Hondata

OK, here's the breakdown....

About 3 months ago, my car started creeping lean at idle every once in a while. First, I cleaned out the IACV and FITV (on my '94 B18B Integra Non-VTEC).

It progressively got worse to the point that I can't really drive the car in partial throttle without going lean, regardless of the entries in the fuel tables. I initially thought it was a fuel delivery problem, so i replaced the fuel filter, injectors, pump, and FPR. Still the same issues.

Someone then suggested I look at the TPS thinking that perhaps the car thought it was at idle when in fact the throttle was open. the sensor tested fine for voltage and changed linearly with angle with no distortion.

I then suspected the O2 sensor and this is where I think at least part of the problem lies. First of all, I can only get the car to idle decent in open loop. If I do closed loop, I get a hunting idle. I took a look at the offset voltage between the wideband voltage output and the interpreted A/F value in Hondata and it appeared to be off. So, I adjusted it, got them to sync up and then a few minutes later, they were mismatched again. It is as if there is some sort of electrical interference or shorting going on causing the narrowband output of the wideband to go nuts.

In addition to this idle issue, when I accelerate and let off the gas, such that the car is in columns 4 and 5 of the MAP, it goes lean. I increased fuel here to no avail as well.

s3d's are posted. Thanks for any input.

What wideband are you using and where is the calibration?

I am using a PLX M300 wideband, currently using the narrowband output (gray wire) as the closed loop input for the ECU. The wideband output (white wire) is running to the D6 (VTEC pressure) on my LS. The calibration will be up shortly. Thanks.

You cannot use vtec pressure as an input for the wideband signal. You must use one of the ones listed in the software. Vtec pressure is not an analog input, it detects an on or off signal.

calibration posted below...

Spunkster wrote:You cannot use vtec pressure as an input for the wideband signal. You must use one of the ones listed in the software. Vtec pressure is not an analog input, it detects an on or off signal.

I lied and spaced out. I use D6 for full throttle shift. I use ELD for wideband in. Sorry for the confusion.

However, the fact remains that I am still having issues at low throttle. I am having a second O2 bung welded onto my exhaust so I can run the OEM narrowband and still watch the AFR on the wideband. This way, I'll be able to see if the wideband controller is bad.

I also noticed a slight leak in the fuel send line at the pump. I have since replaced the crush washers and will find out tonight if they will do the job of sealing.

It may be that the wideband sensor is dying and you need to replace it. They typically read lean when they die.

granted, however, would that explain running lean in open loop?? could there be some screwed up wiring that somehow affects say the 12V switched supply to the fuel pump or something??

I have received a replacement wideband O2 and will be re-installing my OEM narrowband as well to try to determine if the O2 sensor is bad or if the controller is bad.

if the sensor is dying and thus reading lean it won't matter if you are in open loop, closed loop, or on the loop de loop at Six Flags.

perhaps i'm not being clear. open loop means that the O2 sensor is ignored (thus long term and short term fuel trims ignored) and the AFR is purely dictated by the fuel vs. pressure map and various correction factors, right??

So, my question is this: is a dying O2 going to actually make my car run lean in open loop regardless of what the display says? like if it was tuned before and ran fine in open loop, would there be any reason for the dying O2 to cause an actual lean condition in open loop (detectable by the car actually losing power and getting some minor detonation etc regardless of what is displayed on the PLX screen and what Hondata interprets from the O2)??

but how do you know you are running lean if your wideband is not reading correctly?

There are too many factors here to say what is actually happening. Have you done a leak down and compression test? It's possible that you are loosing fuel pressure, have a clogged fuel filter or have bad injectors. If you suspect the 02 being bad, then you have no way of being sure you are actually running lean.

Based off your innitial post you said it was fine in open loop.

yes, i agree. however, i have been able to notice the problem at low throttle in open loop since that first post.

compression and leak down is on my list of things to do and i also highly doubt the injectors are at fault since i just put in the sonic cleaned ones from RC and experienced the same problems. i also replaced the fuel filter 20k ago, so i doubt that's at fault.

i will go ahead and replace the O2 sensors as I stated earlier and see if the issues come back. thanks again for the feedback and recommendations.

OK, so I think I have determined the sources of the problems. One of them is definitely the IACV and there is also definitely something going on between the wideband controller and Hondata.

I hooked up my OEM narrowband O2 and used that as the narrowband input to Hondata while I still had the wideband hooked up for monitoring A/F on the gauge and tuning (which i didn't really mess with this time around).

I noticed that the idle was a lot more controlled with the narrowband O2 (and it got even better when i bumped up the IAC duty cycle) than the narrowband output from the wideband. I also replaced the wideband O2 with a new one, which rules the sensor itself out as the problem.

I believe that whatever is causing this mismatched voltage to AFR interpretation in Hondata is being caused by faulty circuitry in the PLX controller or perhaps some kind of electrical interference at the ECU. However, I do trust the AFR's displayed on the PLX controller screen and have been using that as a measure of the AFR's instead of those datalogged in Hondata.

My updated calibration is posted below. The datalog is apparently too large for this site, so i'll have to get a shorter one later. I am curious to hear your thoughts.