Wideband Lambdasonde: was sie kann und was nicht.
Verfasst: Do 25. Dez 2008, 23:30
Wem etwas langweilig ist kann sich mal hier durcharbeiten.
ich bin noch nicht ganz durch aber wollte es euch auch nicht vorenthalten.
Quellen:
http://www.smartcarby.com
http://forum.nitrous-advice.org/viewtop ... f=5&t=4106
-How a Wideband sensor works and what its limitations are.
The common called ‘oxygen sensor’ is not in fact an oxygen sensing device. Some people are mislead by this incorrect labeling into thinking that the sensor measures the level of oxygen in the exhaust and determines the air fuel ratio based upon that. The mislabeling is even perpetrated by Bosch and other manufacturers. The simple narrow band sensor is virtually useless for measuring the air fuel ratio of an engine other than to achieve a switching based upon stoichiometric conditions present in the sensor. There is no place for a narrow band sensor in performance tuning. However a narrow band can be used in some situations. Some people spend considerable time correlating narrow band sensors to a gas bench and can generate useful data applicable to their engine but that is not generally done.
Wideband sensors are very popular for measuring the air fuel ratio (AFR) of an engine. Many tuners use them to set a desired AFR that they are confident will achieve a desired outcome for a customer. The wideband sensor is modern, fast, accurate and far cheaper than investment in a 4 gas or 5 gas analyzer. Most people are taught that a wideband sensor can accurately display the AFR of the engine. A WB sensor can easily read leaner than truth and in fact even read richer than the truth. The conditions to do this happen quite often on hotrod engines that have not been developed sufficiently. For the interest of Nitrous tuners, a leaner than truth reading would seem not to be a problem, as one may argue that running the engine excessively rich is OK to do. And the other argument would be that a richer than truth reading should be avoided at all times. If you are tuning successfully at the moment then all is well until you decide to increase the power. The point of destruction isn’t obvious until it happens. By understanding how a wideband sensor (WB) can incorrectly read the AFR and performing some simple tests you can determine if the WB is accurate for your tuning purpose. First we need to understand what this number called Lambda is.
Lambda is a number that represents the completeness of the combustion of reactants. The reactants within the chamber are simply any combustible gases and the Oxygen from the atmosphere or the nitrous. Lambda 1 is complete combustion of all reactants, which means all fuel molecules have been fully oxidized and all free Oxygen has been consumed by the reactants. Achieving this state of complete reaction has nothing to do with the type of fuel used. Any fuel can be completely reacted and all the Oxygen can be used at the same time if the fuel and Oxygen relationship is correct. For instance you could fill a container with just Hydrogen gas and Oxygen gas in the exact proportions of 2 Hydrogen molecules to every Oxygen molecule, there is no air involved, and then you can say that the mixture in the container is H2O which is of course water. But it isn’t water yet is it, we haven’t combusted it yet. If we let the container sit at room temperature for 30,000 years there will eventually be no gases present in the container as all the gases will have slowly combusted into water. If we heat the mixture in the container up we don't have to wait 30,000 years to get a drink. As we heat the gases up the reaction happens much faster. By the time we heat it up to 500 degrees or so it’s reacting in milliseconds. The reason why I'm talking about heat and time will become apparent later.
The mixture we filled the container with was a Lambda 1 mixture. If we filled the container with too many Hydrogen’s, for example say we put in 5% too many, then we filled the container with Lambda 0.95. If we sampled some of the gases that we put into the container with a WB sensor, the sensor will use a destructive process called catalytic reaction to measure what ratio the reactants to Oxygen is. The WB sensor will read Lambda 1 for our perfect chamber mix and Lambda 0.95 for our rich mixture example. The wideband sensor can only measure the lambda of gases. There is no need for conventional combustion of our gases to occur in order for the WB sensor to work.
There is no need for air or anything else to confuse us. So thinking in terms of lambda enables us to think independently of fuel type or air quality.
