O2 Sensor Behavior, Is it Toast?

Talewinds · Talewinds's Garage

A fellow Pelican and I are attempting to set the baseline AFR using the "ghetto style" O2 sensor measurement of between 0.2-0.6Vdc. This morning after the car had been driven and warmed up we were getting no reliable reading from the O2 sensor's output (the male pin, not the heater pins), like when a meter is open circuited and it's bouncing around on millivolts while it's reading the atmosphere, not a tangible electrical signal.

Fast forward, later at home he connected his meter to the O2 sensor while the car was cold and after starting the car watched the O2 output go from around 0.6v and slowly sink down to 0v gradually as the car warmed up.

What's going on with this O2 sensor?

Wavey · 04-09-2010, 03:28 AM

Rethinking this morning - if the O2 is disconnected while we're doing this, it can't be causing the issue, correct? So it must be a different part causing this? Can we leave the O2 connected and just probe that wire to get a reading?

Thanks for posting this Talewinds!

Wavey · 04-09-2010, 05:15 AM

DUH, Ok you're saying the O2 may be giving us a false reading as soon as it warms up, therefore faulty O2, might get good readings w/a good O2, maybe no other problem at all.

Related question then - if the O2 is providing readings to the DME that are completely out of range, does the DME ignore that signal and go to a default?

brads911sc · brads911sc's Garage

A follow-up question...

Can the O2 be cleaned? I installed a brand new Innovate Wideband with my EFI and it stopped providing a good reading... I pulled it out and cleaned it, put it back in and it worked.. so in the process of tuning (when the car was very rich), it must have gotten some carbon buildup on the holes? does the wire react to a twist? in twisting in the o2 you could have twisted the wire a couple revolutions.. does that matter? I dont have the answers... just have had intermitent false readings so I wanted to offer my experience. When I cleaned It, I just used brake cleaner. Cant say that this is proper or that it is good for it... If I keep having issues, Ill buy a new one..

don gilbert · don gilbert's Garage

Im not sure about Porsche, but I'm sure it is the same sevice interval as BMW in 88, manufactures sevice interval is 30,000 miles. As the 02 gets older, they work slower, because of the soot build up. I think in 88 it is still a one wire? If it is, you can get a Bosche 1 wire for 20 bucks at O'Riellys. I have cleaned them before, but not with satisfactory results, When I was at BMW, I had access to factory diagnostic equip, and could watch 02 v. output on a monitor, and cleaning never produced the same quick voltage fluctuations as a new one.

Wavey · 04-09-2010, 05:46 AM

This particular sensor only has about 4K on it but has been running in a very rich environment. The '88 is heated therefore 3-wire.

brads911sc · brads911sc's Garage

might want to clean it good before you replace it...

Wavey · 04-09-2010, 05:59 AM

I think I have a good used one in the cat that came off the car, not too many miles on it - I think we'll test w/that next.

Talewinds · Talewinds's Garage

Quote:

Originally Posted by Wavey

Related question then - if the O2 is providing readings to the DME that are completely out of range, does the DME ignore that signal and go to a default?

I have to assume, although I don't know the technical spec of the DME, that at some point the DME defaults into a mode that is no longer dependent on input from the O2, just like when you unhook the O2 sensor. And, the DME works to lean out the mixture, if you're running lean already the DME can do nothing to richen it.....

Wait, I found it in the "Idle Bounce" thread...

Quote:

Originally Posted by Scarceller

You should always set on rich side 14.2 because the O2 sensor circuit is really designed to lean out mixture not richen it. Also the spec says 0.8% CO which is equal to 14.2AFR. I'd use the 14.2 or even 14.0 as the base setting then let the O2 sensor pull it lean when in closed loop. I also found most of these cars like to idle better when on the rich side with the base mixture.

Then the other school of thought is that if the O2 fails they usually go open circuit and this is like not having an O2 sensor installed so to be on the safe side you rather run rich than lean.

Brad and Don, I was getting slow/intermittent response from my O2 sensor last year so I tried the heat/ cold water method of flaking off the carbon buildup, which worked nicely, but it killed the O2 sensor in the process! Ooops. Then when I bought a new O2 sensor I was getting dead-on reliable readings from it and the voltmeter. I am very suspect of the way Wavey's O2 sensor was behaving yesterday.

rusnak · 04-09-2010, 08:44 AM

Yeah...O2 sensors do not like water.

