Warm idle Lambda AFR
 

Engine is warm after 20 minute drive.

First pic. O2 sensor unplugged, FV defaults to 50% as it should, AFR set to about 13.6 as seen here:

http://forums.pelicanparts.com/uploa...1744930436.jpg

Second pic. O2 sensor plugged in. FV previously set by me at 40-45, AFR drives leaner to about 15.8 as seen here:

http://forums.pelicanparts.com/uploa...1744930502.jpg

Is this normal lambda controlled AFR behavior for warm idle?

Secondly, if FV number decreases from 50 to 40, shouldn’t that richen the AFR?

Thanks

Anyone else with stock lambda and CIS have 15.7 AFR at idle?

Thanks

Where is that AFR gauge connected to, the narrowband 0-1v signal?
I guess not as with a narrowband signal it can't read steady AFR 16:1.
So are you using a wideband controller? and if yes, ... what resulting voltage of the controller equals Lambda 1 / AFR 14.7:1

1. If you're using the stock eco, ... use the orig narrowband sensor and read the voltage of the green wire on port 2 of the ECU, ... does it swing between 0.1-0.9v?
If yes then you're spot on Lambda 1

2. Is your exhaust system stock as it left the factory? Cause removed Cat or SSIs do affect the resulting sensor signal.

3. I would not focus on AFR, but on lambda as that is the native reading of the sensor. AFR is just "calculated" from the sensors lambda reading and also depends on the ethanol content. Be aware that with 5% ethanol at lambda 1 the AFR equals to 14.4:1.

I have an SC converted to EFI so I have played a lot with AFR in all conditions. I can tell you that these cars will not be happy idling that lean (15.7). I have mine set at 14 and it idles well at 13.6ish.

While Andrew is correct that a stoichiometric mixture of gas with 10% ethanol has a different actual AFR value, that will be irrelevant since you can just read the AFR as if you had 100% gas with 14.7 being stoich.

Seems his car ist still running the original lambda based k-jetronic and with these models idling at Lambda 1 / AFR 14.7:1 is still ok and smooth. The only drawback with the original ECU is, that idling oscillates a bit within around 50 rpm. As with these old ECUs the 0.5 to 1 kHz swinging narrowband signal will be 1:1 forwarded to the ECUs duty cycle generation.

Original lambda system intact, new O2 sensor, new O2 relay. At cruise throttle AFR waivers around 14.3 goes to about 13.5 at WOT.

I do have SSI exhaust but surely it would not make that much difference at idle. JBell custom AFR gauge that I sent back to him to verify it is calibrated correctly

14.3 at cruise and 13.5 at WOT seems in the ball park but 15.8 at idle seems out of whack with lambda hooked up. I can super richen it by turning the mixture screw to force the system to 14.4 at idle, but then it will be so rich it surges badly and won’t start (too rich).

Still not answered the question:
What provides the voltage to that individual made AFR gauge?
Is it the original narrowband signal (where here I think it's not) or a voltage signal from a wideband controller? If yes, then in the gauge at what voltage is AFR 14.7:1 (Lambda 1) present ??
I ask cause ... maybe the gauge expects a different voltage which equals Lambda 1 / AFR 14.7:1.
So ... 2.35, 2.45, 2.5 or even higher volts.

No you can't, you did not understand what the lambda control of the ECU does --> no matter if you turn the CO screw for enrichment, at the end with closed loop control the ECU forces the Mixture to result in Lambda 1. So if you enrichen the mixture with the CO screw the ECU simply will decrease the duty cycle to make it result in Lambda 1 again.

Surely? :-)
It does make a difference. Cause the sensor in an SSI system at idle will result significantly colder than with its position in the original exhaust system, means much nearer to the exhaust port of the head and especially in front of the CAT which will make the sensor resulting much hotter at idle and therefore running more precise. And I guess you deactivated the ignition retard function at idle by disconnecting the second hose at the back side of the dizzy can? If yes, then your sensor will result even colder, cause the original retarded ignition timing with 5° ATDC at idle will let the exhaust flow result hotter at idle compared to the regular ignition timing using 5° BTDC.

Yes thats common, ... the original ECU only rises the duty cycle to static 65% at WOT, which does not result in the needed AFR 12.5:1 (lambda 0.85) for optimal combustion power, but in 13.5:1. Thats due to keep over all emissions within the limits in these years.