How an AFR number is arrived at, is a simple multiplication of the Lambda value times the stoichiometric value of whatever fuel you are using. However finding out the true stoich value of a commercial gasoline is just about impossible so we just use 14.7. It doesn’t matter if you use Lambda or AFR in the concept of engine tuning, but Lambda has advantages for ECU tuning and jet changing.
If we want to alter the mixture based upon thinking in Lambda its easy because Lambda 0.95 is 5% rich and to alter it to lambda 1 requires 5% smaller jetting flow or injector timing etc. Whatever you’re using to tune, it is easy to calculate changes.
If we place the WB sensor in a container of only CO2 (Carbon Dioxide) gas it will read lambda 1. This is because CO2 is the completely Oxidized state of carbon. There are no molecules that can be reacted with oxygen.
If we fill the container with CO (carbon monoxide) the WB cell will transfer oxygen’s from the atmosphere onto the surface of the chamber as the catalyst reacts the Oxygen with the CO until all the CO molecules have turned into CO2. The number of Oxygen molecules to complete the process of CO to CO2 is counted by the electronics and a Lambda value is displayed. The Lambda value will be infinity rich as the oxygen transfer rate will be the maximum the cell can produce.
If we fill the container with Hydrogen, the same Oxygen ion transfer process occurs and the catalyst will do the same thing etc. The electronics will ‘count’ the number of Oxygen molecules transferring that are necessary to react the Hydrogen to its fully oxidized state (H2O--water) and display a Lambda value. This counting process is very dependent upon time required to react the gases.
Gasoline and other fuels are Hydrocarbons, which means they consist of Hydrogen and Carbon. If we fill our container with Hydrocarbon gas the WB will count the Oxygen’s needed to fully react the HC gas just the same as it did for the earlier examples.
The catalyst in the sensor needs to be heated to a constant temperature in order to achieve a fast enough reaction. WB sensors have a heater element in them to maintain the temperature to 850 Celsius for correct reaction rate etc of the catalyst. Basically when it’s hot, it happens fast enough to be of value as a measurement, so if the heater circuit can’t heat up the cell enough or the cell gets too hot it’s not accurate and you get an error.
So a power supply problem or an exhaust gas condition that is preventing the cell from being within its correct operating temperature range causes the system to be no longer accurate.
Now let’s fill our container with HC’s and Oxygen so that there is a 5% rich mixture in it and see what’s happened after some time to complete the reactions. The H’s have consumed the necessary Oxygen’s to turn into H20 (water) and some of the carbons have completed both stages of reaction to make CO2 and then we have run out of Oxygen within the container. That results in some of the carbons only making it to the CO stage. It may have even resulted in some of the Hydrogen’s not being combusted so there would be H’s left over in that case. In order to complete these reactions we have to supply oxygen from the atmosphere. So now the electronics measures the charge from the oxygen ion transfer that is supplying Oxygen’s to the catalyst to enable the completion of the reaction of the remaining CO’s and H’s to the final reacted state of CO2 and H2O. Once the reactions have been completed there will be stoichiometric conditions within our container. We had to supply some extra oxygen to the mixture by grabbing the Oxygen’s from the atmosphere to do it and that is how we calculated that we were 5% rich. The point is that technically the WB cell alters the air fuel ratio of the environment. This effect is insignificant in a moving environment like a exhaust pipe. However in an engine we don’t know what the actual Oxygen present in the exhaust is. Any available Oxygen’s in the exhaust gas gets used without being counted and that’s where the WB cell falls down, it can’t measure the combustion efficiency. Inefficient combustion has HC’s and H’s and O’s left over and therefore an inefficiently combustion engine is read as either richer or leaner than truth.
Why is there all this talk about combustion efficiency? In an engine with poor combustion efficiency the gases present in the exhaust doesn’t follow the science books, you can have an engine that is rich but it can have Oxygen in the exhaust, or you can have an engine that is lean and it can have unburnt fuel in the exhaust. These engines give results opposite to the science books. The factors that control combustion efficiency are not always optimum in all areas of the chamber and you get fuels exhausting from one area and Oxygen’s from another.
