Tailpipe O2 Sensor for 911

[Jay Laifman]

I don't want to weld a bung into my exhaust. Does anyone know if either of these will work well with the 911? I found a comment that the smaller diameter one might not work well at idle. Or any other suggestions? Or even one to borrow in So Cal (Westlake Village-Simi-Oxnard-SFV area in particular)?

[glaverbel911]

It will not work. Your exhaust will cool off to much in the low rpm zones.. For accurate readings you need to be as close as possible too the port. If you can't weld get a shop to do it. Bungs are cheap and weld them into the secondary section of the exchangers.

[Jay Laifman]

That's a bummer. But I appreciate the info. What is the "secondary section of the exchangers"?

Although I recall now a shop who had my car did do O2 readings on my car to set it before. How did they do it without a bung?

And now that I research it a bit more, I note that in the Bosch CIS Workshop Manual for the 911, it actually calls for a CO sensor in the tailpipe. Does that work better in the tailpipe than the O2 sensors? I can't seem to find any CO testers for tailpipes online. But I do see that some think that AFR can be converted to CO.

"CO Adjustment . Make the CO adjustment immediately after a test drive (engine at operating temperature). Install the CO sensor in the tailpipe. Remove the sealing plug from the air flow sensor housing. Using wrench KDEP 1035, turn the CO mixture screw out (counterclockwise) to a value leaner than specification. Then, turn the mixture screw in (clockwise) to bring the CO value into specification. Always adjust from the lean side when setting CO. After each adjustment, remove the wrench and accelerate the engine to stabilize the mixture.Accelerating the engine with the wrench in p/ace could bend the air flow sensor lever."

[NorthernThrux]

CO testers have long probes that can be placed far into the tailpipe. You ideally need to be a foot in from the exhaust tip to get proper gas mixing and not be affected by atmospheric concentrations. Gas analysers also don't care about the temperature of the exhaust while an O2 sensor does need to be hot (>650 F). That's why modern cars include a O2 sensor heater so they can get into closed-loop faster. LM-2 and similar systems don't have that heater and will give false readings at low flow (i.e. low rpm) as the exhaust is too cool at the exit point. You need to be closer to the exhaust ports.

You can fairly accurately approximate CO from AFR. I have this table from when I used to tune my motorcycles.

Ravi

%CO AFR
0.1 14.72
0.2 14.54
0.3 14.42
0.4 14.34
0.5 14.28
0.6 14.23
0.7 14.21
0.8 14.17
0.9 14.15
1 14.11
1.1 14.09
1.2 14.04
1.3 14.01
1.4 13.98
1.5 13.94
1.6 13.89
1.7 13.86
1.8 13.82
1.9 13.8
2 13.77
2.1 13.73
2.2 13.69
2.3 13.63
2.4 13.59
2.5 13.56
2.6 13.54
2.7 13.49
2.8 13.45
2.9 13.41
3 13.38
3.1 13.34
3.2 13.31
3.3 13.27
3.4 13.24
3.5 13.2
3.6 13.15
3.7 13.12
3.8 13.08
3.9 13.03
4 13
4.1 12.96
4.2 12.93
4.3 12.9
4.4 12.86
4.5 12.83
4.6 12.8
4.7 12.75
4.8 12.7
4.9 12.66
5 12.64
5.1 12.59
5.2 12.54
5.3 12.51
5.4 12.46
5.5 12.43
5.6 12.4
5.7 12.37
5.8 12.3
5.9 12.25
6 12.25
6.1 12.22
6.2 12.18
6.3 12.13
6.4 12.1
6.5 12.09
6.6 12.03
6.7 12
6.8 11.96
6.9 11.93
7 11.89
7.1 11.86
7.2 11.82
7.3 11.79
7.4 11.76
7.5 11.72
7.6 11.69
7.7 11.65
7.8 11.61
7.9 11.58
8 11.54
8.1 11.5
8.2 11.46
8.3 11.43
8.4 11.4
8.5 11.36
8.6 11.32
8.7 11.28
8.8 11.25
8.9 11.21
9 11.16
9.1 11.12
9.2 11.08
9.3 11.05
9.4 11.01
9.5 10.97
9.6 10.94
9.7 10.9
9.8 10.86
9.9 10.82
10 10.79

[Jay Laifman]

Thanks! That's a great chart. Except it sounds like if I use one of those exhaust inserts, it won't be accurate at idle. So my only option is to get a bung welded in - which I'm not eager to do.

[NorthernThrux]

A bung and a plug is like 10 bucks. A mobile welder could weld that into your header/SSI at your house for probably 50 bucks. You can put the bung in a spot where it is not visible. It's the only way to accurately figure out what is going on. You could try attach it with a high temperature epoxy instead of welding. Don't blame me if it falls off at some point in the future (easy to patch the hole though!). If you are lucky you can drill the right size hole and thread a bung in if you can find such a thing.

The exhaust inserts won't be accurate even at part throttle. I can hold the exhaust pipe briefly even at part throttle, which means it's just not that hot. And given how important it is to nail the idle values, I think it is worth the investment.

[Jay Laifman]

Would the bung be best in one of the three pipes from the head, or in the single pipe after they join?

[glaverbel911]

Would the bung be best in one of the three pipes from the head, or in the single pipe after they join?

Weld Into the secondary pipe right after the 3 primary’s.
It’ll be a must to tune your crazy lean burning cis engine.

[Jay Laifman]

Thanks again!

[Jay Laifman]

Here is a postscript for any CIS folks who care.

The reason I wanted a CO or O2 tester was to be able to adjust the mixture correctly. Once I got my WUR and Throttle Position Valve working correctly, the car was running nicely, but the idle would sometimes be stable, and other times it would drop and wobble. I found instructions on Pelican to lift the throttle plate while at idle to find the correct setting. I used that method and it cured the idle completely.

As the Pelican instructions say, lightly lifting the throttle plate when the car was set rich caused a noticeable stumble. A light lift after I rotated the mixture to lean caused RPM to increase - I wouldn't call it a surge.

I had also tried adjusting the decel valve to see if that could fix it. It didn't.

Once I leaned it out and got rid of the wobbling idle, I did have to set it toward the rich side to also make it nice and smooth.

Also for further guidance to anyone else, I had to rotate the mixture screw (not the idle bypass) about 3/4 of a rotation to go from the undulating idle to a lean mixture (where the RPM would increase when the throttle plate was lightly lifted). Then I backed off that - so probably 1/2 a turn got rid of the sometimes-undulating idle to a smooth idle.

I will probably adjust 1/8 of a turn each way to see if I get any improvement. But right now the system is working great without any noticeable hitches. When the range of good, between both ends of bad, is less than 1/4 of a turn, I'm happy to play with small changes to see if I can notice anything better. And maybe one day I'll get a bung and see if it can be any better.