Primary O2 Problems after Euro Header Install

[Icecream]

Glad to see it's coming together. Maybe you twisted the wires and damaged them when screwing then on. GL!

[ATB88]

And plot twist: it’s not the O2s.

I swapped the bank 2 O2 and sure enough, it started switching and behaving flawlessly, and the codes went away. Great! Buttoned everything up with both of the O2s replaced, checked multiple times throughout that everything still works and the sensors are reading and switching as they should, and no codes. Before putting the heat shields back on. After putting the heat shield back on. Before disconnecting the secondary O2s and EGT. After disconnecting the secondary O2s and EGT. Both sensors were working flawlessly.

I pack the car and head off, and sure enough I get an SES. Codes are back but this time only for bank 2! Bank 1 voltage is switching high/low perfectly as it should. Bank 2 now reads 0V. Wtf.

I guess the hint here is that swapping the sensors solved the problem until I started moving again... but I am seriously out of ideas. I feel like I have checked every wire, cleaned every connector. But, down to try anyone's suggestion if they have one! :]

Maybe some signals or wire continuities that I should check?

[BMWahba]

Damn, following this. Header install coming up for me here so I wish you the best of luck figuring this out!

[Cubieman]

Is there any remote chance that the wiring running to the connectors is damaged? I haven't noticed you talking about doing a valve cover gasket lately but I have read that some have cut up their O2/EGT wiring during the process. Man it sure sounds a like a software issue....and get a charger on there when flashing!!
Sorry I couldn't be of more help man, you gave me the confidence in one post to pull my fuel rail a while back! Good luck.

[ATB88]

Thanks! I can't say I've ever heard of anyone having these kinds of problems on a header install before :/ but I can't tell for the life of me what I've done different. I was pretty careful with the sensors. Also, while software is a reasonable candidate for behavior this weird, I think that the fact that I was able to make the problem temporarily go away points to it being hardware? Also the last time my valve cover was off was last October, the wiring was treated carefully and I had zero issues between then and now. Trying to flash again with a charger is a reasonable suggestion, but, I actually don't have an outdoor outlet anywhere where I could plug the car in as I live in grad student housing on a university campus.

I've been working on this since 4am and am quite defeated and confused. Overall, here is a summary in one place of the behavior I'm seeing and what I've tried:

-After installing euro headers and deleting post-cat O2s and EGT, I flash a new tune to accommodate the O2 delete and start seeing error codes 0x90 and 0x91, "function, oxygen-sensor control cyl 1-3" and "function, oxygen-sensor control cyl 4-6"

-When starting the car the O2 sensor voltages start at 0.45V and then within about a minute or less descend to 0V or quite close to it. Bank 1 plunges faster than bank 2. When I turn off the car but keep monitoring the O2 voltages, within a minute or two, the voltages slowly climb up from around 0 back to up 0.45V. So it doesn't exactly seem like the sensors are dead? Eventually after fiddling with connectors, I can get bank 2 to settle at a voltage of 0.15ish V but it isn't switching high/low like normal behavior. Bank 1 still sits between 0 and 0.05V. Codes persist for both banks. Code 0xD "signal, oxygen-sensor cyl 1-3" joins the party.

-I tried switching back to the old tune I had before doing the headers, and connect back the old postcat O2s and EGT sensor. Still have the same problem with the pre-cat O2s.

-I then dig back into the exhaust and decide to replace the bank 1 sensor with a used bank 1 pre-cat sensor I had lying around which I knew wasn't faulty. Also clean all connectors with QD electrical contact cleaner. I turn the car on, and lo and behold, it works! Bank 1's O2 sensor voltage is switching between 0.1V and 0.9 V as expected. Bank 2, which I didn't yet swap out, is still reading about 0.15V constant. Only the bank 2 code persists.

-I then swap out bank 2 sensor with another used bank 2 pre-cat sensor I had, and it fixes that bank as well. So, now, I have both banks showing the expected oscillating signal between 0.1V and 0.9V on both banks, and no codes! Success! At this point, I'm still reverted back to the tune that I was running when I had the stock catted headers, with the post-cat O2s and EGT plugged into the harness.

-I carefully put the heatshields back and get everything put back together. I check again that it's still fixed: both banks behave as expected, and no codes. So far so good. I then flash the tune that deletes the post-cat O2s and EGT. Turn the car on, run it for a few mintues, still fixed!