If we had a mixture in our container that was 5% excess oxygen (lean) the sensor will complete all the oxidization of the fuel molecules and then discover that there are still Oxygen’s left over. So only after it’s completed the reactions will it remove the Oxygen’s from our container and tell us what it did by displaying lambda 1.05. In an engine with poor combustion efficiency a lean mixture can contain unburnt fuel so the catalyst has to react that first with whatever available oxygen present in the exhaust gas and then the electronics removes any excess Oxygen’s and counts them.
The combustion within the cylinder is far from consistent throughout the cylinder. If it was a perfect world and we input a perfectly homogenized stoichiometric ratio there would be lambda 1 mixture everywhere even viewed on a molecule by molecule basis. If we use a simple HC compound such as Methanol and that’s all that we put in there and somehow we stop engine oil etc contaminating it, we can surround that methanol molecule with the right number of oxygen molecules to completely combust it. If we did that with every molecule the energy release would be as per theory. The energy release would occur at the same rate throughout the mixture. In practice we can’t do that because we don’t have time to perfectly mix the fuel and air and we don’t have all the other factors necessary to achieve that perfect combustion.
Let’s just look at Methanol for a little bit, it’s a CO molecule surrounded by H’s. In fact the Carbon element has 3 Hydrogen’s attached to one side of it with 1 Oxygen attached on the other side of it. To that Oxygen is attached another Hydrogen. The way that this simple molecule reacts is not simple. There are many processes or methods of reactions, no theory explains it all. Mankind cant work it all out yet so don’t ask me. Suffice it to say for our purpose here, that a reaction sequence takes place with an unknown order or reason. But we know the end result of a perfect reaction with certainty. The 4 Hydrogen molecules of the methanol will create 2 water molecules and the Carbon and oxygen will eventually end up as CO2. If we don’t get this to happen in the chamber then the catalyst in the WB will finish the job for us and tell us some sort of answer as to how it did it. That answer is of course told to us in the form of Lambda.
The inconsistencies throughout the chamber mean that somewhere there is complete combustion of the Methanol and somewhere there isn’t. So the sensor has to receive the incompletely combusted molecules and complete it. That isn’t always going to happen is it? Any liquids present can’t be combusted until the liquid is turned into gas and we can only react the actual sample within the cell. Anything that bypasses we miss out on.
Why do you get soot in the tail pipe? It’s because Carbon hasn’t been reacted in the cylinder and the temperature conditions in the exhaust pipe are too cold to complete the carbon reaction that you missed in the cylinder. So soot affects the lambda readout.
Even if we forget about that, we then have to react whatever incomplete reactants there are in contact with the catalyst. Any Hydrogen entering the chamber that is not attached to another element is the easiest to react so they start the process first and demand the oxygen first. Any CO in contact gets reacted in a longer time period than the H’s. Any HC’s take longer again. Etc. The more complex the molecule the longer it takes to break it down and react it. In the mean time the environment is constantly changing, there is even oxygen hitting the catalyst. It’s a mess in there.
Any oxygen in the exhaust pipe gets on the surface of the catalyst and reacts with any molecule present that is not completely reacted and isn’t counted by the electronics because it’s not transferring in or out of the cell. So as Oxygen’s arrive they react within this constantly changing environment.
The reality of all this is that you can’t determine the state of the reaction completion within the engine cylinder, which is where you really want to know how completed it is. You can’t tell that from a reaction that is completed due to catalytic action. Assuming a perfect world all you can determine via catalytic reaction to completion is how far from stoichiometric the mixture in the chamber was.
Now this is the important point to understand; a WB tells you nothing about the burn within the cylinder. It actually doesn’t even need a burn. When you get a misfire it reads it as lean but that’s because it hasn’t had time to react to completion. In reality there is a usable AFR in a misfiring cylinder, as an example say its EFI, if the injector opened correctly and the other factors were correct then fuel was given correctly to the cylinder but if the arc failed to ignite the cylinder with sufficient energy the burn will be far from complete or not even a burn at all in the case of a electrical short and yet the WB reads LEAN. No its not, the cylinders not lean, it just didn’t ignite that’s all. The reason an ignition miss shows up as a lean spike is that the HC chains are harder to oxidize. They go unnoticed, and that is what causes the ‘lean’ indication. Remember that the wideband uses the O2 that’s present in the exhaust gas firstly. If the wideband adequately completed the HC reaction at the sensor interface – most of the lean error would go away. Consider that a misfire is not a change of AFR it’s just a non combustion within the cylinder and the wideband is a combustion device so if the wideband was able to complete the misfired gas output it wouldn’t read a lambda change. Hopefully this may make it clear as to the limitations of the WB sensor and give you some concept of its time limitations too. A miss fire is a time fault.