It’s an AEM wideband controller. Used by many others here over many years in conjunction with SSI exhaust. What’s a bit different about mine is the JBell gauge, but he has made dozens of them for many here that use the same controller.

Are you saying 15.8 could be normal idle reading with SSI’s?

Out of town for the weekend. When I get back I will take a cold start video to get your impressions.

If my FV is set to 45% with O2 connected, should my AFR gauge read 14.7 at idle?

Thanks

Do all These users drive that controller in combination with that gauge in a lambda controlled k-jet car? As said, its all about the voltage wich equals Lambda 1, if the controller delivers a voltage where the gauge expects a different one for AFR 14,7:1 then ... the reading won't be correct, ... but I just assume, ... as ...the orig ECU with original exhaust system should let result a reading of lambda 1 at idle and normal driving. If not, then oxygen fluctuations in the exhaust flow could make readings get messed up, or the position of the sensor does affect its resulting voltage.

No, ... in my case with SSIs at idle the duty cycle rises more to match lambda 1, means the sensor reads to lean. In your case it's the opposite, ... the final lambda 1 reading is a bit too lean.

So, ... check the sensor, but it's a new one, as you said, and here ... is that new one even faulty as well? IN times of today sometimes with new spare parts you get even more into more trouble.
Or, ... the ECU could be faulty, targeting a wrong value.

Hard to figure out. Been through at least 4 O2 sensors, 3 O2 relays and I have a spare O2 ECU that reads exactly the same. Even changed the O2 sensor-to-ecu connector.

The AFR makes sense in open loop ie, WOT or before O2 kicks in. In closed loop too lean.

Another example. Using Gunson Gas Tester I set the CO to 1.5 on Wednesday. This should be a bit rich for my car. Plugged in the O2 sensor AFR went straight to 15.8

The ecu does not know what exhaust I have or where the O2 is plugged in or whether I have unplugged the ignition distributor retard hose. All it knows is once it receives O2 sensor input is to vary the FV to achieve 14.7 AFR. Mine is either doing that and the indicator is wrong, or the gauge is right and the ecu is seeking the wrong value.

I did send the gauge back to John Bell and he double checked the gauge was calibrated properly for my AEM controller.

I had a hard time getting the gauge wired up. It is a split voltmeter/AFR gauge (see pics above). It’s been a couple of years so I will go back and check my grounds and verify what wire goes where. But as I said, in open loop it looks to be working properly.

Lastly, I recently bought a used Lambda ECU wiring harness from another Pelican. It’s possible I could have a short or broken wire somewhere but I am grasping at straws there. It’s literally the only thing I have not changed.

Thanks for everyones input. Always appreciated.

With hot engine and lambda system operating in closed loop state, ... now at idle turn the CO screw in a way so the duty cycle swings around 50 and 55 %.

With that setting what AFR is shown on your AFR gauge at a) idle? and what AFR at b) WOT / High load

Thought too simply ... The ECU itself surely not, but the oxygen amount in that area of the exhaust system where the sensor is installed to can be affected by several things, as described. Like air leaks in the exhaust system etc.. And that "will" affect the sensors reading, .. as well as a non proper operating temperature of the sensor, even with a heated wideband model. Cause heater of heated sensors do help for faster self initializing, but the target operating temp will be achieved and kept by the exhaust flow temperature.

Is this possible? I set AFR to 14.7 at warm idle and FV reads 11%?
Thought I read the FV range is 20% to 70%.

http://forums.pelicanparts.com/uploa...1745379123.jpg

http://forums.pelicanparts.com/uploa...1745379314.jpg

The range of the original ECU is from 5 to 95%.

At open loop the initial set up of the CO screw with the original ECU should result 0.4-0.8% CO, then (normally) when changing back to closed loop the Duty Cycle rises a bit till it swings around 55%.

Your workflow above shows clearly that the sensor reads a value which does let the ECU think its too rich for lambda 1 (AFR 14.7:1) and so it pulls the duty cycle down.

How is your setup in detail?
- Does the AEM Wideband controller –beside the signal for the gauge– provide also a simulated narrowband signal (Like Innovate controller do) which you use as signal for the ECU?
- Or do you use two sensors, one original sensor with signal for the ecu and another wideband with signal for the AEM controller?

This is how it is set up.

So I can set mixture in closed loop with AFR or FV but they do not agree.

