The truth about widebands. Let the discussion begin

[blownbluez06]

While many of you are aware of this, I am also acutely aware that many are not, so this is for the group that is not. Throughout the past few years, I've had many discussions with friends and various forum members about widebands. There is a lot confusion about how this works so I figured I would post up my .02 on the subject. The misunderstandings all seem to be relatively the same and when a misunderstanding exists, the person thinks they should be looking for different stoich values on the display when they deviate from pump gas, or change the target because they have a percentage of ethanol in their pump gas. This is not the case. If you run Methanol, E85, E100, Propane, or CNG, you will want to see 14.7:1 on your wideband display. The wideband is designed to take the sensor voltage that is equal to Lambda 1.0 value and convert it to the number that represents stoich for pump gas, 14.7:1. If you empty your tank and fill it with Ethanol, you would change your Stoich value in your Fuel>general tab to whatever you agree stoich actual is, be it 9.7-9.9:1. Doing this will multiply your actual yielded injector pulse width by whatever is necessary to achieve that actual true AFR. In other words if you went from pump fuel with a value of 14.68 in that position and you changed to E85, placing a 9.9 value in there, it would multiply your resultant pulse width by 1.48 without changing the values in the VE or MAF. If you want a .85 lambda at WOT with your alternate fuel, you would look to achieve a 12.5:1 on your wideband display. All the wideband is doing is taking a sensor voltage that equates to a specific lambda and displaying a value that is appropriate for gasoline. The sensor has no idea what your actual AFR is. It is only counting oxygen content in your exhaust in an attempt to provide feedback on your AFR. The resultant display is not a fact. Why? Because many other factors will skew your values. Bad plug wires, plugs, detonation, a bad injector in one cylinder can alter your reading by a proportional amount, making a tuner think they are slightly lean. This is why I like to still view the narrow bands. You can see when one bank is having trouble at WOT. It is important to keep in mind where your sensor is placed when tuning. Most of the cars I tune have the wideband sensor placed before the X pipe, so it only picks up one bank. I've also seen the wideband bungs placed so early in the collector that I would expect one specific cylinder to have a huge emphasis over the remaining 7.

[Pekka_Perkeles]

A bit off-topic, but in addition to Wideband sensor, individual exhaust temp sensors are really nice with heavily modified engines. Just recently I was tuning a car which had a temp sensor for each cylinder. After a minor problem with leaking injector o-ring, I continued tuning and, well, AFR values were off in the areas I had already tuned and it did not make sense. Pretty soon I checked the exhaust temps and one of the cylinders was way behind others. Again, a minor electrical problem was detected, fixed and AFR error was back in the reasonable values. This with 3 bar CoS, which does not detect P0300 anymore..

I also agree that narrowbands are very useful.

I was planning to have dual widebands in my car, i.e. one for each bank, but nowadays I'm more inclined to have exhaust temp sensors for each cylinder. Overkill? Maybe, but I still consider to have them. There are some nice gauges that can show up to four cylinders in the single gauge and have memory function as well, including red warning leds.

[RACELS1]

So if I'm understanding this correctly. even if I'm commanding and running at 13.5:1 my NGK/AFX display will still read 14.7:1?

[Pekka_Perkeles]

So if I'm understanding this correctly. even if I'm commanding and running at 13.5:1 my NGK/AFX display will still read 14.7:1?

Well, I think no. Or it of course depends what you mean with the text above. I'm not sure if I understand it with similar way you intended.

[superwagon]

If the fuel you are using has a stoich value of 13.5:1 which = lambda 1 then your ngk afx should display 14.57:1. Your ngk displays 14.57:1 at lambda 1 not 14.7:1 like some other wideband may.

[MikeOD]

So if I'm understanding this correctly. even if I'm commanding and running at 13.5:1 my NGK/AFX display will still read 14.7:1?

No, you're misunderstanding...you can command anything you want...so long as you're close enough to the actual stoich value of your fuel that the fuel trims can correct you if you didn't command what stoich really is, then your NGK will show 14.57 (NGK AFX doest not show 14.7 at Lambda 1.0)...however, when it shows you that value, your AFR may not be whatever you commanded, it will be whatever that fuel is at Lambda 1.0.

[5_Liter_Eater]

Dealing in AFR when using anything other than pump gas (not E85) can be confusing, and even then ethanol content blurs it as well. All widebands are reading lambda but different widebands do different things as far as converting lambda to AFR. Some are hard coded to 14.7:1 for lambda = 1. NGK is 14.57:1 for lambda 1. Others like the Innovates you can configure to display whatever you want for lambda 1. Different fuels have different lambdas/stouchs and whatever AFR the wideband is configured to show as stoich it will, regardless of the fuel. So if you have stoich as 14.7 and run E85 it will show 14.7 when the actual AFR is really more like 9.76. Even if you're running gasoline and have stoich set to 14.7 then it will show 14.7 for stoich even though its really more like 14.2 due to the ethanol blend. So if possible you want to configure the wideband to be close to the actual stoich of the fuel. Becuase of this it's really easier to start tuning in terms of lambda error instead of AFR error.

It's confusing but whats more important is that you define stoich in the tune to match whatever your wideband thinks is stoich, AFR-wise. If the tune and wideband don't agree then while you can dial it in, your fuel numbers aren't accurate and your airflow numbers wont be either. This may or may not be a big deal for you as since most of us big boys need to scale their injectors anyhow we don't have accurate numbers anyhow.

