|
it's generally not any of our responses to tell people they're dumb or put them in a spot
most of us just point things out factually, like "this is better" or "that's not good"
our community may not be super active but we're not like NASIOC, for sure
if someone posts something dumb, or generally something not interesting that could be answered by properly searching, nobody responds. glossed over, boring.
it takes a while to look at everything in a Carberry ROM, so as I'm going from bit to bit reading, I'm also writing this, so you can follow the table names as they come in order in the rom. This method of looking at the ROM might also help others.
Knock CEL:
I don't remember what 0.00 does for both values, but I do know that putting in positive values will result in a constant CEL. You want to put in negative numbers here, like if you see -2.11 knock and have it set to -1.05, it will show at anything greater than 1 degree of timing being pulled, so two degrees would definitely flash. I think that 0.00 results in a constant CEL, do you have one?
Anti-lag/Launch Control/Flat Foot Shifting/Flex Fuel/Alcohol Injection: ignored
TGV Control: looks like what others have used, nothing to see here
Global Map switching: ignored
Fuel Pump Duty: not much to see here. looks fine for the setup.
Boost Control - Target:
per gear is fantastic should you choose to use it. there is one table, with one cell for each gear (except 6th, which shares the same value as 5th) that affects percent total compensation
target boost - doesn't look bad, seems realistic if duty settings and corrections are in line. I like running from 0.60v to 3.6v in 0.6v increments. This permits getting really good tip-in enrichment, which I'll talk about later
Boost Control - Limits:
the most applicable table later on here is overboost fuel cut. the others can be used to disable boost depending on either IAM or IAM and fine learned knock, but as it is now you need to drop to 5 or less on IAM in order to care about boost control disabling wastegate duty altogether.
Boost Control - Wastegate:
Looks like min/max are swapped. You've got minimum WGD at higher values than maximum. I'm not sure how the ECU responds to this. Is your current logged WGD whatever max table says? Ideally, WGD starts out at minimum and checks boost error to determine if turbodynamics (TD) is needed, and if so, how much and what type (proportional or integral) to add, up to maximum. Again, I recommend scaling the X-axis volt values out wider so that lower throttle areas of cruise have greater accuracy in boost error (less is better, always) specifically because of the improvements to tip-in enrichment when boost error is not 2 PSI+. For posterity, all throttle voltage related tables are scaled the same, such as all manifold pressure axises use the same values, depending on what table function is being used and the overall limits of resolution in the table (table size)
other table values look like stock group N, you might find changing PWM frequency for the BCS to 15hz if you're using the stock boost controller.
Boost Control - turbodynamics:
I'm pretty sure all these tables are irrelevant given that max WGD is less than minimum. Then again, I've not purposefully run lower max than min just to see what happens.
Manifold Pressure Sensor:
looks stock, no problems here
Primary Open Loop:
people can argue about AFR all day. if the knock resistance of the fuel is good, this table looks like transitional area to full boost will be lean enough to promote quicker spool but not be richer than needed.
top end of the table for peak boost (stock level to balls out) is in the fat and happy area, a little richer than needed if timing is not too low. Too low of timing and EGT goes up, meaning if timing is low you will actually want a richer AFR like this shows. Of couse saying this now, and remembering you implied having a stock downpipe, it might benefit the safety of the life of the catalysts to reduce a bit from 11.2 down. but then again my lack of experience in tuning with catalysts doesn't tell me what they will tolerate as far as how lean.
the other transitional tables look like what others run, pretty normal to see this
Closed Loop:
looks typical. base = 14.45, plus load comp of 0.23, equals closed loop is always 14.68:1 gas AFR
CL/OL Transition:
most people drop base pulse width determination to zero. 2707.09/scalar * g/rev = BPW. This basically means for stock injectors, you'd need to see just before 1 g/rev for this switch to consider the open loop transition, which is fine. maybe you want to run a richer AFR outside of closed loop in the higher rpm vacuum range of manifold pressure, lowering this to zero would enable running what the OL table calls for more clearly.
for throttle determination you'd need to be above 50% to consider switching, which is entirely possible you'd want to if you're seeing enough boost at just before half throttle. lowering this to zero would of course make open loop table be more directly used. basically if you don't spend much time cruising in higher rpms or spend too much time on boost under 50% throttle, you're fine.