-I drive to the grocery store and get an SES light. Check the O2 signals: bank 1 is still working perfectly, bank 2 is reading a flat 0V. Two codes: "function, oxygen-sensor control cyl 4-6" and "signal, oxygen-sensor cyl 4-6".

-Carefully check everything, all the wires and connectors still look the way they should. Very puzzled why bank 2 decided to take a dump all of the sudden.

-Start fiddling with the connectors trying to gently re-orient them hoping for a miracle. Disconnect and re-connect a few times. Now bank 1's acting up again too, and its code is back. I am completely defeated.

It seems like it could have something to do with the connections to the harness, but it's not clear. Especially when replacing each sensor one by one made them work again, at least for a time, before going bad again, after running long enough on the cat-delete tune. Also when you turn the engine off but keep the ignition on you can see the voltages come back up. This makes me think that maybe something's coming out of the engine that's messing the sensors up? But what? Or it could be the harness connections for all I know.. but how do I troubleshoot that?

Desperate for more suggestions :/

[Cubieman]

I do not know where the o2 sensors terminate at in the DME connections but you were in there for the MAP and/or IAT wiring, maybe double check in there for anything looking out sorts. I believe the 02 sensor wires come in the though the rear most grommet in the DME box.

[ATB88]

Yeah, good suggestion. I was dreading having to bust out the Bentley for wiring diagrams.. but, luckily I tried one more thing that has shown me that the problem is, indeed, the tune, which seems to be making the O2 sensors permanently fail. Either I somehow royally screwed up my tune/program version combo (I definitely consider this the most likely scenario), or there's something subtly funky going on with Martyn's tool that caused this. Now I need help figuring out which.

Here's the reasoning that allowed me to isolate and fix the issue:

-The pattern I noticed: the O2 sensors were totally fine on the original catted tune, and then seemed to become irreversibly f***ed after running them on the cat-delete tune for more than 10 minutes.

(details of that pattern:

-I noticed that the pre-cat O2s that I installed with the headers were working perfectly upon install before I switched to the cat-delete tune I made with Martyn's tool. After switching to the cat-delete tune, they were working fine for the first 10 minutes or so, then everything tanked and both stopped being responsive. I switched back to the original catted tune I had before doing headers and the primary O2s still were unresponsive.

-Still on the original catted tune, while troubleshooting earlier today, I decided to swap in the extra used set of pre-cat O2s I had lying around. Like mentioned before, this seemed to completely fix the issue. The pre-cat O2s were now perfectly responsive. To sanity check, I swapped back in one of the other pre-cat O2s (that went dead after the initial header install and cat-delete tune), and it went back to being unresponsive. This made me believe that the O2 sensors themselves were the problem. Like I mentioned before, as I re-assembled the heat shields etc I kept running the car and checking that the O2s were still behaving, and they were: switching voltages as expected, and no codes.

-Once I was satisfied that I thought the problem was resolved by replacing the primary O2s with my spares, I then decided to re-flash the cat-delete tune that I made with Martyn's tool. After flashing, I turned the engine on, monitored O2 voltages, and they were behaving properly, and no error codes. So, this further reinforced my thinking that the problem was just the other O2 sensors, and not the tune. All makes sense.

-I start driving and a few minutes in, I get that SES and find that bank 2 O2 is now dead. I get very depressed, start trying to troubleshoot again, and during the process of troubleshooting, bank 1 sensor eventually dies too. Both sensors dead. Swapped the other sensors back in, they're still acting dead too)

-Hypothesis: the cat-delete tune is somehow killing O2s.

-Test used to validate hypothesis: Get a fresh set of O2 sensors, and run them for a while on the catted tune and see if they die. If they don't it would appear that it's the cat-delete tune that is causing the problem. I happened to have a pair of BNIB never used post-cat O2s. The bank 1 post-cat sensor actually happens to have an identical connector to the pre-cat bank 2 sensor, so I can use that one no problem. I go back to the catted tune. I swap in the sensor, and, what do you know, it fixes that bank -- voltage is switching again and no bank 2 codes. I want to swap in a fresh sensor for bank 1, but the bank 2 post-cat sensor's connector doesn't match the bank 1 pre-cat's. F*** it, I'm so close to figuring it out, so I get out an angle grinder and remove any trace of distinguishability from the new sensor's connector so I can plug it in to bank 1. Bam, voltage looks great, both banks are now behaving perfectly. No codes (other than for EGT and post-cat O2s, since I deleted those and this is the catted tune).