[smithers]

Becuase of this it's really easier to start tuning in terms of lambda error instead of AFR error.
.

I agree.

[MikeOD]

This may or may not be a big deal for you as since most of us big boys need to scale their injectors anyhow we don't have accurate numbers anyhow.

Nah...the newer big boys have a 127 lb/hr injector size limit...and 4000 g/sec on the MAF table...lots of power before scaling.

[Blingole]

It's confusing but whats more important is that you define stoich in the tune to match whatever your wideband thinks is stoich, AFR-wise. If the tune and wideband don't agree then while you can dial it in, your fuel numbers aren't accurate and your airflow numbers wont be either. This may or may not be a big deal for you as since most of us big boys need to scale their injectors anyhow we don't have accurate numbers anyhow.

This makes sense now as i used 14.68 in mine with e85 and all my Ve numbers were higher than normal.

[5_Liter_Eater]

Nah...the newer big boys have a 127 lb/hr injector size limit...and 4000 g/sec on the MAF table...lots of power before scaling.

SHOOSH! I'm jelly of the newer PCM's. Still limited to 2.5 bar, 12k Hz and #/hr on my E40.

[DSteck]

Use lambda. Forget AFR. Your life will be a lot easier.

I set up flex fuel on a D1 Procharged 408 GTO last night. With fuel alcohol content varying from 10% to 55%, trying to use AFR would have been impossible.

[jimmyblue]

My LM-1 will display lambda; other makes / models may or
may not. Lambda does free you from needing to know the
chemistry - you just aim for the bull's eye, wherever it is.

[DSteck]

I have four NGK AFXs. I've never in my life looked at the display. Lol.

[MikeOD]

I put electrical tape over mine lol.

[LSxpwrdZ]

Same here... the NGK display is useless anyway with its slow refresh rate. Only thing I verify is the readings I get in my scanner when the NGK powers up to verify my stoich and rull rich points are spot on.

Lambda is your friend for sure... I stopped using AFR a few years ago and makes life so much easier.

[onspeed]

If you remember high school chemistry, here's another explanation that might help it make sense (although I'm a chemist so maybe it makes sense only to me.)

Stoichiometric ratio is simply the ratio of substances to make a product with no excess starting material. For instance, O2 + 2H2 > 2 H2O if we were to have an "oxygen to hydrogen" ratio (equivalent to air to fuel ratio in my example later), it would be 1:2. You can scale the reaction by any number and the ratio will be the same (i.e. scale by multiplying by 2 = 2 O2 + 2 H2 > 4 H2O ). But the amount of stuff has to be equal for both side of the reaction. So if you add up the oxygen on the left, 2, it equals the amount of oxygen on the right, 2 and likewise for hydrogen. It can't disappear or be created from nothing.

If we had excess oxygen, 2 O2 + 2 H2 > 2 H2O + O2, then our ratio now becomes 2:2 = 1:1 vs the ideal 1:2 from above. We have left over oxygen since there's not enough hydrogen present to convert to water.

So for AFR, that's simply the ratio, by mass, for the air to the fuel. The ideal combustion reaction (meaning everything on the left side of the reaction is converted to something else) is something like 12.5 O2 + C8H18 > 8CO2 + 9H2O. (But a little more complex since C8H18 is octane, and gasoline is composed of more than just octane) Oxygen reacts with octane to produce carbon dioxide and water. Air to fuel ratio here is 12.5:1. Running rich means we have excess octane after the reaction, and running lean means we have excess oxygen after the reaction. If you're still following me, then you can see that we change the composition of the fuel, say to ethane (C2H6), then our coefficients will have to be different. 3.5O2 + C2H6 > 2CO2 + 3H2O. Air to fuel ratio here is 3.5:1. The 14.7:1 AFR we shoot for is calculated for typical pump gas as the ratio we want so that all the gas and oxygen in the combustion chamber is consumed. Once we start changing fuels the molecular formula of the gas is changed and the reaction numbers will be different as I showed with octane versus ethane. Hence the magical 14.7 goes out the window since it only applies for pump gas (and non-ethanol pump gas for that matter).


When we tune with a lambda reading, we're not really concerned with the composition of the fuel. Lambda is just an arbitrary name we give to a certain reference point. The O2 sensor is a zirconia membrane, that oxygen ions can pass through. On one side of the membrane is exhaust gas, on the other is a reference cell with a predetermined oxygen content. A current is applied through the membrane and resulting voltage is monitored. When the exhaust is rich, there is excess fuel and a lack of oxygen, causing the oxygen in the reference cell to pass through the membrane to the exhaust gas due to simple diffusion (moves from high concentration to low concentration until it's equal). This movement of ions through the membrane decreases the voltage. When the exhaust is lean, the oxygen ions move in the opposite direction and the voltage increases. When the exhaust has no excess fuel or excess oxygen, the voltage is at some point (.45v on narrowbands I think) that's already determined and that voltage is 1 lambda. The wideband sensor works a tad different (essentially two membranes) but the basic concept is similar.

TL;DR - AFR readings is a calculated result based the lamda reading and a certain fuel. Lamda reading is more of raw data from the O2 sensor and does not take into consideration the fuel used, only if the exhaust gas has excess or shortage of oxygen. 1 Lamda will always be the stoichiometric ratio for any fuel.

[DSteck]

^^ I like this guy. Especially since he flagged his own post as TL;DR. Haha.

[fullthrottle]

My head hurts. :lol:

[blownbluez06]

I agree it's best to start with lambda if you aren't already used to doing something else.