Well, ... then one of the units (ECU vs AEM) does interprete the sensor differently or one of the sensors give different signals at lambda 1.
Or ... if the sensors are installed one on the left side and the other on the right side, then maybe the mixtures/combustions are different on each side, ... but then the reading of the plugs would show it clearly

So ... without going more and more into theoretics ....go to an official emissions check point or service garage with testing equipment and let an official calibrated device do the test. There you can see which one is on spot, .. the ECU or your AEM Controller.

BR

Do you have both the OXS ECU and the gauge hooked up to the same O2 sensor, or separate sensors? If separate, you may be getting different readings on one bank of cylinders versus the other bank, due to uneven fuel distribution, or intake leaks, or exhaust gasket leaks.

I would also put a digital voltmeter on the output of the O2 sensor to see whether your gauge is calibrated. At Lambda=1, it should be about .45 volts.

When the O2 sensor is unplugged, the FV goes to 50% as it should. I also have a spare
ECU which has exactly the same readings. Both of these lead me to think the ECU is OK

What do you think of these other readings? All normal? Should Vmax be 6.48V be higher?
http://forums.pelicanparts.com/uploa...1745444016.jpg

Thanks

They are on separate banks. I have recently warmed up the car, unplugged O2 sensor so that FV goes to 50% and noted the O2 reading. Then shut off the car and removed the OE O2 sensor and put the AEM sensor in the OE cavity and restarted the car. Difference was only .4 AFR right side was higher. Meaning that I can look at the AFR gauge and subtract .4 to get what the OE sensor is outputting.

The difference here is much greater than that. I can set mixture in closed loop with FV or AFR gauge. If I use the FV to set mixture at between 45%-55% the AFR at warm idle reads 16.8% (see pic above) If I set AFR on the gauge to 14.7% (then subtract .4 to get 14.3% actual) FV reads 11.6%.

“I would also put a digital voltmeter on the output of the O2 sensor to see whether your gauge is calibrated. At Lambda=1, it should be about .45 volts.”

What is the best way to do this? Back pin the O2 connector (?) or is there a better place to check it with the car running?

Thanks

I already know that both of your original ECUs do give same readings.
But as already mentioned above -> your original ECU(s) and the AEM unit both do read the signal of their sensors differently.

The ECU results in Lambda 1, and the AEM controller system says it's a way off.
To check if the ECU really results in lambda 1, follow PeteKz's advice and with your oscilloscope read pin 2 at the ECU if the voltage signal swings between 0.1 to 0.8 volts – if that is the case, then your ECU behaves as it should and you're on spot with Lambda 1.
Then you should ask yourself why "on the other side" of the engine, the reading is different than Lambda 1, and here the options are a) AEM readings are off, or b) the mixture is different than on the other side of the engine.

As its the test port output, for me its ok

https://youtube.com/shorts/AP-rX-AY0SE?si=NcbrvHMg7vEfM43e

Looks ok

b) From my previous post above:
“They are on separate banks. I have recently warmed up the car, unplugged O2 sensor so that FV goes to 50% and noted the O2 reading. Then shut off the car and removed the OE O2 sensor and put the AEM sensor in the OE cavity and restarted the car. Difference was only .4 AFR right side was higher.” I believe mixtures are fairly close between banks.

This leaves me with a). Currently looking for a used AEM 3.14” AFR gauge I can use to verify the custom gauge.

In the mean time I will set the mixture using the FV setting on the oscilliscope.

Thanks

Go to a service garage or emission check station where an official calibrated testing device gives you a credible reading

Looks ok

Florida stopped emissions testing in 2000. Really hard to find a shop to do that.

I have a Gunson and can set CO .8 myself.

I can tell you when I set mixture using AFR of 14.7 at warm idle, at WOT (closed loop) the AFR drops to about 12 and the car TAKES OFF. Like a turbo (admission: I’ve never driven a 930). If I set AFR with FV of 45% or use CO .8, WOT AFR drops to about 13.5 which is not as exciting.

When at WOT, the OXS CIS system goes to "open loop."

But this would give you the only true reference.

And then?
The GUNSON is inaccurate – here btw in a ± range where you're within your off readings. So won't help you.

Second, an Initial CO seting of 0,8% does not equal to Lambda 1(ARF 14,7:1).
It's simply an initial setting where the ECU and its FV do pull the duty cycle up to match the final needed Lambda 1.
You can set it where you want, the ECU will compensate with its duty cycle to match Lambda 1.

See here PeteKz's comment.


