How does the PCM know what 14.7 is?
02-27-2007, 05:21 PM
#21
TECH Resident
iTrader: (3)
Join Date: Nov 2001
Location: Richmond, VA Where fast cars, well......are hard to come by.
Posts: 979
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by horist
I'm not sure 100% on how it works... but the key thing to remember is that it is NOT measuring 14.7:1 AFR... it is measuring the amount of oxygen left over based on the amount of oxygen in outside air (20%)
For a given fuel source (hydrocarbon based), there is an ideal amount of oxygen for combustion, however after combustion oxygen is also a by-product... stoich is when the byproduct is what is expected for an ideal mixture, anything above or below is non stoich
As for how we know it works... I personally don't know, but I know i'm not qualified enough or educated enough on the subject to invent my own sensor to compare to the narrowband... so I trust the experts and the fact that over time the average AFR of a car that's properly metering incoming air, matches stoich (14.7:1 for 100% gasoline) based on results of several companies producing wideband controllers , some of which use different sensors than others
I'm sure if you search enough or if someone here is much more well-versed in chemistry than I could explain the scientific reason
but I think it's basically something like:
xHC + yO2 = zO2 (this value has to be constant for a stoich burn no matter what kind of fuel is used)+ aC + bH20 (water is a result of combustion) + cHC (unused hydrocarbons) (xyza and b are variables, I don't know the values)
(so i guess it'd be aHC + cO2 + bN = dO2 + eNO + fHC + gH20 in this formula the the O2 after combustion is a constant for stoich, HC is left over hydrocarbons, hydrogen fuses with oxygen to form water, there are obviously other gases that are formed, CO (carbon monoxide) etc... but in the end we know that at stoich there should be a set percentage of the resultant exhaust that is pure oxygen not combined with nitrogen, carbon, or hydrogen)
the end output you want is a known value for z a and b, in order to have this a (amount of O2) must be the ideal value
This ideal value is based on the chemical makeup of the hydrocarbon used (also has to account for the large amount of nitrogen in the air and other gases)
so for example, pure gasoline (no ethanol) to obtain the value of oxygen after combustion that would be stoich... we know that it has to be:
1HC + 14.7O2 + other gases = Constant O2 + Other Gases
If this constant O2 varies, then the O2 sensor provides an electrical output in that direction
Now take E10 with a 14.13 stoich
1HC + 14.13O2 + other gases = Constant O2 + other gases
The amount of starting O2 is different but the Constant O2 is the same as pure gasoline when stoich
E85:
1HC + 9.x O2 = Constant O2 + other gases
But if we add more fuel (HC) to these examples, the constant O2 changes because now there's more Hydrogen and Carbon molecules for it to bind to, thus decreasing the Constant... if there's not enough fuel, there's too much O2 and not enough binding occurs thus O2 constant goes up
Note... any of the above may be wrong... I've edited a few times adding more things I've thought ot... seems to make sense to me but I may have mistakes
so back to the original question (well not original but azzhauler original)
The O2 sensor knows what stoich is because it knows what that Constant O2 value is... the engineers that developed it developed the electrical circuit so that it would output 450mv when this constant was reached and vary if the amount of oxygen deviates in either direction (but their main focus is to provide accurate results at or close to stoich, they didn't design it to measure the amount of oxygen, only to measure it in relation to the ideal constant)
For a given fuel source (hydrocarbon based), there is an ideal amount of oxygen for combustion, however after combustion oxygen is also a by-product... stoich is when the byproduct is what is expected for an ideal mixture, anything above or below is non stoich
As for how we know it works... I personally don't know, but I know i'm not qualified enough or educated enough on the subject to invent my own sensor to compare to the narrowband... so I trust the experts and the fact that over time the average AFR of a car that's properly metering incoming air, matches stoich (14.7:1 for 100% gasoline) based on results of several companies producing wideband controllers , some of which use different sensors than others
I'm sure if you search enough or if someone here is much more well-versed in chemistry than I could explain the scientific reason
but I think it's basically something like:
xHC + yO2 = zO2 (this value has to be constant for a stoich burn no matter what kind of fuel is used)+ aC + bH20 (water is a result of combustion) + cHC (unused hydrocarbons) (xyza and b are variables, I don't know the values)
(so i guess it'd be aHC + cO2 + bN = dO2 + eNO + fHC + gH20 in this formula the the O2 after combustion is a constant for stoich, HC is left over hydrocarbons, hydrogen fuses with oxygen to form water, there are obviously other gases that are formed, CO (carbon monoxide) etc... but in the end we know that at stoich there should be a set percentage of the resultant exhaust that is pure oxygen not combined with nitrogen, carbon, or hydrogen)
