where actually just needed for flipping the manifold, I did need a little height bump.
The thermostat was sticking and causing issues so I pulled it, but Ill get one back in, probably a fail open style. My temps are staying between 160 and 205F sitting still, not sure what normal op is, but that seems ok. Car is only run in the summer, and temps have stayed between 70 and 90F. I thought I was pretty clever on the pipe route. I tossed the AC and HVAC system, its a shorter and more direct route, although likely makes little difference....its unique. Like having the dual scoobysports.

RR related question:
MAF Rescale....I have the MAF table open in RR, logger going....log overlay works fine, I get all my dots going and interpolate and when I hit update MAF table, it says it cant find it.....any advice here? Im likely eliminating the MAF sensor, since I can do full SD, but just curios it wasn't able to link up for some reason.

...same issue with injector one.

I did, however, clever me, figure out how to edit the cars.def to put in my cars weight and gear ratios and tire stuff, so that should make the dyno a little more accurate.

Ive also been editing the logger def, but so far have some issues. For example, if I want to change AF Learning #1 to just say LTFT, or Manifold Absolute Rel Sea Level to just say Boost, or whatever...basically just shortening the name so the logger is a little more compact.....do I have to change the name in several spots? I must be missing a step.

As of right now I'm running just Shell V Power 93 OCT, The station I use doesn't list as a pure gas so I am assuming Im running ~10% eth. @45psi, right now I have the injector scaling at 966.82
Latency
[Table2D]
6.5 9.0 11.5 14.0 16.5
2.652 1.712 1.26 0.992 0.8

Everything seems pretty stable for a starting point so far. Im really just logging and letting him do his thing his way, just trying to learn whats and whys and connect the dots in my head. Looks like we are starting with tuning the VE table first, I havent seen any changes made elsewhere in the various map iterations. So I can assume he is comparing target AFR and my fuel trim corrections and using that to raise or lower VE? In a setup/scenario like mine...how would other experienced tuners approach this?

yes, in order to get fueling down he's adjusting the VE based off fueling error

wideband for open loop, upstream oxygen sensor for closed loop

short term fuel trim + long term fuel trim = closed loop fueling error
final fueling base / wideband = open loop error

I haven't used the romraider MAF scaling program in a while, but I believe you need to have opened the MAF scaling and have it up for display. Of course, I'm not sure if the program works with Carberry or not, I actually don't prefer it. I use another piece of software.

since it's not actually using the MAF scaling, it'll be calculating against the VE table so you'd really want to manually plug in the data by logging MAF volts, mass air flow g/s (which is what VE will come out with) and fueling error related stuff. Then you can put fresh data into the MAF scale, and potentially switch over to that.

When I did blow thru I had two air straighteners, one on either side of the MAF housing portion of the cold side charge pipe. you might have wonky fueling or need to find a way to adjust the angle of the MAF sensor to the air charge going past it.

In either case, using the IAT of the MAF sensor in this location is fine, and staying speed density is fine.

Failsafe is good if the manufacturer is decent

198 to 205 is normal for being heat soaked around idle, stop and go traffic
187 to 192 is normal operating temps
160F might not be in enrichment (primary enrichment table is always active) but it's not hot enough to do good things with the crank case oil

personally I ignore all fueling data from under 187F engine coolant temp (ECT) unless I'm working on warm-up or similar tables

not being a daily it's probably not a big deal though about having the thermostat if it's not driven a ton
if you do, you'll want to inspect your oil more often for excess fuel and water content.

When using my WB02, its post "high flow cat" on the GS downpipe. I can punch it out, but feel bad doing so. But, it appears to be a roughly 3% difference from the front o2 reading, and about 10 rows delay. Should I just run a 10 row 3% reduction to keep it more accurate or....whats the norm?

I assume that I want to pool ranges of data to match the cells in the map and average them out? For example, if my VE RPM axis is 800 1200 1600, I can assume 0-1000rpm falls into the first cell, 1000-1400 into the next, and the ECU interpolates? etc?
And My MAP axis 5.03, 6.96, 8.90 10.83...a little harder to do but roughly 0-5.5 psia, 5.5-7.5 etc would group together and then average the error % out and reduce or add that percent?

I imagine there is already a tool/spreadsheet that takes a log and does this? The MAF VE Calc on the MAF scaling tool?

Sorry, might need my hand held a little while things click. Can I post up logs and maybe get feedback what people see, what they would do and I can use that in my learning? Ill get permission for maps, I dont think he will mind. Seems to me maps are very unique to each engine and combo of parts so I cant imagine mine being useful to anyone else.

But, so far, my process would be, load the log, and if say tuning closed loop, I'd auto filter, keep just the 8 status, kill off anything under say 180F coolant temp first. Make a column that combines STFT and LTFT, and then split the errors into "bins" of say RPM 0-1000, 1000-1400 and MAP PSia. Then average out those errors, and adjust that cell + or - that error in VE?

wideband should be before any catalysts, and the front oxygen sensor becomes erroneous above enough engine load (exhaust back pressure) so comparing a sensor which doesn't read correctly to a sensor that doesn't read correctly (after a catlayst) is no comparison to make

having catalysts isn't impossible, just get an o2 sensor bung welded just after the turbo on the downpipe. angle is important.

yes, the ecu interpolates. if your lowest cell is 800 rpm and you're at less rpm than that, it's using just that range of rows. if you're at 1000 rpm and your cells are 800/1200, then you're at half the value between 800 and 1200, interpolated. if you maxium rpm is 6800 and you're at anything above that, it's just using the 6800 rpm row.

yes, there are tools which take logs and data and make calculations, doing it manually with the data isn't necessary if you're just rescaling VE and you're already speed density, we've got software for that.

The downpipe has a bung at 12 oclock as it bends downward after the turbo/wg....it just seemed a little too close...like I'd be buying a new sensor every 3 months, so I used the post cat bung. I can punch out the cat....or add a bung just before it, I guess either/or. I was just looking at the data graphed, and they matched perfectly between AF Sensor 1 and WB02, minus 3% and the 10 rows, doing that I can get them to basically overlap as is. But, sounds like thats no bueno.

Once you've gone through and adjusted as many cells as you can for VE, do you re interpolate by rows or colums or blocks to smooth the VE table and then move on to adjust something else? I can only assume a choppy VE map viewed in 3D is also no bueno. But I anticipate as the cells are adjusted they will smooth themselves out.....

I def. need to make sure Im getting accurate data. part of why I moved the IAT to before the TB and switched to SD. Not sure if having the headers/up pipe helps the front O2 read a little better higher in the load range or not. At what point does the front o2 generally stop being useful?

more than 1.3 grams per revolution of engine load, or 4000 RPM, whichever happens first

having the wideband just behind the turbo is still a couple of feet of exhaust between the heads and the wideband

I had my Bosch LSU 4.2 wideband in for at least a year with no problems. You could add a copper heat sink/plate between the sensor and bung for some extra heat control.