SAFC II Rewired = Full lean idle

[dm1133]

Bought the car 4 years ago, no mods aside from AEM UEGO Wideband and SAFC II (all signal read fine, but adjustments would never affect the AFR).


At bottom there is a TL;DR

So no with more money and time, I'm looking to finally do some actual modding to the car, not to mention lean out the rich(~14) idle a bit :P. Other than that, car ran like a beaut, perfect AFRs on boost, rich on idle and cruise though. Reason probably being that the O2 (disconnected by previous owner, as they put AEM Wideband in original O2 bung, and didn't route NB signal to ECU, reading constant voltage of 0.31 (lean).


As mentioned, I had a look at the ECU and wiring etc... SAFC II was wired mostly right, except 1 cable. The AFM signal read cable they simply spliced into the MAF sensor, and original MAF sensor cable was still directly connected to the ECU. One very odd thing I found, was the MAF ground cable coming from the ECU was cut. Turns out they connected the SAFC modified AFM signal cable to the MAF ground. I was surprised the car even worked in this configuration. Anyways, I simply cut the MAF wire right after the SAFC II splices into it. Then I took the modified AFM signal coming from SAFC and spliced it with the other part of the MAF cable going into the ECU as it should've originally been, and obviously reconnected the MAF ground wire to the ECU MAF ground pinout. (Let me know if I possibly messed up somewhere along here).

Reconnect battery, start engine up, runs beautifully first 10 seconds, right at stoich. Adjustments to the SAFC are making significant AFR differences even at idle. But then it keeps consistently leaning out idle, and after about a minute it leans out the idle to over 18, until it finally shuts downs. Only way to keep it alive was to fool ECU into thinking 24% more air is coming in through SAFC, in idle.

Anyways, I think the simple fix would be just to connect my WB to give a NB signal to ECU, so I end up doing that, reconnect etc... Again, at cold start up, keeps beautiful stoich idle, and then after a bit, RPMs start dropping, leaning out like crazy. I brought base RPM all the way to 900 RPM. Running the AEM in P04 mode just to confirm, ran it in P03 to see if I have a different model, but that's definitely wrong signal.



I hooked up my consult, to check voltages. Again, only way to keep the car from stalling out or sputtering like a mofo at idle, is by increasing fuel injector duty using consult or +30% adjustment on SAFC. It turns out the voltage for O2 sensor in P04 mode is going over 1v, seen it hit all the way to 1.4v, should be 0-1v, so for some reason it's adding that 0.31V and whatever else, to the NB input my AEM is giving. Anyone know what the fix for this is? When I disconnect my AEM, it goes to the stale 0.31V again, and all that really does is require a bit less adjustment from my SAFC to keep the car running. The real problem seems to be the MAF signal coming from SAFC. For input at idle, it's showing 1.1V(direct MAF voltage) which is good I believe (2000 rpm = 1.8V good according to manual). However, for the modified voltage coming from SAFC, it's 1.02-1.04 base adjustment is 0%. So there's a discrepancy in signals, the SAFC is returning a lower air value than is actually coming from MAF sensor :S, bad wiring???


I'm think it's these voltage issues causing the idle to be ****, so I'm wondering if anyone has suggestions on how to fix these, grounds bad or what??? Since I've had the car, tach would always go crazy and barely ever shows anything correct, the RPM signal from ECU, showing in SAFC and consult is perfect, so maybe bad grounds???


I don't think it's AAC or any of those, one reason being, the car worked fine before the rewire, and I'm sure if I rewire back to the MAF cable going directly to ECU, it'll work fine again. But I want the ability to change A/F signal for fine tuning ofc. However, all suggestions are welcome, at the current moment the car is undriveable if I set the 800 rpm adjustment to +30%, as it will go pig rich when trying to get into first gear.

List of mods
AEM WB
SAFCII
HKS SQV (not SSQV) http://www.upgarage.com/upgarage/gaz...22001802s0.jpg
Australian air intake(forgot the brand lool, rated best suction and 2nd best filtering)

SAFC II settings
6 cylinder, rising
Sens. Cal 1:1

Attempted HW 1:1 / 4:4 / 2:5 all returned with lean idling conditions
HW 2:6 caused super rich idle condition, again starts stoich, moves down to 10.0 rich.

Should I be using HW?? I read someone used Flap and that fixed it for them?!?!

Sorry for the really long-ass post, but I'm trying to give you as much info as possible.

For TL;DR
******************
Rewired SAFC II to correct specifications according to manual, settings correct as far as manual concerned. Car ran perfectly before while it was only monitoring MAF signal. Now the 0% adjusted signal coming from SAFC is different voltage than the one from MAF, SAFC giving 1.02v while MAF is sending 1.1v signal.

Also O2 sensor consistently showed 0.31v since owning car(No O2 sensor connected, harness is sitting in bay unplugged). Cut wire from O2 harness, and spliced AEM UEGO NB signal into it. O2 consistently showing rich mixture, now because the 0.31v signal simply added w/e NB signal is giving, showing voltages anywhere from 0.6v to 1.4v.