-Testing WB accuracy.
If the combustion efficiency is poor then altering the ignition timing can alter the Lambda readout of a WB. If the cylinder is insufficiently vaporized the combustion has to progress through more stages to get to completion. Altering ignition timing alters the completion of the combustion within the cylinder. Sometimes advancing the ignition timing improves completion sometimes it reduces completion. So when you do a test and alter the timing and the WB reads the Mixture has changed and you know you haven’t changed the mixture, what you are really seeing is a change in the reaction completion loading of the cell. You have no way of knowing what gas changed, to do that you have to use a gas bench or better still, a gas spectrometer.
Now we put together a hotrod engine and start it up. It’s got more power than the stocker we started with, as it has a higher fuel and air flow rate. We install a WB and measure the Lambda and we decide we don’t want it so rich so we change the fueling of the engine. Wrong--- we didn’t check the combustion efficiency first, we didn’t find out how complete the combustion was. That’s why you can sometimes change jets, etc. and things don’t change as they should. You can spend a long time tuning with a WB and never see what is really happening inside the cylinder.
The main factors that control combustion efficiency within the cylinder are vaporization, homogenization and distribution. The lambda measurement of the catalytic reaction is only based upon gases present and time available.
How different gases cause problems with the WB sensor
Hydrogen (H) makes the sensor think the mixture is richer than it is The Hydrogen molecule is small and fast to react and requires 2 Oxygen’s to finish the reaction this gas creates high oxygen demand. Free Hydrogen’s are present in exhaust gas.
Hydrocarbons (HC) make the sensor think the mixture is not as rich as it is. HC creates low oxygen demand in a given time frame. HC is hard to react so that’s seen by the cell as low oxygen demand. The cell simply can’t react the HC before its moved on.
Carbon Monoxide (CO) only takes 1 Oxygen to fully combust and is faster than HC but not as fast as H. The CO content is a very good indicator of AFR, but it’s like the other gases in that some engine combinations can produce unrelated amounts.
In summary it is the balance of the gases and their particular response that creates the signal and when people modify engines to various performance levels the combination is not always going to arrive at a normal; whatever that is, burn. It’s when this happens that the WB falls a long way short of a 4 or 5 gas exhaust analyser. The gas bench displays the gases and that enables the tuner to identify specific areas of burn sequence, you can’t do that from AFR information only. WB's are great time savers but they can equally be great performance by-passers. It’s not the WB's fault as such; its peoples lack of understand of their limitations. A wise move is to tune with a gas bench and a WB, when the two are in agreement you can be confident that the WB will serve you well for future atmospheric adjustments.
-Using a WB on Alcohol
The sensor works the same way – in a basic way it is independent of the fuel - except there are differences in the way it responds to different fuels. This is because it has to first catalyze the unburned gases in the exhaust, and then make the measurement of oxygen demand. If you change the fuel of the engine you change the engines ability to condition the fuel prior to ignition, therefore altering the combustion efficiency of the engine - you change the catalytic loading requirement of the lambda sensor. Basically unless the sensor can complete the oxidation reactions first, it will be biased leaner than truth. This is fuel independent - strictly an issue of how much unburned gases (H2, CO, HC) there are in the exhaust gases.
-Wideband sensor verses a Gas bench and how that relates to combustion efficiency improvement.
The wideband (WB) sensor is fast and capable of determining a result for an individual combustion event of an engine even when mounted in the exhaust collector. But its limitation is that the result is only the Lambda of the gases. This lambda is converted to an Air fuel ratio (AFR) by multiplying the Lambda by the stoichiometric value of the fuel used. Some people prefer to use Lambda others use AFR it doesn’t matter, whatever you’re happy with will do.