-Finally, I go and drive for an hour. To see if the pre-cat O2s continue to behave, or if they die like the last two sets did. They keep working. They still work now. Still on the catted tune.

Based on the above, my expectation is that if flash the cat-delete tune and go for a drive, both sensors will die, just like the last two sets did under the cat-delete tune. I don't want to try my luck until I can figure out what's going on here. I also expect that if I swap in any of the sensors from the last two sets that died, they will still be dead. This would pretty much absolutely confirm that the tune killed the sensors. I'm going to do that in the morning when the car's cool.

So, assuming that my reasoning holds up... why is the cat-delete tune killing the O2s? The tune is literally just the tune I was running before installing the euro headers (an 0401-PD31 partial), which was working perfectly at that point, loaded into Martyn's tool to have EGT disabled, secondary O2s deleted, and cat protection disabled.

Possibilities for my user error in dealing with the tunes:

1. The full binary I'm running is Terra's modified CSL program version, which I believe is based on the 0401-PD11 program version. However, the tune that I'm running (both the catted tune that worked fine and the cat-delete tune that is apparently not fine) is identified in Martyn's tool as 0401-PD31 program version. Not sure how this happened or if this "mismatch" matters and can cause the problems I'm experiencing.

2. The only modification I've made to my tune other than through Martyn's tool is modifying the cam offsets in TunerPro to match non-CSL spec cams. I did this using the 0401 XDF file found in the github stickied in the coding section. This is a pretty simple modification, and I believe that I did it correctly (which seems to be confirmed by the fact that I don't have any VANOS or timing related codes).

So... any ideas on how to figure out what's going on with the tunes? Also, how could a tune permanently ruin an O2 sensor in 10 minutes?

Also, in the meantime: I'm running a tune that thinks I have post-cat O2s and an EGT sensor, but I don't. So it's throwing codes for those, and that's fine. Would this cause any other drivability issues? Would this keep the fuel system from getting into closed-loop or anything like that?

[ATB88]

Well, the shitshow continues:

Went for a drive again this morning to check if everything was still working, and things still seemed to be great: both primary O2s behaving normally, and no codes for primary O2s. Then 10 minutes in I get the 'ol SES light. Check the scanner: code 0x91 (function, oxygen sensor control cyl 4-6). Check the sensor voltages and indeed, bank 2 has tanked to 0V, and bank 1 is still working perfectly. This is still on the pre-cat delete tune that I had been using before this entire debacle started.

So, this seems to also rule out the "cat delete" tune I made with Martyn's tool.

Quite frustrated now. The only thing that seems to be an apparent fact is that swapping in new primary O2 sensors always seems to fix the problem temporarily, but they eventually go bad again. And once they've "gone bad", they never seem to work properly ever again -- swapping in any of the sensors that have already gone bad doesn't improve anything, they stay bad. But swapping in a new sensor always fixes the problem, at least temporarily.

So, now I need to ask: what running conditions could cause this to happen?

I'm going to re-make my entire tune from scratch, to see if I can rule out errors I've made in flashing/modifying binaries. But beyond that, any other suggestions?

[bigjae46]

Sounds like you keep frying the O2 sensors. I’d try and take a resistance measurement of the heater and sensor circuit of a new and old O2 sensors. That will let you know if you have an open circuit.

I had a similar baffling issue with a misfire which ended up being a bad DME. Had to bust out the multimeter. Resistance tested every wire in the ignition harness.

Remember... a multimeter is your best friend and will save you a lot of time and money.

[ATB88]

Sounds like you keep frying the O2 sensors. I’d try and take a resistance measurement of the heater and sensor circuit of a new and old O2 sensors. That will let you know if you have an open circuit.

I had a similar baffling issue with a misfire which ended up being a bad DME. Had to bust out the multimeter. Resistance tested every wire in the ignition harness.

Remember... a multimeter is your best friend and will save you a lot of time and money.

Yeah, good call. It's starting to look like I need to dig into the wiring side of things if the tune isn't the issue, cuz that's really the only thing left. I'll try that when my new O2s come in on Tuesday! I should at the very least be able to confirm that the old O2s are all in fact dead by comparing to new ones.

