AEM Wideband reads leaner than factory widebands during WOT under boost

[rcmikey]

Can anyone explain this? 2014 GT with 2.3 Whipple, 20lbs boost, AEM wideband located opposite of factory wideband in same pipe. Steady state both factory and AEM wideband read the same. They will mirror each other until boost is introduced into the mix. At WOT the AEM will read 8% leaner than factory. I have my ideas but I am looking for input that might lead to a real cause. Any thoughts welcome.

[veeefour]

One is calibrated to 14.08 and second to approx 14.7 I believe.

[Ortiz]

Is the AEM calibrated for 14.7 or for 14.08? Reading it through lambda or afr?

[Ben C]

It's my understanding that when wideband sensors get old they read stoichiometric correctly however they "exaggerate" rich and lean readings. Can't find where I read that though..

[rcmikey]

Lambda

[rcmikey]

One is calibrated to 14.08 and second to approx 14.7 I believe.

This would not explain why at steady state they are the same. Only WOT under boost does the AEM read leaner

[rcmikey]

It's my understanding that when wideband sensors get old they read stoichiometric correctly however they "exaggerate" rich and lean readings. Can't find where I read that though..

I get the same result with new sensors, including a new AEM. Same result with different sensors. Still puzzled.

[jsllc]

So rather than open up the whole debate of WBs there are at least 10 different treads that cover this subject. There are many factors that go into the accuracy of the WB under various circumstances. Maybe browse those for more info. Back pressure, speed of sensors, type of sensor, position in exhaust and frequency of polling are just a few. We would have to know all factors to know which one is more accurate. I will pic one, polling the AEM X-Series is considered the quickest in the industry. We do not know how the factory ECM polls or calculates the data. So for example if AEM is polled every 10ms and has a 19ms response time the results will be much different that if it is polled every 50ms and has a 104ms response time.

In short I would believe the AEM if installed correctly and is newer. Ask Dr. Mike on the threads. He has spent more time on this than all of us.
2018-01-05_18-29-20.jpg

[rcmikey]

So rather than open up the whole debate of WBs there are at least 10 different treads that cover this subject. There are many factors that go into the accuracy of the WB under various circumstances. Maybe browse those for more info. Back pressure, speed of sensors, type of sensor, position in exhaust and frequency of polling are just a few. We would have to know all factors to know which one is more accurate. I will pic one, polling the AEM X-Series is considered the quickest in the industry. We do not know how the factory ECM polls or calculates the data. So for example if AEM is polled every 10ms and has a 19ms response time the results will be much different that if it is polled every 50ms and has a 104ms response time.

In short I would believe the AEM if installed correctly and is newer. Ask Dr. Mike on the threads. He has spent more time on this than all of us.
2018-01-05_18-29-20.jpg

I can agree there are many factors that can and will effect the accuracy of different wideband sensors. 1. Back pressure: tested with and without cut outs open with the same results. 2. Speed type and position of sensors/polling: this should not make an 8% error between the two because they are dead on at cruise. I could see if they were different all the time. During a dyno pull the entire time under boost the AEM is reading lean. Sensor is located same as stock sensor just 180 degrees out. Dyno sensor reads 8% lean also. Not to say you are wrong. I appreciate the input.

With the same sensor setup on a non boosted coyote the sensors are dead on for the entire WOT pull to include the dyno wideband.

Do you think I could be experiencing blowthough at that boost level and cam timing? I know that the copperhead will calculate within the speed density logic to compensate for blowthrough. So the factory wideband reading would be compensated for this. Any instance of blowthrough will create a lean condition post combustion than actually occurs in the cylinder. My thoughts are the AEM has no input of correction therefore will read leaner than the factory corrected sensors. The dyno sensor has no logic to compensate so it would read the same as the AEM. Thoughts????

If my thought process is correct this may be a solid indicator of blowthrough for boosted coyotes. Maybe help in speed density tuning?

[jsllc]

Sorry,

Wished I was better with these Dr. Mike can tell you about most all of the 3rd party WBs. Factory readings are subject to their undocumented formulas/modifiers. We just do not know what they are using to create the numbers. My bet would be speed of sensor, polling and math. None of which we will ever know. I would just go with the AEM for tuning AS LONG AS YOU ARE SURE of the install.

I understand the Coyote. I had SN# 12 of the Ford Shelby F150 to tune. It Sucked. I was encrypted by Ford Racing and Whipple. I got it after it blew up at 6500 miles with factory tune. Went round and round with Whipple/Tuscany saying it was too lean. They all told me I did not know how to run a Coyote. They would not pay to fix it for the customer and would not help with tune. Worked to HPT to break encryption which eventually got done. Customer gave up put a third motor in it and sold the truck $150,000 down the drain.

Something you said. They two sensors ate 180 degrees opposite? Thee sensors should both have a downward angle. With 0 being top 280 to 80 would be max but 315 to 45 would be better.

[veeefour]

This would not explain why at steady state they are the same. Only WOT under boost does the AEM read leaner

It would as PCM lambda is altered with "transport delay" and this parameter is not linear.

[rcmikey]

Sorry,

Wished I was better with these Dr. Mike can tell you about most all of the 3rd party WBs. Factory readings are subject to their undocumented formulas/modifiers. We just do not know what they are using to create the numbers. My bet would be speed of sensor, polling and math. None of which we will ever know. I would just go with the AEM for tuning AS LONG AS YOU ARE SURE of the install.

I understand the Coyote. I had SN# 12 of the Ford Shelby F150 to tune. It Sucked. I was encrypted by Ford Racing and Whipple. I got it after it blew up at 6500 miles with factory tune. Went round and round with Whipple/Tuscany saying it was too lean. They all told me I did not know how to run a Coyote. They would not pay to fix it for the customer and would not help with tune. Worked to HPT to break encryption which eventually got done. Customer gave up put a third motor in it and sold the truck $150,000 down the drain.

