Puzzling mixture and O2 meter question:

[sammyg2]

Edit, problem solved, details in the last post on this thread.

Original post:
I just installed an air/fuel ratio gauge and I don't know if I should believe it even though it is supposedly higher quality and not a cheapo. here's the background:

I had the mixture adjusted so that car ran well, then hooked up the O2 sensor to the lambda control. The car started bucking very slightly during cruise. I richened it up and it got worse, leaned it out and it ran smooth and crisp. Also when it was richened up the idle dropped so low engine almost died when I would let off and push in the clutch.

When I hooked up the A/F meter it showed extremely lean. I pulled a couple of plugs and they were white, no hint of tan at all.

I disconnected the O2 sensor and it richened up a but, but still leaner than Stoichiometric on the gauge.

I dcontinued to richen it up a bit at a time, but I had to adjust the idle every time because it would drop. I finally got it rich enough so that the gauge showed stoich all the time and got close to the green side rich) but never made it there even under full throttle.
When it was richened up a bunch (still stoich on the gauge but closer to rich) the reading was very steady, never wavered from partial throttle to full throttle.
As soon as it cools down a bit I'll go read a plug again.

I don't get it. the car runs better when the gauge says it's too lean. I would assume the O2 sensor is bad but the white plugs verify the gauge reading.

Why would the car run better lean, and not want to idle well when it shows rich? Isn't the CIS supposed to richen up the mixture under full throttle? According to the gauge it isn't.

back in my carburetor days we used to adjust the idle mixture for highest idle speed then lean it out until the idle dropped slightly. I guess I expected the same thing here but the engine idles much higher when leaned out, it almost dies when I try to richen it up on the gauge. Any suggestions? I'm contemplating getting a gas tester but it that's money I was saving for other modifications to the car.

Oh, one other thing: i spliced the connection for hte gauge on the wire coming from the O2 sensor. I since heard that that is a no no, i should have spliced in after the connection to the computer since there is something weird about the wire or jacket. Is there anything to that?

[Thack]

"Oh, one other thing: i spliced the connection for hte gauge on the wire coming from the O2 sensor. I since heard that that is a no no, i should have spliced in after the connection to the computer since there is something weird about the wire or jacket. Is there anything to that?"

You can't run one O2 sensor for two devices. You bleed off voltage. Try putting in another O2 sensor. It's not hard but you will have to take it to a muffler shop that can weld in the fitting.

[stlrj]

How do you know your A/F meter is calibrated correctly? How can the plugs tell you anything about the mixture? All my plugs come out white regardless if they came out of my 86 3.2 or my Grand Caravan. Are 911's supposed to have tan plugs because they're air cooled even though they seem to be fuel injected with 02 sensors like any other car?

[ischmitz]

Originally posted by Thack
You can't run one O2 sensor for two devices. You bleed off voltage. Try putting in another O2 sensor. It's not hard but you will have to take it to a muffler shop that can weld in the fitting.

Well, that's not necessarily true. The Oxygen sensor generates a voltage in response to Oxygen in the exhaust gas stream. Anything above 470mV means rich and anything below that means lean. This signal is feed into the DME. The input is low-impedance to be resistant to electric noise.

The A/F meter is most likely a simple voltmeter with a total range of 1.00 volts and fairly high impedance of 10kOhm or more. It won't pull down the voltage of the oxygen sensor noticeably. If it did it would cause the engine to run richer. So I don't think that's Sammy's problem.

I would hook up the car to a good exhaust gas tester to find out what really is going on.

[Thack]

"Anything above 470mV means rich and anything below that means lean...The A/F meter is most likely a simple voltmeter with a total range of 1.00 volts and fairly high impedance of 10kOhm or more"

Exactly. If the meter is using available voltage the O2 sensing circuit may think it's lean. So it richens up the mixture even though it's already perfect. He leans it out and the car adjusts the mixture to ideal stoich. If I understand Sammy and his observations.

"I had the mixture adjusted so that car ran well, then hooked up the O2 sensor to the lambda control. The car started bucking very slightly during cruise. I richened it up and it got worse, leaned it out and it ran smooth and crisp."

[ischmitz]

Thack,

when I read Sammy's post it sounds like when he hooked up the sensor to the lambda control the car got too rich. When he adjusted to more rich he had worse results. He had to lean it out to get it running properly. If the meter would have dropped the O2 voltage the car would have run too lean. Leaning it out more as Sammy did would have made things worse, not better.

You need to think of voltaqe as pressure and ampere as flowrate. A resistor is an orifice. The smaller the orifice the the higher the resistance. The voltmeter has a very internal high resistance and measures the pressure, hence the flow throw it is neglectable. The internal resistance of the oxygen sensor defines by how much the voltage (pressure) drops if a flow (current) takes place. An ideal source with zero internal resistance would have no voltage drop as a result of flow.

To be safe I would disconnect the meter from the sensor and then adjust the mixure with a gas analyser. Once everything is chery watch the gas analyzer and reconnect the meter. If something changes the meter is the cultprit and needs its own O2 sensor. If everything stays the same the meter can run parallel to the DME off the same sensor.

I am wondering if the meter does not require a wide-band O2 sensor to begin with. Maybe that's the reason its reading is wrong on a narrow-bad sensor such as the one found in our 911.