Injectors:
stock scaling, stock latency
stock group N injector timing, it works but I've played with this alot and ended up making it effectively be the same time as when the valve would be open for non-avcs.
per injector compensation: it's actually based off the volumetric efficiency difference of the cylinder runners while operating, not having anything to do with fuel. Sure, there is an air/fuel relationship, but it's stock intention is to compensate for runner differences. You might see all zeros on ej207s that have different heads, no tgvs, and different exhausts. For stock ej205 manifold with or without tgv deletes, I run the stock value or slightly less. Look to a stock MAF rom for the values.
cranking:
looks like stock group N. it is intended for a different engine setup with 20% larger injectors. A+C are much fatter than necessary for these injectors and would probably benefit cranking quality greatly, by being lower in value.
take 552/420 as a factor, and multiply that against these tables on the values only, not the axis.
tip-in enrichment:
this is the sort of thing that will be new to you, how would you know right? Stock tip in enrichment is super dumb. Since boost error will almost always be high when figuring in the target boost table and realistic running conditions for cruising and throttle transition, it will basically add more fuel over a longer period of time than needed. Of course this happens in the closed loop area, so short term fuel trims will try to compensate.
decel fuel cut:
stock-ish. play with this later if you want, after doing other things concerning drivability.
warm up enrichment:
if you've got TGVs, run the stock tables from an 04 ADM STi, I prefer these over stock USDM EJ205 tables.
if you go without TGVs, starting with JDM EJ205 settings is usually recommended, but I say screw it and start with group N or S203 warm up enrichment and go from there.
if you want to know more about how these values work and what it all means, I can expand on it. Basically the table description does a good job of describing it, but some people refuse to accept the true math. It's clear to me, I'm bitter. it's a factual black and white thing. for sure though, warm up enrichment is way more complicated than it needs to be.
air/fuel learning:
stock MAF roms are based off mass air flow grams per second airflow ranges, and carberry bases it off of manifold pressure. having trims applied to open loop means that if open loop is being requested, and there has any learned compensation for trim D, it is applied to requested pulse width. So if your MAF scaling is off when in trim D, then open loop will be off as well. If everything in the range of trim D is off by the same percentage as everything in the MAF scale above trim D range, then cool. This is basically what compensates for fuel types, since not all pump gas is equal. But MAP axis based MAF fueling is a weird combo, so most people disable it and tune the MAF really well for open loop, and then never enable it again. If you always use the same pump and there isn't huge changes in region/season fuel (1 through 4, not all regions use all four, and only hawaii uses #1 seasonal fuel all year round) you may not necessarily need to enable it.
MAF/SD blending:
set to MAF of course. the load smoothing table is stock for Carberry, and IMO I think it is roughly based off of Cobb's table. I actually like Cobb's table better for MAF on hybrid speed density. 87.5% across the board for 100% MAF, linear interpolate to all 0s at full SD. Which is technically different at the blended and full percentage, but from what I've used, this is what I like.
speed density:
stock group N stuff, intended for a different engine with a completely different VE profile from it's sum of parts and group N restrictor plate on the compressor inlet. not really what your VE would look like.
mass air flow:
stock tables for USDM ej205, you might find that rescaling the MAF is necessary to keep requested fueling accurate to wideband
Timing Advance:
note: stock MAF ROM uses base timing plus knock correction advance (also known as dynamic advance multiplier) and carberry uses base timing with knock correction retard. On the OEM, if IAM is full at 16, then base timing + KCA value = total timing (without other compensations) and half IAM at 8 would present base timing + half KCA value as total timing without other compensations. Inversely, if IAM is 16 on Carberry, then base timing = total timing (without other compensations) and if IAM is half at 8, then total timing is base timing - half the KCR value for that operating cell. It's the maximum that dynamic advance can pull if you were to go down to IAM = 0. Values less than the one set to activate rough correction (IAM correction) will result in an area where rough correction will not happen, except maybe just after reflash when the ECU is more likely to use rough correction over feedback or fine learned, even if proper settings permit the other two to be active.