the end output you want is a known value for z a and b, in order to have this a (amount of O2) must be the ideal value
This ideal value is based on the chemical makeup of the hydrocarbon used (also has to account for the large amount of nitrogen in the air and other gases)
so for example, pure gasoline (no ethanol) to obtain the value of oxygen after combustion that would be stoich... we know that it has to be:
1HC + 14.7O2 + other gases = Constant O2 + Other Gases
If this constant O2 varies, then the O2 sensor provides an electrical output in that direction
Now take E10 with a 14.13 stoich
1HC + 14.13O2 + other gases = Constant O2 + other gases
The amount of starting O2 is different but the Constant O2 is the same as pure gasoline when stoich
E85:
1HC + 9.x O2 = Constant O2 + other gases
But if we add more fuel (HC) to these examples, the constant O2 changes because now there's more Hydrogen and Carbon molecules for it to bind to, thus decreasing the Constant... if there's not enough fuel, there's too much O2 and not enough binding occurs thus O2 constant goes up
Note... any of the above may be wrong... I've edited a few times adding more things I've thought ot... seems to make sense to me but I may have mistakes
so back to the original question (well not original but azzhauler original)
The O2 sensor knows what stoich is because it knows what that Constant O2 value is... the engineers that developed it developed the electrical circuit so that it would output 450mv when this constant was reached and vary if the amount of oxygen deviates in either direction (but their main focus is to provide accurate results at or close to stoich, they didn't design it to measure the amount of oxygen, only to measure it in relation to the ideal constant)
After reading it and thinking about it some more, it seems to make sense now. And thinking about my old LM1 software, I understand more on how it works (or at least I think I do lol). All the sensor does is tell the controlling computer "this is how many extra parts of o2 I'm seeing" (by changing its output voltage) and the computer, which has been told the stoich rate, says "ok my fuel should take 10 parts o2 for 1 part fuel and there should be , but I have an extra 2 parts left over, so I must be getting 12 parts o2 instead". Which is why in the LM1 software you have to tell it which fuel you are running, in the pcm it's coded.And then after thinking some more, that methodology makes it kinda hard to read anything richer than stoich, as all of the o2 should be consumed. Attempting to rectify this problem I found a couple sites that give more detail on o2 sensor function, which appears to be rather far from simple, not to mention the different types. Since obviously the combustion can't be perfect there will always be unused fuel and o2 at stoich, so maybe the years of data say a stoich mixture, while it can't be completely burned, shows x amount of o2 left over. And as we get closer to 0 we can determine that we are overly rich. Maybe past that point it's just splitting hairs that won't make a difference.
One of the interesting sites I saw that explained one type of sensor pretty well was: http://www.crystec.com/staoxye.htm Maybe interesting if you ask "How do it know?" A+ for horist and co.
02-27-2007, 09:39 PM
#22
From a controls standpoint, I'd call this a shifted window averaging complex multiple proportional integral derivative control scheme. The window here is the range to the + and – (the error) of the target voltage (set point) that the controller is attempting to maintain (...about 450 mV.... that whole swing within itself is the 'window'). The voltage observed in this case from the O2 is the process variable. The controller attempts to maintain this window with another PID, the fuel trims, also calculating error. The resulting fuel addition is the action used to drive this second PID’s offset. The feedback from the O2s is latent enough that it really mandates this slow averaging scheme to have the best result. Really all analog controls are averaging though; you just have to break the time portion down small enough to see it. In this application it's readily apparent.
The window here is the range to the + and – (the error) of the target voltage (set point) that the controller is attempting to maintain (...about 450 mV.... that whole swing within itself is the 'window'). The voltage observed in this case from the O2 is the process variable. The controller attempts to maintain this window with another PID, the fuel trims, also calculating error. The resulting fuel addition is the action used to drive this second PID’s offset. The feedback from the O2s is latent enough that it really mandates this slow averaging scheme to have the best result. Really all analog controls are averaging though; you just have to break the time portion down small enough to see it. In this application it's readily apparent. Last edited by Frost; 02-27-2007 at 09:50 PM.