Honestly, ... your approach won't bring you a satisfied result.
Go where a real Lambda gas analyzer exists, anything else ist just guessing, trying, guessing.

Good luck.

Updates!

In order to absolutely verify my custom JBell made dual Voltmeter/AFR gauge was accurate, in early May I bought and AEM digital gauge compatible with my AEM 30-0300 controller. It showed exactly the same number as my JBell gauge, which eliminates the gauge as a source for my high AFR (lean) idle readings. Thanks to JBell for an awesome accurate analog looking gauge!
http://forums.pelicanparts.com/uploa...1749664994.jpg

In mid May I bought an ECU wire harness from a fellow Pelican. This to verify my O2 wiring harness was not grounding somewhere. One of the many reasons a Cab is better than a Coupe is shown here:
http://forums.pelicanparts.com/uploa...1749665247.jpg
Was able to run it through the unzipped window and leave my OE harness in place, unplugging the original and plugging in the test harness at each fitting.
http://forums.pelicanparts.com/uploa...1749665247.jpg
Ran the engine and again, warm idle AFR was nearly 16 and the car ran otherwise as before. This told me the O2 wire harness was not causing these lean readings.

Pulled and cleaned and set the throttle position switch. The one between the throttle body and the air plate underneath the rubber “hat”. I had convince myself somehow this could cause a lean idle. Not fun to get this out and of course cleaning the contacts inside had no effect on my idle, but at least I know that now.
http://forums.pelicanparts.com/uploa...1749665954.jpg

There is a happy ending, finally. See next post for the fix!

In 2021 I had the motor out and on a stand, and added SSI’s at that time. Knowing that some day I would put in an AFR gauge had a bung welded into the passenger side at the local muffler shop. Never occurred to me that most folks put them on the drivers side. The welder did a hideous job on the weld, but functionally it worked so I pressed on.

After all was back together and AFR gauge installed I began trying to get the mixture numbers displayed to match the idle frequency valve and CO numbers in the Bentley manual. I could never do it or even get them to be close. AFR (with O2 sensor plugged in) was always way too lean compared to FV or CO. Why did I care about this if FV and CO were ok? I never felt comfortable running the engine hard with the AFR showing off scale lean at idle, even though cruise and WOT were good.

I had inquired a few times over the years on Pelican to those with O2 on each bank. My question was what is the AFR number difference between left and right sides. A few people said they were nearly the same with a spread of .1 to .5 between sides. This small difference would not account for the 15.8 on my car to the 14.7 I was looking for.
I verified this by putting the AEM sensor in the Lambda O2 bung on the drivers side and it was within .5 of the pax side readings, just as others had said.

Today I took the car back to the same muffler shop and had them weld a bung into the drivers side just behind the Lambda sensor. They did a better job this time:
http://forums.pelicanparts.com/uploa...1749668142.jpg.

Look here! Three years later and AFR is 14.7 at warm idle!!

http://forums.pelicanparts.com/uploa...1749668260.jpg

Anyone with an O2 lambda system (SC 80-83) planning to add an AFR gauge, especially if you have SSI’s and a 2 in 1 out muffler, it needs to be on the drivers side behind the Lambda O2 sensor.

Thank you to everyone who had input on this over a half dozen threads since 2022. It was a frustratingly easy thing in the end.

I took my car out this afternoon and flogged the p*ss out of it.

Regards

Fantastic.

Any insight into why L/R side report different readings?

Phil

Not yet. I never found out much about the cars of the people that responded.
Could have been 2 in 2 out or some other setups that did not mirror mine well.

Or maybe I have an exhaust leak on that side. Actually just did another smoke test a couple of days ago by putting smoke up the tailpipe. Looking specifically for exhaust leak. There was a small leak from the SSI-to-muffler connection on that side so I will be replacing that gasket. Everything else looked clean.

Neil

Good report ...

beside checking the exhaust system on that side, ... did you check the fuel flow amount of all injectors so all do match?
https://nineelevenheaven.wordpress.com/adjusting-the-fuel-distributor/

March 2023:
http://forums.pelicanparts.com/uploa...1749732835.jpg

Pretty close for popsicles test tubes!

Here is a pic of the pax side. Old sensor location now plugged and arrow shows the slight exhaust leak. About 10 inches away and also downstream of the sensor. May have had enough effect to throw off the readings.

The small leak might also expand at high temps?

Phil

Yes good thought.

Forgot the picture!

http://forums.pelicanparts.com/uploa...1749763941.jpg