ignition dwell - stock table for grey coil packs. black coil packs actually handle a wider range of operable loading time (dwell time) and tuning here isn't really required
x-axis - you can set it to engine load, but if your MAF scaling is off, it's better to start out in MAP timing. you don't have MAF error making changes to timing like engine load would if it were wrong. If you do follow through and switch it, the process does no conversion of the units in the table. The values are whatever were in those cells before, and do not reflect what your engine would have as an equivalent to converting the prior MAP axis based table to run the same in g/rev for that engine. That's more complicated of a process and requires getting a ton of VE data.
base timing - there is a large drop in timing across a small amount of manifold pressure increase. more than 8 degrees for less than 10kPa difference? that's a big jump. you might horizontally blend 14.7 to 24 PSI and work around from there.
base timing idle, all three - looks typical. 14 degrees for idle. you can tweak this later with decel fuel cut to have a better transition from engine braking to cruising when fueling with no throttle.
Ignition Timing - Compensation:
the tables look like stock Group N, which uses the IAT in the MAF housing for everything fueling, so is probably just as reasonable to start with as your IAT is also in the MAF. EJ207s with manifold IATs used them for the intercooler auto-wash, not fueling. You can expand on these tables later if you want. Zero them out and make your own values, centering them where you like, or look at other ROMs to see what Subaru did with other vehicles and engines.
In here I like settings that are quicker to call it knock, remove less timing and consider to add or subtract quicker. More sensitive, more active, less hold and wait. At full balls tuned out for a VF turbo, pulling two degrees and then adding it back in during the same pull can be pulling and adding back in 20 ft/lbs, which you can feel. I feel like group N stuff is better than stock, but it's still kinda lazy. I'll share my recommended settings later.
Knock Sensor Parameters:
more sensitive in some cases than stock USDM MAF would show, but is easy enough to understand. there isn't a big need to change anything here.
Variable Valve Timing:
provided you have the heads equipped with AVCS solenoids, AVCS cams and cam gears, along with a JDM or ADM 16-bit ECU with provisions for AVCS, then these tables would matter but alas they do not. I zero these out for posterity.
Misc:
auto-trans logic can be used to make MT and AT tables be switching tables (not super useful), along with being an optional switch for other features. one just needs to have an on/off switch for grounding pin E1. Anti-lag? Flex fuel? you want to make an economy map type of switch?
I'm not sure how much the defrost being on effects engine calculations, and I've not messed with the pin for defrost, but keeping this on and using the defrost button itself for activation of a table switch totally works.
neutral position safety switch is opposite for JDM transmissions. reverse this if using the JDM transmission.
rev limit is for fuel cut, as described
CEL Fixes:
on an non-avcs USDM ecu, how the ROM comes freshly flashed after patching to the Carberry ROM, will show a few DTCs under CEL status. You can use 'learning view RC3' or romraider logger if using the right definition for the logger to see what they are, and then disable all the codes related to the parts that we don't have going on for our engine setup.
Secondary ignition, OCV, and Cam sensor B should be turned off
evap purge or boost controller circuit DTCs off if you remove them later
Sensor Scalings:
as it looks, coolant temp and intake temp sensor scalings. stock intake temp sensor is in the MAF body, it's the thermistor you can see on the outside of the chamber. The thermistor for mass air is inside the chamber of the MAF sensor body.
Thresholds:
Gear determination thresholds is probably the first table you should check. Your transmission's gears are probably different than the group N came from, so yeah. This helps for per gear compensations
Radiator fan mods is fun. Like since I don't have A/C, I've turned my A/C switch into a fan-on switch. The fans still work depending on coolant temp and road speed, but hitting the a/c button forces both fans on.
intercooler auto-wash only matters for those ECUs that have it
Idle Control:
most of these tables don't need changed, but change as you may to tailor your experience
o2 smoothing:
turn it on, set VSS threshold to 0, and you now have a much more active closed loop correction system
OBD2:
force pass readiness monitors, it force passes OBD readiness monitors
Diagnostic Trouble Codes:
TGVs, rear 02s, whatever you remove that you don't need a DTC/CEL for.
_________________
if you're generous, feel free to donate.
venmo @ jon7009, 1047 when asked
jedilley@gmail.com for paypal
Last edited by jon7009 on Mon Feb 26, 2018 5:23 pm, edited 3 times in total.
|