Something you said. They two sensors ate 180 degrees opposite? Thee sensors should both have a downward angle. With 0 being top 280 to 80 would be max but 315 to 45 would be better.

I agree completely about the undocumented modifiers and background math that we may never know. The car has been down the quarter mile over 100 times last year commanding .715 lambda with no issues. On board wide band shows commanded, AEM would show .78 lambda. Plugs always looked great. After three runs there would be at least 10% of the cad coating on the bottom face of the plug. Always wanting a bit more power I added 2 lb boost and monoblade. Timing was left alone and commanded I leaned out to .78 and picked up right at 100 hp. (875hp) On board wide band shows commanded and AEM showed .82 lambda. Leaner is meaner but getting into the not so safe zone from my perspective. My mistake was not to do one thing at a time.

Best move will be for me to go back to the commanded .715 to keep in on the safe side until I know for sure how the onboard calculations are made.

The location of the sensors are not exactly 180 apart. They are across from each other at the proper downward angle.

[jsllc]

If you have the time and the inclination you may want to swap locations between the two different WBs just to see.

So according to Kenne Bell (huge ford guys), which I am sure has made more logged passes than any of us, Lambda .78-.80 can be run all day long IF you have everything dialed in and are sure that it is consistent. That is where I set up cars for the "glory" run on the dyno. then I go back to .78. Max power is at .85. So if you are the one monitoring and driving .8 would be as high as I go (USING AEM AS the reference). Here is an article about dyno runs. http://kennebell.net/tech/tuning-tec...alls-problems/

[superman07]

Great info, new favorite forum!

[dr.mike]

To be safe, I would pay attention to the leaner reading device, until I could verify it. If you want to verify for yourself, you can use BAR97 HI test gas. It should read 0.883 lambda, at STP.

[rcmikey]

To be safe, I would pay attention to the leaner reading device, until I could verify it. If you want to verify for yourself, you can use BAR97 HI test gas. It should read 0.883 lambda, at STP.

Appreciate the input. The AEM only reads lean under boost. Don't know how I could test the sensor that way. Both the AEM and the factory wide bands read exactly the same until boost is introduced.

[realspeeddan]

The Ford sensor readings are adjusted by estimated exhaust temperature. If your exhaust configuration is not stock and you haven't adjusted the exhaust temperature and/or Temp Loss to O2 tables then the ECU is "correcting" the reading based on factory estimated temps, which may not be the same as actual temps.

You can test it. If you change the Temp Loss to O2 reading drastically you will see the reading get closer or farther away from the aftermarket sensor reading, depending on if you raise or lower the temperature tables.

[rcmikey]

The Ford sensor readings are adjusted by estimated exhaust temperature. If your exhaust configuration is not stock and you haven't adjusted the exhaust temperature and/or Temp Loss to O2 tables then the ECU is "correcting" the reading based on factory estimated temps, which may not be the same as actual temps.

You can test it. If you change the Temp Loss to O2 reading drastically you will see the reading get closer or farther away from the aftermarket sensor reading, depending on if you raise or lower the temperature tables.

This makes complete sense. Car does have long tubes and runs FTW 99+ fuel. In my mind the temps in the exhaust temp table should have been lowered. In my current tune the temps had been raised 200 in the WOT load area. I believe the higher value was mistakenly entered using misinformation or just not knowing what the table was for in HP Tuners. If I take into consideration the fuel I use and the headers combined they values should be lowered. I will have to strap it down again and do some experimenting. Stock temp values are a good place to start. This might finally put this to rest.

[GrannySShifting]

This makes complete sense. Car does have long tubes and runs FTW 99+ fuel. In my mind the temps in the exhaust temp table should have been lowered. In my current tune the temps had been raised 200 in the WOT load area. I believe the higher value was mistakenly entered using misinformation or just not knowing what the table was for in HP Tuners. If I take into consideration the fuel I use and the headers combined they values should be lowered. I will have to strap it down again and do some experimenting. Stock temp values are a good place to start. This might finally put this to rest.

Do you have access to EGT you could plug actual values in? Doesthis show that you cant necessarily rely on the OEM widebands Id ask??... Ive never had a fresh good O2 read so off because of a few hundred degrees in EGT change, even in turbo cars. Cant see 200 degrees REALLY skewing widebands reading that much?

How would you ACCURATELY change these exh temp tables? You would need 3 widebands, two outside the car to make sure they read agreeing with each other and one isnt wacky, then just fudge the exh numbers until the OEM ones report the same thing? Sounds like so much fudgery

[GrannySShifting]

I agree completely about the undocumented modifiers and background math that we may never know. The car has been down the quarter mile over 100 times last year commanding .715 lambda with no issues. On board wide band shows commanded, AEM would show .78 lambda. Plugs always looked great. After three runs there would be at least 10% of the cad coating on the bottom face of the plug. Always wanting a bit more power I added 2 lb boost and monoblade. Timing was left alone and commanded I leaned out to .78 and picked up right at 100 hp. (875hp) On board wide band shows commanded and AEM showed .82 lambda. Leaner is meaner but getting into the not so safe zone from my perspective. My mistake was not to do one thing at a time.

Best move will be for me to go back to the commanded .715 to keep in on the safe side until I know for sure how the onboard calculations are made.

The location of the sensors are not exactly 180 apart. They are across from each other at the proper downward angle.

What fuel is this? .715 lambda On gas thats really rich, on E10 even. Little rich even for E85?