1MeanZ

iTrader: (9)

I've searched and not had luck so I'm going to break down and ask for help at this point.

I've got an 88 MAF car I'm fighting with. When its in open loop or highway mode, the WBO2 is steady. For example highway mode commanded AFR is 16.2, and it will stay there +/- .3 or so. it holds about the same range when it's cold running in open loop. As soon as it warms up and goes into closed loop it swings wildly all over the place from 12-17. It's centered around stoich, but it varys wildly. Amazingly the engine doesn't behave that poorly while this is happening, but a discerning butt can feel it.

I even went so far as to copy all the tables (minus timing and maf tables) from my GTA that runs and drives beautifully and it made no difference, the O2 still swings wildly all over the place. My IROC doesn't do this either, nor has any other MAF car I've tuned to date. I'm at a loss at this point.

I've gotten to the point I'm actually wondering if the ECM is flaky, and I'm usually the last guy to blame the ECM. This morning pulling in to work the car randomly wanted to idle at 9:1 afr. I let it idle pig rich and started checking underhood for obvious problems, none were found. By the time I got the laptop fired up the car had straightened up and nothing on the datalog looked out of place. At lunch I had to open the throttle about 25% to get it to start but then it idled like usual. On the way home it ran ok till it went into closed loop, at which point the AFR was stuck in the low 16s. Then after it went into highway mode it straightened back out and ran like normal with the fluctuating AFR. Looking at the datastream while driving the INT is just all over the place. I swapped in a known good '165 ECM and PROM and it seemed a lot better, but I just can't believe that the ECM is bad....

I've not SD swapped any of my cars, or tried the EBL, but I'm getting frustrated....

1MeanZ

iTrader: (9)

Updates after tinkering with it yesterday and this morning.

I swapped in the ECM and MEMCAL from one of my other cars and re-used the PROM I was using. Closed loop WB readings were more stable, with only the occasional flash way rich or way lean, more often way lean.

So then I reduced the Upper O2 threshold by 10mv, raised the lower O2 threshold by 10mv. I also reduced the whole Integrator Delay Multiplier vs Error table down to 25 from 99. Car acts the same except now it goes lean when I first roll in to about half throttle, if you stay in it after about a 2 count it comes back to stoich, so I suspect I shot myself in the foot by slowing the Integrator down that much.

Final observation, the first time the car went into highway mode this driving cycle, the wideband maintained its dance while the commanded AFR was 16.2 for a minute or more. Then all of a sudden while still in highway mode it leaned out and stabilized like it should. BLM at highway speed was 122-128 today.

MrWillys

I would say what you are seeing is closed operation of crossing over from rich to lean in it's design. This is normal.

1MeanZ

iTrader: (9)

LOL, I've been tuning these cars for 10 years now. I've lost track of how many I've done custom tunes for. I'd like to think I'm well aware of what is normal and what isn't, my other 2 cars don't behave this way, nor have any of my customers complained of this with any of my tunes.

MrWillys

My apologies but your explanation sounded like crossing over. Let us know if you figure it out?

1MeanZ

iTrader: (9)

no worries man. This thing crosses over to vast extremes very quickly, not normal behavior compared to my other cars.

RBob

iTrader: (1)

It is likely that the ECM is bad. This is how the '7165s fail, they get intermittent. For further testing drop the ECM down (leave connected) and remove the upper case. With the engine running gently press on various areas of the ECM board.

Note if the engine changes in how it is running. You may find a spot that when pressed the engine runs great. With other areas being pressed the engine goes away.

RBob.

1MeanZ

iTrader: (9)

Thanks for the response RBob. I feel like I'm to the point of needing help from guys of your caliber, this is the first time I've been completely stumped.

Updates: I swapped ECMs from my red IROC to the black one. The red IROC runs perfectly with the suspect ECM, just like it always has. The black IROC runs exactly as it has with the known good ECM.