Now car runs like **** at idle, beautiful cold-start up at 14.7 stoich, but keeps leaning out until full lean. Bad wiring or what else could it be, don't think it's vacuum leaks, boost gauge showing 20 on the vacuum. Any idea for fixes?

P.S. Thank god I copied this into my clipboard before submitting, turns out it took so long to write I got logged out .

P.P.S. Even after writing all this I forgot an important piece of info that it's a R32 GTS-t Type M. So RB20DET.

[Skym]

Simple answer to the biggest question I have ever seen. Put it all back to stock with wiring and don't use a SAFC. Then clean, redo grounding cables, fit new stock o2 sensor and check for intake manifold flange, exhaust manifold flange leaks, as they can cause wideband to produce lean result's, from memory it can go into 17's.

Also, should be alot rich than 14.7:1 when engine is cold and wideband is 0-5V, not 0-1V like stock narrowband.

[dm1133]

I know I can put it back to stock wiring, but as I said, I want the tuning ability of the SAFC, it's already there, might as well put it to use, for the mods I'm planning on doing I won't need something to the scale of a FC Power.

Yes, I know WBs are 0-5v, but the AEM come with an additional wire that can be simulated as an NB, because the wire returns WB reading in a 0-1v scale. You're basically telling me rip out everything you have, including WB and SAFC... first thing anyone recommends to get in a turbo car is a WB.

The car DEFINITELY runs lean at idle, RPMs drop and you can hear the engine isn't getting enough fuel, as I said, after a point it will stall out how lean it gets, it passes 18 in lean reading.

So I'm wondering if there's any fixes for the current system I have in place, with the voltage discrepancies in the O2 and SAFC MAF signal output. Redoing grounds??? And which grounds specifically if anyone knows.

[dm1133]

Did some more tinkering, with HW setting at 4:5 I was able to get a very close modified MAF signal from SAFC, only about a +-.02. Well, at least it was so for the first few minutes, as the car warmed up, there were sudden drops in the modified SAFC MAF signal coming into ECU. While the actual MAF signal stayed steady at 1.14v, the modified MAF signal would be at 1.13-1.15v and every 2 seconds or so do erratic jumps to 1.05v and go back up, causing it to lean out at times to 16, but still keeping mostly stoich when it returns.

So it's obviously a wiring issue, I would think. Are my connections bad? They aren't soldered, as I didn't wanna make semi-perm connections before testing it out. Is the SAFC screwy since the previous owner ran the modified single into the ground wire coming directly from the MAF? Bad groudns???

If someone with electrical knowledge could advise me on this and save me some hassle, I'd greatly appreciate it. The car has not had any mechanical issues for 4 years of my ownership, but it seems the electricity has always been a bit screwy, tach going crazy, and at times certain buttons on the A/C control not working (although the A/C blows mighty fine ), so could the culprit be simply bad grounds, if so can I find out which specific wires are grounded bad, I'd rather not have to redo every single ground... it's kinda reinventing the wheel.

[Skym]

I'll elaborate, grounding incorrectly can damage Apexi products (in some manuals there's 2x ways to ground their products, depending on features used). That's why I was saying to remove SAFC (before it damages engine). It can happen with their turbo timer if (from what I can remember) incorrectly earthing it to ignition barrel, which permantly damages turbo timer (causes it to not switch to turbo timer and shut engine off). They recommend to replace product and ground correctly. With what a new SAFC cost's and what it does, it's not worth it. If have to adjust tune, do it properly with Nistune or similar, as stock ECU can screw up a tune with SAFC if switches to knock maps, etc, which ECU probably will with running that lean.

I'm familiar with AEM wideband adjusting signal, so can be used by ECU, just don't trust 0-5V to 0-1V conversion. Rather have 0-1V from stock narrowband.

[dm1133]

Oki, now the dilemma is, if I connect the narrowband, it'll show voltages ranging from 0.3v to 1.3v , thereby almost consistently indicating a rich mixture to the ECU. Even with narrowband d/c, it will show a base voltage of 0.3v.

[Skym]

Sounds like intake manifold airleak, as o2 sensor voltage can bounce up and down (hitting lean, rich voltage triggers on ECU) and revs rise (lean), fall (rich).

[dm1133]

Didn't know if I cut the O2 sensor wire, leaving only a 1 inch covered wire going into the ECU, that I presume should be showing 0v, but instead showing .31v indicates an intake manifold leak :P.

It's the wiring I believe is the problem, but I'm no electrician, so I have no clue why a 1 inch wire going into the ECU would still be showing a 0.31v reading. Bad ground or bad wire or????

[Skym]

Just the voltage fluctuating up and down that much (0.3V to 1.3V) sounded similar to a intake manifold airleak. The voltage triggers are set not at 0V, but before 0V (I think is 0.3V = lean), same with before 1V (I think is 0.7V = rich). Unless someone has adjusted them on ECU side that you don't know about or narrowband is faulty, thus lazy to recover from going too rich.

The constant voltage (0.31V) is normal when unplugging stock narrrowband o2 sensor, as my car did similar with voltage. It's the ECU using a default value. It does the same with other sensors (I think was MAF, TPS, etc).