A gas bench is actually a gas species (species meaning type of) measurement tool. There are different ways of measuring gas concentrations and species etc but the method used for automotive exhaust consists of 2 processes. A gas bench for auto use has an infra red sensor that measures the Carbon dioxide (CO2) the Carbon monoxide (CO) and the Hydrocarbons (HC) present in the exhaust. There is another device called a chemical cell that measures the Oxygen (O) and a different chemical cell that measures the level of Oxides of Nitrogen (NOx). The gas from the exhaust is firstly passed through a water trap (to prevent condensation occurring inside the unit) and then filtered to remove carbon so the sensors are protected from soot buildup. There is a pump to draw exhaust gases into the unit. The gas transport time to achieve this slows down the reaction time of a gas bench. So the hose length etc all affects the sampling time of a gas bench. In addition to that there is the actual sampling time inside the infra red chamber and the chemical cells. A typical gas bench can sample HC, CO and CO2 in less than 2 seconds and NOx and Oxygen in around 4 to 5 seconds. If the gas concentrations are not moving around much it is faster than that. You have to add the gas transport time to those specifications. So they are pretty good but not capable of measurement on a per combustion event basis like a WB is. Not all WB controllers are capable of that performance either but one brand is.
So why am I talking about a WB verses a gas bench? It’s because there is a big difference in the usefulness of the information given by these machines. The WB only reads out the AFR, the gas bench reads AFR as well, but it also shows you what that AFR is made up of.
Knowing the AFR in the engine is only one aspect of tuning. It is common place to read in magazines and even some tuning or engineering sites that setting the AFR of the engine is the only thing you need to do. What is important is to improve the combustion efficiency of the engine not just set an AFR. By reading the individual gases present in the exhaust, the gas bench information can be used to determine combustion efficiency whereas the WB can’t tell you that. AFR is not an indicator of combustion efficiency.
-What influences combustion efficiency?
In order to answer that question its best to explain how the preparation for combustion is important. If the preparation is poorly done then the combustion efficiency suffers.
When an engine is running, the fuel is delivered to the intake as a liquid. Liquids do not burn so the liquid fuel has to be converted into a gas in order to burn. However that is not the only consideration, the ratio of gas to oxygen has to be within the combustible range for the fuel gas present as well. To understand what combustion efficiency is its best to view the engine as an energy transmission machine. Essentially that’s what engines are, some people say they are air pumps but that’s never going to explain how they release energy from fuel. You can pump air all day but its no use unless there is fuel combusted so really its best to understand how energy is moved etc.
I will start from the beginning of the process; this is hard to do in a simple document so bear with me. You may need to read and absorb information by a paragraph at a time etc.
Gasoline or petroleum contains potential energy in the form of chemical energy. It is stored as a liquid and starts to vaporize when subjected to approximately 39C; it will be completely converted to gas form at approx 220C. So why is something that is a liquid requiring such a wide range of temperature to vaporize it? It’s because petrol is not just one chemical, its around 480 to 520 different chemicals. The exact number of chemicals contained in petrol is not even known by the petroleum refiners. Approximately 20 of them are unknown Hydrocarbons that scientists cannot isolate yet. The first chemicals to boil into a gas are the ones that boil at low temperature, these are called light Hydrocarbons. The ones that boil at high temperature are called heavy Hydrocarbons. There are no strict rules of what temperature determines what classification. If you just conceptualise that light HC’s turn into a gas first you will be on the way to understanding a very important point about combustion. This will be revealed later on.
The engine has to turn the liquid fuel into gas in order to burn it; there are a few things that determine the energy necessary to achieve gas.
