O2 sensor confusion

Hey All,

I had to get Tom Celica through emissions a couple of weeks ago. I did, but I got lucky.

I've had an O2 sensor code forever. Finally welded a new bung in to replace the oddly placed O2 mount on my cheap header. I used my Innovate wideband's "analog output" to send to the stock ECU. And it worked. First time I've had a working sensor in that car with the wideband gauge, so I finally got to see what it should be doing. With that info, it became a lot easier to tell when it was working and when it wasn't.

Long story short, it worked, then didn't. Day I went for the test, it started working, so I went for the test. I passed, but leaving the testing place, it stopped working again and it's never worked again since.

Now, when I say "stopped working" I don't mean like a bad connection. It's weird. I can monitor the output of the sensor with my multi meter. At idle, the gauge will show it getting leaner and leaner. The volt meter shows me a constant .01 volts which, as I understand it, should be telling the ECU to richen things up. But it doesn't. Now, if I just unhook the O2 sensor, it runs pretty much as expected. A little rich, but runs well. So I figure that's telling me a couple of things. First, nothing serious wrong with the engine. Second, the connection from the sensor to the ECU is good because it is changing how the engine runs. Although, I assume with the sensor unhooked, it's running in open loop.

Okay, with the sensor hooked up, driving around, it's lean, lean, lean, then boom, rich, rich, rich. No in between, no cycling back and fourth. Makes it hard to drive. Real slow when it's all leaned out then power comes in when it goes rich.

What the heck do you suppose is going on? I'm going to do some more diagnostic work, but really don't know where to go next with it. I want to say the ECU isn't working, but it did work just fine for a bit and got me through emissions. My first thought is to just run a more direct line from the output of the wideband to the input of the ECU. Right now I have a jumper going from the wideband to the stock O2 sensor plug in the engine bay. So there's a lot of wire I can eliminate if I just run it from the gauge to the ECU only about a foot away. Other than that, I don't really know what else to try.

Thanks for any help,

Ed

Aaaaand, just like that, it's working again. Weird.

Somthing in the wiring harnes?

dean1484 wrote: Somthing in the wiring harnes?

Yes. It has to be wiring. I just discovered something.

Went to drive the car back to it's garage (I rent a garage a few blocks away) and it started acting up again about half way there... right when I turned on the headlights. Turned them back off and, sure enough, it was fine. I did it four or five times just to be sure, 'cause it's so weird, but it was totally repeatable.

I'm going to go with bad ground to the gauge, somehow making the "analog" signal dirty. I know that's what it is because it's always the ground. It's funny when I think about it now. I had it working, then that night it didn't work. Then, the next morning, just like that, it worked. Then, late in the day, after the emission test, it didn't work. It was always late afternoon when it quit working. I must have turned on the lights, but I never made the connection before now.