One idea I've had on how I could be killing the O2 sensors: poorly routed wiring. The O2 sensor cables have heatproof sheathing, but the headers get extremely hot. If the wires were poorly routed such that they sit physically against the headers or some other very hot surface while driving, would that be enough to wreck the wiring and cause an open circuit? I've been trying to be careful with how I route the wiring, using the clips on the heat shields, but maybe they're still shifting around enough while driving that the cables touch the headers and fry?

[Cubieman]

One idea I've had on how I could be killing the O2 sensors: poorly routed wiring. The O2 sensor cables have heatproof sheathing, but the headers get extremely hot. If the wires were poorly routed such that they sit physically against the headers or some other very hot surface while driving, would that be enough to wreck the wiring and cause an open circuit? I've been trying to be careful with how I route the wiring, using the clips on the heat shields, but maybe they're still shifting around enough while driving that the cables touch the headers and fry?

This^
I noticed when I did euro headers that when I routed the pre-cat 02 wires into the clips on the heat sheild one wire was damn near touching the header and the other was < 1" away from the "flat spot" on one of the header pipes.

I noticed this and poked a hole into the kind of corrugated looking heat shielding sitting right at the fire wall and then zip tied both pre-cat o2 sensor wires securely with high temperature zip ties. This moves the wires significantly further away as opposed to running them through the heat shield clips.

[Bostonbruck]

I’m almost done with my euro header install. I’m really hoping I don’t have any issues.

[nextelbuddy]

i used your same process with Martyns tool on my car with the modified CSL tune and did not have these issues. i had stock manifolds but in my tune i still had rear o2s and CAT DTCs disabled leaving only the primary o2 sensors enabled and functioning. i ran my CSL airbox wth the CSL tune set exactly how you have it set for over a year with no problems. i also did several other CSL tunes wit that same tune for 3 other local people and had no problems from them either.

the only difference is that I was not using the latest Terra fixed CSL PROG file for reading error codes. i was using the original Paffy mssflasher modified CSL PROG and TUNE files with stock CSL cam offsets.

so theres 2 different variiables between your setup and mine software wise.

- i was using Paffy CSL PRO file not terras fixed error code file
- i was using stock CSL cam offsets


try changing only one thing at a time.

i would start by going back to stock CSL offsets. dont remodify it with tunerpro and tat XDF just go back to a stock paffy modified CSL tune file and then use martyns tool to change it to 6MT and mark the options you want and flash it.


if that doesnt help then flash the paffy CSL PROG file and see if that helps,

[ATB88]

i used your same process with Martyns tool on my car with the modified CSL tune and did not have these issues. i had stock manifolds but in my tune i still had rear o2s and CAT DTCs disabled leaving only the primary o2 sensors enabled and functioning. i ran my CSL airbox wth the CSL tune set exactly how you have it set for over a year with no problems. i also did several other CSL tunes wit that same tune for 3 other local people and had no problems from them either.

the only difference is that I was not using the latest Terra fixed CSL PROG file for reading error codes. i was using the original Paffy mssflasher modified CSL PROG and TUNE files with stock CSL cam offsets.

so theres 2 different variiables between your setup and mine software wise.

- i was using Paffy CSL PRO file not terras fixed error code file
- i was using stock CSL cam offsets


try changing only one thing at a time.

i would start by going back to stock CSL offsets. dont remodify it with tunerpro and tat XDF just go back to a stock paffy modified CSL tune file and then use martyns tool to change it to 6MT and mark the options you want and flash it.


if that doesnt help then flash the paffy CSL PROG file and see if that helps,

Good thoughts, thanks! The first thing I'm going to try tomorrow is another fresh set of O2 sensors with the wires very carefully and deliberately routed with zipties to make sure there's no way they could come into contact with the headers while driving, and use a fresh flash from scratch with the terra modified bin. If that still results in sensor failure, then I'm going to do what you suggested and see what happens. Perhaps it's possible that this is a weird interaction between Terra's modified binary and Martyn's tool, although Martyn has been very helpful in comparing my files to what they should be and hasn't found anything wrong.

[zivagolee]

I kept getting SES due to my O2 wires touching headers (it also didn't exhibit issues right away.. it became an issue at a trackday). Apparently, they are routed differently on EU cars so I had my shop reroute them via the same way they did.