• Different chemicals require different amounts of energy, and there are many fuels to choose from, they are all different but street fuels are regulated by government to have vaporization characteristics within a narrow window. Race fuels are not regulated. You can inquire with race fuel companies via the net and download data of their various fuel mixture distillation points. This data is given as the temperature necessary to boil away a mass fraction of the fuel. Here is an example and a link to VP fuels. http://www.vpracingfuels.com/spec_sheets.asp
Distillation: for VP fuels C12 (C12 means the fuel has an average carbon molecule length of 12 carbon atoms linked together)
10% evap @131.0°F (this means if the fuel is held at 131F eventually 10% of it will vaporize into a gas but the rest of it will stay there. When you try to start an engine and its cold, these HC’s help a lot, without them starting is difficult, as methanol is )
50% evap @194.0°F
90% evap @228.0°F
E.P. @233.3°F this is the temperature needed to fully vaporize all the fuel in the test tube. What is interesting is that most race fuels vaporize completely at around this temperature but unleaded street fuel requires 420F. Makes you think doesn’t it. It’s much harder to vaporize street unleaded than it is racing fuel.
• Fuel mass; as the fuel mass in the cylinder increases (like when you add the extra nitrous fuel load) more energy is required to vaporise it. If this energy is not increased then less fuel mass will be gas at any equivalent point in time. This is just like it takes longer to boil a kettle that’s full compared to half full.
• Surface area of liquid exposed to energy input. If the fuel is in large droplet form it has a higher mass to surface area ratio and it takes longer to vaporize the fuel. The liquid droplet absorbs heat via its surface area and the surface area of a sphere compared to the volume of a sphere gets better for heating the volume of the sphere if the sphere is smaller. A really tiny drop has a high surface area and hardly any mass so it increases temperature faster.
There are 2 ways you can change the state of the liquid fuel, one is to heat it up and boil it away, and the other is to place it under less pressure. Heating it up is obvious, that’s what compression and water temperature and exhaust gas retention etc does, but the inlet manifold vacuum is less understood. The boiling point of a liquid is affected by the pressure exerted upon it. For fuels etc this is measured as the vapour pressure or Reid index. The higher the vapour pressure the more atmospheric pressure is required to maintain a liquid state. When you run along at part throttle the manifold vacuum is high meaning the absolute pressure in the manifold is low. This low absolute pressure lowers the boiling temperature of the fuel chemicals and more of the fuel is vaporised within the intake manifold. This is important to get right because there is less compression heat in a part throttle engine. The fuel chemists do that blending for us. For a full throttle engine its still important, if the carby or injector throttle body is too large not enough vacuum is generated in the inlet manifold and virtually no vaporisation occurs there so all of it has to be done in the compression. Of course there is radiant heat etc and heat picked up by contact with the engine surfaces and left over exhaust gases in the chamber, it all helps. That’s what an engine is; it’s an energy input balance machine.
When thinking about your engine you must realise that it’s just a machine that is inputting energy into liquid fuel in order to create the most effective amount of fuel gas possible at ignition time. The AFR is delivered to the engine in liquid form but the engine has to convert it to use it. Here is an important point, at the point of fuel ignition the gas concentration is not always 100% of the delivered AFR.
By this point you should be thinking about how to find out what the gas concentration at ignition is. It’s easy, you pull the sparkplugs out. If you use a WB or a gas bench and either tells you that the AFR is in the 12’s for instance then the plug should have tan coloured deposits on it. But many engines will actually have fairly clean porcelain even though the overall cylinder is rich. (Unleaded street legal fuel with a rich AFR actually tans the same as the old days when you vaporise it correctly around the spark plug. There is no difference in sparkplug reading with unleaded fuel, the only difference is tuners don’t generally get it to vaporise as early in the time period of the engine).
Now this is where the WB fails and the gas bench shines. What is happening is the WB is reading the end of the burn just as the gas bench is, they are after all both reading the exhaust gas but the WB just says its running at 12.8 AFR for instance. The gas bench will calculate the same answer (usually) because it’s sensing the same gases (remember the article about WB sensors and how they work) but you can look at the screen and see that there is maybe oxygen present. Hang on a minute here, why is there oxygen present when the chamber is rich? There is oxygen present because the environment around the sparkplug is not fully vaporised at ignition time. That is why you get white plugs on rich engines. Its actually lean at ignition time, its not a lean cylinder as an overall package but what burns is only the gas and if there is lean gas mixture it will burn lean then use the heat from that burn to gas the remaining fuel and then it burns rich at the end of the burn sequence. Doing that is not the way to maximum power.