In post #1 I mentioned that I took my GTA's tune and just changed the fuel and timing tables looking for a solution. Then in post #2 I mentioned that I was slowing down the integrator reaction and changing the O2 thresholds. This began having various undesirable affects such as lean PE, seemingly unresponsive AE adjustments, and didn't really get rid of the problem. Suspecting a corrupt .bin, I got rid of the modified GTA tune and went back to one of my previous known good tunes and all the undesirable side affects went away, PE and AE came right back in line, but I still have the wildly swinging O2 values at part throttle in closed loop.

So at this point I don't really suspect the ECM anymore. I think I had a corrupt .bin complicating things during my diagnosis(I've only ever had that happen one other time, not sure how they get corrupt). On my way home tonight I'll datalog the injector pulsewidth and see what it shows. IF the O2 reading didn't stabilize so nicely in open loop and in highway mode I'd suspect the O2 sensor. I'll also double check the TPS voltages, coolant temp readings and MAF output.

I just don't know what data input or mechanical problem would cause the ECM to vary the mixture so wildly. The one thing that keeps sticking in my head is the fact that a previous mechanic spliced one of the fuel injector wires with crimp on butt connectors. One of the injectors is really hard to get to with the super ram intake, and for some reason it looks like they either lengthened the wire or had to replace the injector plug. If that connection was faulty, I'm not sure what affect if any that would have on the ECM, seems like it would cause problems all the time.

RBob

iTrader: (1)

In open loop the MAF and CTS have the biggest affect on injector PW. The MAF measures airflow to calculate the injector PW. While the CTS affects what the commanded AFR is.

Once in closed loop the above is done with further adjustments made due to O2 sensor feedback. Can check the INT & BLM for the amount of adjustment.

RBob.

Larry

I'm thinking ignition problem causing intermittent misfire. Or if an injector didn't pulse it would cause a lean spike too. If you have access to an ignition scope it should tell you. A misfire will will show up as a lean spike even when its not lean and add fuel making it rich. I've been chasing one of those for a while. Intermittents suck.

1MeanZ

iTrader: (9)

I'll dig deeper into my timing map and make sure I'm not ramping in too much timing at light load where it's acting up. The distributor/cap/rotor and plug wires are one year old, the plugs were new in the spring and it has a brand new MSD 6AL box on it.

1MeanZ

iTrader: (9)

Also, I have an MSD ignition coil mounted on the firewall above the heater box, the main ECM wire harness runs directly over top of it. Anyone think that could be part of the issue? How much EMI does a coil put off?

RBob

iTrader: (1)

Um, a lot, lot? (sorry)

RBob.

1MeanZ

iTrader: (9)

So should I attempt to relocate the coil or should I shield the harness with copper tape? What are the odds that this could be the issue? I'll stop further diagnosis until I've relocated the coil or shielded the harness.

RBob

iTrader: (1)

I would relocate the coil. Every time it fires a spark plug the magnetic field that is built up during dwell collapses. Inducing a high secondary voltage that then goes on to the spark plug.

RBob.

Larry

Interesting. Let us know if it makes a difference.

1MeanZ

iTrader: (9)

so for a brief test, it was much easier to temporarily re-locate the ECM harness by placing it in front of the dipstick tube at least 12" away from the coil and coil wire. The relocated harness made absolutely no difference, I pulled over, re-clipped the ECM harness to the firewall near the coil and drove back home, again no change in behavior.

I can graph the O2 voltage from the ECM's datastream and compare it to what I see on the wideband gage. On a flat highway with the cruise set, the indicated AFR was swinging from low 12s to about 18, the narrowband datastream seemed to mirror that with millivolt readings swinging all over the graph from over 950mv down to below 100mv. Once highway mode kicked in, gage shows 16.2 and the O2 datastream tightened right up as expected.

So at this point I am confident that the readings I see on the gage, matches the data the ECM is seeing. I've swapped ECMs and relocated the wiring harness to no effect. The O2 is steady any time the ECM is in open loop (cold, WOT, Hwy mode).