03-01-2007, 09:39 PM
#23
Staging Lane
Join Date: Aug 2006
Location: Powell, TN
Posts: 68
Likes: 0
Received 0 Likes
on
0 Posts
So, how does a car run rich(non PE)? If the PCM is constantly adding and subtracting fuel based on nbO2 readings, to keep a car at stoich, how is it anything but stoich even if the MAF and fuel trims are completely out of whack? Does the fuel addition or subtraction of fuel have a limit?
This is for maf enabled.
This is for maf enabled.
Last edited by 2000c-5; 03-01-2007 at 09:49 PM.
03-01-2007, 09:57 PM
#25
Staging Lane
Join Date: Aug 2006
Location: Powell, TN
Posts: 68
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by slow
yes there is a limit to both, short terms, and longterms.
If the maf is wrong, the car will be correcting, until it hits the limits.
Ryan
If the maf is wrong, the car will be correcting, until it hits the limits.
Ryan
Later tuning with a wideband took the lesser part of an afternoon.
03-02-2007, 10:00 AM
#26
TECH Enthusiast
Join Date: Sep 2004
Location: Omaha, NE
Posts: 564
Likes: 0
Received 0 Likes
on
0 Posts
You guys are making a simple question too complicated. At 14.7 A/F, the O2 sensor swings between approx .2 to .8 Volts several times per second, called crosspoints. If these voltage swings average .45 volts, the fuel trims are at zero. If the voltage swings spend more time above .45V then below it and for example, we now have a new voltage average of .55V, the STFT will indicate an approx -10%, and will pull fuel until the O2's voltage swings average .45V again, on and on and on. That simple.
03-02-2007, 10:38 AM
#27
TECH Addict
iTrader: (10)
Join Date: Jun 2004
Location: Commerce Twp, MI
Posts: 2,918
Likes: 0
Received 0 Likes
on
0 Posts
The STFT's are causing the swings above and below 14.7, which is measured at a .550mV switchpoint (for airflow greater than 32grams/second). The switch point drops as airflow drops to compensate for O2 performance I believe. Either way, the STFT's increase or decrease to be (I believe) .200mV above or below the switch point. If it notices the STFT's are averaging less than more than desired, the LTFT's update and the averages are re-analyzed.
Think of a magnified view of the AFR in closed loop in the form of a sin wave where the bottom is rich and the top is lean. The STFT's add in and subtract out 2~3% to create that wave. If you were to draw a line straight through the middle of those fluctuations, that's your average. If that average is consistently above 0%, the LTFT's are increased slightly to shift the average back down to where it's supposed to be. If they're consistently below 0%, LTFT's are reduced slightly to shift the average back up.
There are filters involved and minimum step sizes that restrict the movement of the fuel trims...both on the short term and long term levels. That's why, if you stare at your LTFT's long enough, you'll see them fluctuate sometimes +/-0.8% for similar driving conditions. It's not the ambient conditions that are changing in that short period of time. It's the fuel trims falling between the limitations of the PCM trying to find the sweet spot.
Think of a magnified view of the AFR in closed loop in the form of a sin wave where the bottom is rich and the top is lean. The STFT's add in and subtract out 2~3% to create that wave. If you were to draw a line straight through the middle of those fluctuations, that's your average. If that average is consistently above 0%, the LTFT's are increased slightly to shift the average back down to where it's supposed to be. If they're consistently below 0%, LTFT's are reduced slightly to shift the average back up.
There are filters involved and minimum step sizes that restrict the movement of the fuel trims...both on the short term and long term levels. That's why, if you stare at your LTFT's long enough, you'll see them fluctuate sometimes +/-0.8% for similar driving conditions. It's not the ambient conditions that are changing in that short period of time. It's the fuel trims falling between the limitations of the PCM trying to find the sweet spot.
Last edited by SSpdDmon; 03-02-2007 at 11:53 AM.
03-03-2007, 04:21 AM
#29
TECH Junkie
iTrader: (3)
Join Date: Nov 2001
Location: TEXASS
Posts: 3,202
Likes: 0
Received 0 Likes
on
0 Posts
Originally Posted by Frost
There is a finer cyclical control that causes the swing.... how else could trims stay constant and you still see the swing?