IF it was fully vaporised at ignition time there would be insufficient oxygen present around the sparkplug and the initial flame kernel would be rich resulting in unburnt carbon deposits.
CO is partly combusted carbon, the final stage of carbon combustion is CO2 but it requires sufficient heat and also oxygen supply to achieve CO2. If the heat is low or there is not enough oxygen to go around then CO is the result.
HC is a general term for any molecule with Hydrogen attached still to a Carbon. They can be any type of species. HC is unused fuel gas molecules.
O2 is unused Oxygen. Low O2 levels indicates good usage of the available Oxygen.
NOx is Oxides of Nitrogen a general term of 4 different combinations of Nitrogen Oxide that are possible. NOx is formed at very high combustion temperatures and can reliably used to indicate peak combustion temperature. Things melt and detonate when NOx goes too high (like over 1000ppm as a general guide).
Here are some examples of different gas readouts but they are all the same AFR.
(ppm means parts per million.)
CO% CO2% HC ppm O2% NOx ppm AFR
6.7 7.0 3000 6.5 1900 12.8
6.7 9 1000 3.1 2500 12.8
4.3 11.75 150 0.1 90 12.8
Note that the first burn has a lot of HC unused and there is also unused oxygen. These gases are a real world EFI engine I actually measured. The WB read that mix as 11.8 so what use it that. If you based your tuning decision on the WB data you would lean the engine off but looking at the gas bench data you can see that there is significant unused oxygen and lots of HC as well so you can tell that the engine has a vaporisation problem. Investigating what is causing that will fix the engine and it will be much more powerful. There is no way you could run Nitrous on that engine tune, it would melt everything. It’s obvious there are lean places in the cylinder, that’s why there is all that oxygen. There is plenty of fuel in there so what’s adding more going to do? Yet the WB said alls OK go and add the Nitrous. Makes you think doesn’t it.
Note the second set of gas readings, there is more CO2 and less HC and its using more of the oxygen. It’s still the same AFR and the WB read that as 12.8. It read the same as the gas bench in that example. It was the same engine, same injector duration, just more advanced timing. But the timing didn’t fix the entire problem so you have to dig deeper. Incidentally there was only 1 KW difference in power output between gas set one and two. So the dyno was no help.
Note the third set of gases, there is low CO high CO2 hardly any HC and just about all the oxygen is used and there isn’t much NOx. This is a good rich burn of a good efficient engine.
Here is another set of examples of 14.7 AFR
CO% CO2% HC ppm O2% NOx ppm AFR
0.18 13.85 0 0.13 90 14.7
2.0 12.0 500 2.0 3000 14.7
2.37 11.5 2600 6 0 14.7
The first set of gases is a near perfect 14.7 burn, very hard to achieve
The second set is a poorly vaporised 14.7 burn. Note the HC is up and the CO is up and the oxygen is not being used and there are high levels of NOx. The NOx indicates its high combustion temperature somewhere in the chamber and the CO is indicating there is a cold area somewhere in the chamber. So the gas bench allows us to deduce that the heads need to come off and get fixed. A WB would just tell us the AFR is 14.7 and you should be happy mate.
The third set of gases tells us that its poorly vaporised incorrectly distributed and too retarded. There is CO because it’s retarded and even more HC because it’s retarded and lots of wasted oxygen because it’s retarded and no NOx because it was a very cold burn. We can even perform an oxygen balance on the gases and discover that there is a lean misfire so it’s poorly distributed. It’s even possible the engine has the wrong cams in it. You can’t do any of that with just a WB.
Hopefully you should be realising that you need to see the gas relationships rather than just the AFR. But you must sample the gases before a catalytic converter, not after it as the cat alters the efficiency of the combustion. That’s what cats are for to fix the poorly combusting cylinder so that the government is happy. Cats do not alter the AFR readout of either a gas bench or a WB but they alter the combustion efficiency of the engine out gases.