As an aside, when installing and removing the O2 sensor, make sure it is disconnected on the plug side first so you don't twist the wires.

RWebb · RWebb's Garage

btw - are you using an analog meter to check it?

Wavey · 04-09-2010, 11:53 AM

No, digital in both cases. But the reading are so far off the charts it probably doesn't make any difference. Prepping for a DE this weekend, will get back on it next week.

Talewinds · Talewinds's Garage

Quote:

Originally Posted by RWebb

btw - are you using an analog meter to check it?

Digital. It's a very high end meter with a digital/ analog scale for visual purposes.
It's given me very good results on an O2 in the past.

I know what you're getting at with the digital vs. analog meter though, as my trade was industrial electricity/electronics.

RWebb · RWebb's Garage

ok - sounds like the thing to do is to just replace it - the $$ are likely < than the hassle of debugging...

ischmitz · ischmitz's Garage

The 3.2 DME simply compares the O2 input signal against an internal reference voltage. For the time the sensor voltage is below the reference voltage (indicates lean mixture) the mixture trim value increases with a given rate (rate depends on RPM) and for the time the value is above the reference voltage (rich mixture) the mixture trim value decreases with the same rate.

The rate of change is about 5% per second at 3000 RPM. The trim value can modify the mixture by plus or minus 20% in part-throttle. If the O2 input voltage is zero all the time (sensor damaged or not connected) the mixture trim will simply walk all the way towards richening the mixture by 20%.

So the bottom line is without the sensor connected you run rich in your part-throttle range. Newer DME's from the 964 onwards detect a damaged or missing sensor and keep the mixture trim a zero.

Ingo

Talewinds · Talewinds's Garage

Quote:

Originally Posted by ischmitz

The 3.2 DME simply compares the O2 input signal against an internal reference voltage. For the time the sensor voltage is below the reference voltage (indicates lean mixture) the mixture trim value increases with a given rate (rate depends on RPM) and for the time the value is above the reference voltage (rich mixture) the mixture trim value decreases with the same rate.

The rate of change is about 5% per second at 3000 RPM. The trim value can modify the mixture by plus or minus 20% in part-throttle. If the O2 input voltage is zero all the time (sensor damaged or not connected) the mixture trim will simply walk all the way towards richening the mixture by 20%.

So the bottom line is without the sensor connected you run rich in your part-throttle range. Newer DME's from the 964 onwards detect a damaged or missing sensor and keep the mixture trim a zero.

Ingo

Now THAT'S good stuff! Thanks for your input!

Wavey · 04-09-2010, 02:18 PM

Quote:

Originally Posted by Talewinds

Now THAT'S good stuff! Thanks for your input!

No kidding! Thanks Ingo!

Randy - I have a known "ok" used sensor, so we'll screw that in and get a reading right there. No problem getting a new one, just want to be sure that's what's causing this rich condition. There's a lot more to the story, started with a dyno run showing a very rich condition, but from Ingo's info it certainly could be a bad sensor. But the thing only has about 4K miles on it, so ???

don gilbert · don gilbert's Garage

I dont know where you are, but you don't necc. need a heated 02 if you are in the southern region. (it will just stay in open loop for a few min. longer.) I have put the one wire on lots of gm cars w/heated 02,(some have four) works fine for 20 bucks, unless you just want to shell out the cash for the heated one. (or ones)

Wavey · 04-09-2010, 02:46 PM

I've been wondering about that Don. When I fired this up cold, later yesterday, I got a slightly high, .69? normal reading at first, then it steadily dropped below .00 over about 15 seconds. Wondering exactly what the purpose of the heating is? I do drive it year-round, sometimes at pretty cold temps.

don gilbert · don gilbert's Garage

Strictly emission reasons. Just gets it into closed loop faster, esp. for cars in cold climates. I use a heated one on my sc engine with cis. only because of have SSI and they let the 02 cool down real fast between short stops. I like closed loop as fast as poss. when im just putting around. (fuel mileage) On a properly tuned cis car, you can hear and feel the gentle fluctuation of the lambda control in action, not so much on motronic though.