RomRaider

Seems doing a Hex search for 3F000000 BF000000 4F22C79A F808D59A finds the correct 3F000000 BF000000 values?

Hi Oren!

Requesting address for AZ1J500J , rom and def found here
viewtopic.php?f=34&t=6993

32640

Thanks Matt! I'll check that out - it's immediately followed by a routine if I recall correctly, so you're probably right (since most routines start with stuff being pushed to the stack).

I use a similar process when I define ROM's without IDA(HxD + Frontpage to modify the XML values) to confirm the routine is correct and not something unknown etc.

I checked it on a few of your original post ROM's first to confirm I had it right but the line underneath is the same every time as well so it is handy to have a backup confirmation

My new Innovate MTX-L will be showing up today so I am finally ready to test the theory that with the values zeroed out, there is no effect on ARFs from the missing rear o2 sesnor.

I need a little help to establish a baseline for the issue before I get make the change. I know that in general, the ECU runs AFR's richer with a missing rear o2 - but I don't have a good idea of how much and when. Can I see it at idle with the car fully warmed up? Right now, it usually targets 14.6 at idle. I have no cats in case that might already be affecting things.

If I warm the car up, remove the rear o2 sensor, and plug the hole, what ARF's should I expect to see at idle? Will it be immediately apparent or take some time to see the ECU's adjustment? Is it an immediate correction, learned correction, or both?

I'll do a before/after test with the zeroed out values and post back with the results.

Thanks!

Well I pulled the rear o2 sensor and replaced with my wideband. Even without the AF3 correction limits zeroed, it was still targeting 14.6 at idle. If I blipped the throttle, it would target 13.8 for a bit and then rise back up to 14.6.

I zeroed the limits and re-flashed. It still targeted 14.6 and sits on 13.8 for a bit after a blip.

I went out for a long drive and found that generally for afrs are at 14.7 (wideband & front o2) while cruising but every so often, especially after accelerating and going back to cruising, my ARFs would hang out around 13.6 for a while and eventually go back to 14.7.

My logs indicate that it was targeting around 13.6 based on corrections it was applying. I didn't actually log the fueling target.

Do you guys think my car is reacting to the missing rear o2 sensor? Based on other logs, it would seem like it is but again, I don't really know the full symptoms of rear o2 corrections.

Well I ended up running the MTX-L narrow-band signal into the ECU input for the rear o2 sensor. It doesn't target the rich idle/cruise ARFs anymore. I was also able to leave the rear o2 completely unplugged. There is some debate out there about having to leave it plugged in but it would see that as long as your have the heater circuit codes disabled and the ECU is seeing the proper voltage on the input, it targets the correct AFR.

It sounds like some of the other rear o2 sensor related tables that have yet to be defined might provide the control we need to allow a missing rear o2 to be completely ignored by the ECU.

Did you do wire in the analogue outputs at the ecu or the rear o2 sensor wiring? I am looking to do this as well.

Cheers,

Rich

Yes - since all my wiring for the MTX-L was under my radio, I wanted to hook up at the ECU rather than going back out the firewall and under the car. I couldn't find the service manual for my year so I used a multimeter to find the input at the ECU, cut it and connected it to the MTX-L output.

The missing sensor or malfunctioning sensor with all 08+ cars will force the car into a limp mode of sorts, where the car targets a richer AFR. I have seen it generally around the 13.3 mark. Its probably using an alternate map with richer AFR targets or an offset.
It usually starts this within about 5 minutes of running without the sensor or after a rear o2 code is detected. Deleting DTC's does not help. Also using the mechanical cel fix or running a DP like Agency power designed, will do the same thing, it throws codes for post catalyst too rich or too lean and goes into this limp mode of poor fuel economy.

Hi may I have address for A8DH100P? It's forester xt 06 EDM.

I believed this too based on posts from you and several others and this is exactly the behavior I observed when I removed the rear o2 sensor and had no input at the ECU. Once I added the input at the ECU from the MTX-L, it immediately stopped this behavior. I've logged my fuel targets and wideband readings in all sorts of scenarios and it never deviates from the target AFRs. I've been running this way for a month now.

It sounds like you have first hand experience trying to using a simulated narrow band input, is that correct? Did you verify the voltage input was correct and what the ECU was expecting to see?

On an unrelated (I think) note - I do experience rich fuel targets immediately after starting car, seemingly independent of ECT or anything else. It lasts for maybe 10 seconds or so after the car has started. I can see that the ECU is actually trying to target AFRs as low as 12s before going to 14.7. It never seemed to be an issue running on gas, but with E85, if I try to drive while its running that rich, I will throw a misfire CEL about 70% of the time. If I wait until the fuel target settles out, I don't have misfires. I can't find any tables that control this. Do you have any insight on this issue? Can this be tuned out?

I haven't tried simulating the narrowband but with over 2000 customers with 08+ WRX's, I have seen just about everything. If someone gets a CEL they email me.

The misfire problem at startup is due to the FPR being overloaded by the fuel pump running at 100% at startup for the first 30-40 seconds.
If you were to check your fuel pressure at startup the needle is bouncing back and forth quickly() abut 15 psi higher then it should be.
On gas its not an issue because it will fire at the rich lambda, e85 is such a rich fuel it won't fire at that rich lambda.
when you first start up the car the oxygen sensor isn't fired up so it can't make the corrections. IT also doesn't completely run off maf at start up.
My little trick to get rid of this is to force AFL at idle to be about -15% by scaling the maf higher in this range. IT won't help the first time you start but after the car learns to pull that -15% then it will start up properly almost every time. It's not perfect but it does help a lot. With an AP I can also 0 out the hot start enrichment and that seems to fix it as well since I no longer get misfires with my car is running close to 0 AFL on speed density and changing this table to 0 fixed it. Unfortunately, we don't have this table defined yet for opensource. It runs rich because of the fuel pressure issue but apparently not rich enough to misfire.

There are tables that define initial (fake) load and maf values to be used before any real readings become active. Maybe these tables could be revealed and altered to deal with rich E85 start-up problems.