I then doubled the INT delay and halved the INT % correction and it made no difference. Car ran and behaved exactly the same.

So at this point I'm completely stumped. I feel like I've pretty confidently eliminated the harness location, ECM and the O2 sensor and controller. I appears at this point that the ECM is causing the O2 swings that I'm seeing, and it only does it when it's looking at the O2 for correction.

RBob

iTrader: (1)

Try another NB O2 sensor, the ECM changes the INT/BLM based on feedback from it. Also check the O2 sensor ground lead from the engine to the ECM. It is an isolated ground for the O2 sensor amplifier in the ECM.

So both the O2 signal and the O2 ground have to be good.

RBob.

1MeanZ

iTrader: (9)

I'm 100% on board with checking the ECM grounds, specifically the one for the O2. Does the diagram in the GM service manual say which ECM pin is the ground for the O2 amplifier? D3 or D6? I assume it's one of the grounds on the back of the cylinder heads? (I don't have my manual in front of me at the moment, I'll double check when I begin troubleshooting)

I'll also double check where I have the ground for the WBO2 controller.

You mentioned trying another NBO2? I don't have an NBO2 in the car at all at this point. I can put one back in for testing, but right now the ECM is being fed the channel 2 NB output from my Innovate controller.

RBob

iTrader: (1)

The O2 ground is ECM pin D6. And does go to the cylinder head ground lug point.

I didn't realize that you were using a simulated NB output of the WB controller. If it is easy to put a NB O2 in for testing it is likely worth it.

RBob.

1MeanZ

iTrader: (9)

I was concerned about the O2 controller outputs not matching. In other words, what if the gage was showing something different than what the ECM was being fed?

This past weekend I was datalogging while driving. While connected through the ALDL, I graphed the o2 data the ECM was seeing. It seems to mirror what the gage shows. I had the cruise set and when the gage showed erratic O2 readings so did the ECM, then when highway mode was activated, the gage sat on 16.2 and the ECM reflected that as well.

For whatever it's worth, I have the exact same Innovate WBO2 sensors on my other 2 cars. I've been feeding simulated NB data to '165 ECMs for 10yrs on one car and 5yrs on the other and they've never had an issue.

I'll thoroughly check the grounds for the ECM and the wideband and report back. Thanks for the help RBob.

1MeanZ

iTrader: (9)

I immediately discovered that the ground wires that are supposed to be on the back of the passenger side head were spliced into with crimped on butt connectors and screwed to the firewall. I'm not sure how I previously missed this before. I cut the connectors off, soldered new wire on with proper ring terminals and bolted the grounds to the back of the head where they belong. I then checked resistance from ECM terminal D3 and D6 to the passenger cylinder head. I had about .6 ohms from the ECM plug to the bolt on the cylinder head.

I drove the car after this and no change, still get the wild O2 readings. I narrowed the O2 R/L upper and lower thresholds by .010mv just to see what it would do and it appears to have had no affect.

I noticed during datalogging that sometimes when the ECM goes into highway mode and calls for 16.2 AFR, the readings on the wideband don't change. It's as if once the criteria for highway mode are met, the commanded AFR changes, but the ECM doesn't respond, or doesn't appear to by the O2 readings. After I've stopped or slowed down and come out of highway mode one time, it usually goes back in the next time the criteria are met.

84Elky

As you've stated, it's frustrating and something does not make sense. Here's why. From a code viewpoint, you are guaranteed to have entered into Highway fuel mode 10 seconds after all criteria have been met. It is then the commanded AFR is recalculated to a value higher than the Calibration Stoich AFR value based on Calibration Hwy fuel table values and Stoich AFR. This causes the BPW to be reduced due to the higher highway AFR which should immediately be reflected on the WBo2. It appears you see this at times, and sometimes not. So seems to be signal and not actual fuel related unless by coincidence the recalculated Highway-AFR-based BPW is identical to that provided by Stoich AFR, which is unlikely. It seems you've changed and swapped many things. How about the WB controller since that is providing the signals? Just a thought.