oxygen sensor experience?

[RogerD]

Does anyone have any experience using an O2 sensor to check/adjust mixture on carb'd cars? My 72 spider with dellorto 45s seems to be jetted right but has symptoms of rich mixture (back firing, gas odor in exhaust, sooty plugs). I know I'd need to have an attchment point welded to the exhaust. Any recommendations on products and procedures? Thanks.

[Alfajay]

RogerD:

Yea, I'm working on this too. I may be a couple steps ahead of you. Here's what I have done:

I bought a basic 1-wire narrowband oxygen sensor for $15 or so. Use Google to search for suppliers - I mail ordered mine. I bought a meter in kit form from Jaycar Electronics for $14.95 Australian. Spent another $30 or so having a bung welded into my exhaust pipe just behind where the two branches come together. OK, so for $60+ bucks and a little time I have a "rich-lean" indicator.

What I learned is that my Weber DCOE's just operate at the extremes of the meter. When I tested the meter by connecting it to my 164's O2 sensor, it displayed in the mid-range, so I'm confident it is working OK. But on the carbureted car, it jumps from the richest to the leanest LED based on whether I'm accelerating or cruising. It isn't as simple as that it indicates consistent lean while cruising - it just jumps around.

I believe that these simple meters can be useful on a carburated car, but I also believe that they are intended for vehicles with EFI, and that Webers (or any carbs for that matter) just aren't that precise. It's sort of like using a micrometer to measure wood that you are cutting with an axe - yea, it'll measure it, but does it help you to cut more accurately? I suppose that a more expensive wideband sensor + wider scale meter would be more useful for setting up carburetors.

[Gubi]

I've never done this myself, but my understanding is that you want a wide-band O2 sensor for this sort of thing. If you use a typical narrow-band sensor, once you get off stoichiometric the sensor is going to peg one way or the other which isn't very useful for tuning.

They sell wide-band air/fuel meter kits specifically for this purpose, but having EFI I've never had the need for one.

[alfaparticle]

I am using exactly the same setup as Jay. It has been very useful. My engine has RJ racing cams and 40DCOE116/117 Webers with 34mm venturis(and other improvements). The meter first led me to go to leaner idle jets, then leaner main jets and finally different emulsion tubes. The car is running much better and is getting good gas mileage. My guess is about 150 hp at around 6500, good power to 7000 (my limit), gobs of torque above 4000 and 23-24 mpg. It will not pick up from below about 2200 in 5th.

Mine too tends to jump from very lean to pretty rich, but it shows exactly where the lean and rich spots are which is a big help. I had been running with 135 main jets and it was rich everywhere except at the transition between idle and main circuits. I dropped to 130's and got rid of some richness around 4000 rpm but got a big lean spot at 6000. Dropping air correctors did not have much effect. Then I swapped in F34's instead of F41's. Now it is not too rich anywhere and has only the transition lean spot at about 3500. I am now increasing air correctors to see if I can lean it out a bit at the top.

I have learned a lot about Webers in the past few months with the aid of the mixture meter.

[JRatto]

I have used the NGK wideband AFX sensor with very good results in tuning (and finding cause of lean hesitation/hot start problem) my 48mm Webers. It was very affordable as well. After 17 years of "tuning by ear, nose and throat" this tool makes it so much easier and more accurate. I found, after my fuel delivery problems were solved, that my jet stack was a tad rich, and was able to lean it out, while finding more response from motor.

[RogerD]

Thanks to all for your answers!

[Tifosi]

:2cents: from the peanut gallery:

The things certainly will give you a much better idea of what's going on, and is a lot more accurate (and less time consuming) that doing a bunch of plug chops.

The 'usual' pattern of lights in the gauges is the lean side is form 15:1 down, the center section (also the chunk with the most lights) is the stoich range of roughly 14.7:1 while the rich end of the spectrum starts at about 13:1 and goes to around 11:1 - 10:1.

Of course with narrowband, all that you'll see is voltage #'s rather than ratio #'s, but those voltages translate fairly well to the LED display.

If you go wideband, most all will spit out an actual A/F ratio # on the display so you can get a much better idea of wha's going on. (IIRC, 12.5:1 is supposed to be the sweet spot to aim for AFA performance is concerned)

Regardless of what setup you get, it's worthwhile to spend another few $$ and get a 3 or 4 wire O2 sensor as for (1) they don't cool off while idling, which means you can actually tune the idle mix to avoid over fat or over lean conditions there, and (2) they 'come online' faster than the single wire jobs especially if headers are involved, (under a minute with heated) and in the case of the 4 wire, you have a specific and well secured ground for the sensor as opposed to relying on the sensor body and the exhaust pipes (and thier rubber hangers...) to keep a good circut.

[baldy]

Hi, Ive used one on my sprint to set each cylinder mixture up at idle,and there very good. my car also has mapable ignition so you can see when the main jets start to work at and set timing advance acordingly. mine is the basic LM1 unit. Lee

[ToonRboy]

Sorry for the belated replay. If you were jetted right, you wouldn't be so rich - common problem for folks to install too much carb. Anyway, wideband is probably the best way to get A/F data on your secondary and WOT circuits ('cause you can see while your driving it around). Int'l sells a tool called "Colortune" that is inexpensive and easy to use. No welding a bung, wiring, or any of that.

Hope this helps,

Oh, and by the way, if you can put the car on a dyno (to create load), you can realistically check the accel, mid and WOT (while keeping the car stationary to sight down the device).

[baldy]

Hi, I dont know if you have seen a 16v boxer engine before, but colour tune, is totally unsuitable as the plugs are burried in the heads inbetween the cams,so this would not be practicle. but it may work on an engine with easier access. Lee

[ToonRboy]

Good point - No Alfasud's in the US baldy.