Ingo

[Thack]

Maybe the O2 sensor is weak, not putting out enough voltage regardless.
Then the Lamda Control would think it's lean, correct?
The one thing that is unclear is that Sammy doesn't state how he achieved the good running setup. Did he do it by trial and error or did he use the A/F meter. If he used the A/F meter alone (no lamda control hooked up) then why does the lamda think it is rich when he hooks it back up? And does he have both hooked up while he is reading this lean condition.
To me it sounds like Sammy gets perfect Stoich, then he hooks up lamda control. He experiences rich condition because lamda "sees" lean condition. He leans out the mixture. He hooks up A/F and sees it is lean although car is really running right. He disconnects O2 sensor and the A/F gets more voltage and meters richer than before although not perfect, because he leaned it out earlier. He then slowly enrichens it until the meter reads perfect stoich.
Your electricity knowledge is correct although 10K ohms will bleed off voltage slightly. When your talking .470V as near perfect ANY drop in voltage will affect the lamda control. I know your knowledge is outstanding in this field but I can easilly see Sammy is going in circles here. It makes sense doesn't it?
This is a classic situation of "I do this but my car responds oppositely". Well it sounds like you are confusing your car to me Sammy. Maybe you can put in a switch so you use either the meter or the lamda control but not both at the same time.

Mike

[sammyg2]

I really need to learn to explain things better.

I adjusted the mixture by trial and error with the lambda control disconnected. I fiddled with it until it ran like I thought it should (no cat, turbocharger header took it's place).
Then I hooked up the lambda sensor and I could feel the lambda control making adjustments, swinging back and forth. That's what I meant about bucking. It was very subtle but I could feel a periodic change as the freqency valve changed the mixture. It was probably about every .5 seconds or so. That seems like a slow response to me.

OK, after I hooked up the lambda control and felt the mixture swinging I assumed the mixture was off and the O2 sensor was making large adjustments and cyling so I richened it up and the swings got worse. Light bulb goes off in my head so I adjust it the other way and it smoothed out. I was thinking I got it close enough so that the lambda brain was now making small adjustments. Maybe so, maybe not.

Anywho, I hooked up the gauge with the lambda brain controlling and it indicates extremely lean (.25 volts).
I disconnect the lambda brain and it jumps up to around .35 volts on the gauge. I tried to verify this with my voltmeter but it was too slow to react so that didn't work.

Right now my gut feeling is that the O2 sensor is weak so I snagged a new one off e-bay yesterday. I'll try it and see what happens. That may solve all my problems but just for giggles I bought another A/F meter (evil-bay) to compare it to. I can sell whichever one I decide to not use and recoup most of the investment so not much loss there.

The only thing worse than no data is bad data. It looks like I may have to jump off my wallet and either buy a good gas tester or pay someone to test it for calibration purposes. Oh well, I guess I can live without a new carpet kit for a little while longer.

As far as reading plugs, I used to be pretty good at it back in my hot-rodding days but gas was much different then. I assumed that plugs would still color up so that they could be read after a few miles, maybe things have changed.

[stlrj]

I assumed that plugs would still color up so that they could be read after a few miles, maybe things have changed

Back in ancient times times we fueled our cars with leaded gas that left lead residue on our spark plugs so we could read them and stoichoimetric mixtures were unheard of.

Today, with tightened emission standards together with unleaded gas, white is the color of the day for plugs and anything other than that, you have a problem.

Why would the car run better lean, and not want to idle well when it shows rich?

What your meter reads as lean was problably close to stoichoimetric which is what your engine likes to see and explains why it runs so well at that setting.

It appears your engine is telling you something and your meter is telling you something else. If I were given the choice, I would definitely listen to what my engine was telling me.

Cheers,

Joe

[ischmitz]

Sammy,

good call on the O2 sensor, when they age the first get "slower". They won't respond to changes fast enough. As a result the feedbacl loop (the frequency valve in your case) overcorrects. This is what you might have been experiencing.

To properly read the output of an O2 sensor you'll need an oscilloscope. A DMV is usually too slow and won't revel how long it took the sensor to get to its output. The scope allows you to measure the rise time, when you goose the throttle. I have that stuff and the old Bear exhaust gas tester. Maybe it's time for a roadtrip to Mikey's house, hm? You'll supply the beer, I have the equipment and Mike does the entertaining.......

Ingo

[stlrj]

the O2 sensor, when they age the first get "slower".

An overrich mixture will also make a good 02 sensor appear slow and lower your crosscounts ( the number of times the mixture crosses 450mv in 2.5 seconds) since it takes much longer for the feedback loop to ramp down from too rich to lean in its effort to keep it at stoicoimetric.

Therefore, having your base mixture at stoicoimetric will speed up the process, increase your crosscounts and increase performance at the same time.

Joe

[sammyg2]

Problem solved. If antyone out there decides to hook up an air/fuel ratio meter, make sure you run a seperate dedicated ground wire all the way to the battery. At first I was trying to use the ground wire under the dash that was originally for hte clock. Evidently voltage was bleeding back or wasn't grounding well enough causing the lean indication. After grounding directly to the battery the gauge now reads rich, just as I suspected. Cool, now for some serious tuning and boost cranking

[Joe Bob]

Following ya from board to baord....%^B

RTFB....