It’s common for tuners to think about the engine only from the power stroke view but the power stroke is dependent upon the conditioning done on the fuel prior to ignition. It is better to correct the distribution and vaporisation of the fuel prior to ignition than it is to try and mix it after it has ignited.
So for Nitrous tuning it is important to add the additional fuel in a way that promotes distribution and vaporisation within the cylinder. It’s hard enough getting a rich power mixture vaporised on some engine combinations let alone having to deal with the additional fuel for nitrous as well.
A good rule to observe is that adding Nitrous or supercharging a bad engine combination will not fix the engine’s problems. It only makes it worse and at best limits the power addition without blowing the engine. Its not all doom and gloom as there are many successful tunings of nitrous and supercharging etc being done but it’s when racers decide to push the boundaries and go one better than the other guy that you have to start investigating engines in this way.
Bruce Robertson
ich bin noch nicht ganz durch aber wollte es euch auch nicht vorenthalten.
Quellen:
http://www.smartcarby.com
http://forum.nitrous-advice.org/viewtop ... f=5&t=4106
-How a Wideband sensor works and what its limitations are.
The common called ‘oxygen sensor’ is not in fact an oxygen sensing device. Some people are mislead by this incorrect labeling into thinking that the sensor measures the level of oxygen in the exhaust and determines the air fuel ratio based upon that. The mislabeling is even perpetrated by Bosch and other manufacturers. The simple narrow band sensor is virtually useless for measuring the air fuel ratio of an engine other than to achieve a switching based upon stoichiometric conditions present in the sensor. There is no place for a narrow band sensor in performance tuning. However a narrow band can be used in some situations. Some people spend considerable time correlating narrow band sensors to a gas bench and can generate useful data applicable to their engine but that is not generally done.
Wideband sensors are very popular for measuring the air fuel ratio (AFR) of an engine. Many tuners use them to set a desired AFR that they are confident will achieve a desired outcome for a customer. The wideband sensor is modern, fast, accurate and far cheaper than investment in a 4 gas or 5 gas analyzer. Most people are taught that a wideband sensor can accurately display the AFR of the engine. A WB sensor can easily read leaner than truth and in fact even read richer than the truth. The conditions to do this happen quite often on hotrod engines that have not been developed sufficiently. For the interest of Nitrous tuners, a leaner than truth reading would seem not to be a problem, as one may argue that running the engine excessively rich is OK to do. And the other argument would be that a richer than truth reading should be avoided at all times. If you are tuning successfully at the moment then all is well until you decide to increase the power. The point of destruction isn’t obvious until it happens. By understanding how a wideband sensor (WB) can incorrectly read the AFR and performing some simple tests you can determine if the WB is accurate for your tuning purpose. First we need to understand what this number called Lambda is.
Lambda is a number that represents the completeness of the combustion of reactants. The reactants within the chamber are simply any combustible gases and the Oxygen from the atmosphere or the nitrous. Lambda 1 is complete combustion of all reactants, which means all fuel molecules have been fully oxidized and all free Oxygen has been consumed by the reactants. Achieving this state of complete reaction has nothing to do with the type of fuel used. Any fuel can be completely reacted and all the Oxygen can be used at the same time if the fuel and Oxygen relationship is correct. For instance you could fill a container with just Hydrogen gas and Oxygen gas in the exact proportions of 2 Hydrogen molecules to every Oxygen molecule, there is no air involved, and then you can say that the mixture in the container is H2O which is of course water. But it isn’t water yet is it, we haven’t combusted it yet. If we let the container sit at room temperature for 30,000 years there will eventually be no gases present in the container as all the gases will have slowly combusted into water. If we heat the mixture in the container up we don't have to wait 30,000 years to get a drink. As we heat the gases up the reaction happens much faster. By the time we heat it up to 500 degrees or so it’s reacting in milliseconds. The reason why I'm talking about heat and time will become apparent later.
The mixture we filled the container with was a Lambda 1 mixture. If we filled the container with too many Hydrogen’s, for example say we put in 5% too many, then we filled the container with Lambda 0.95. If we sampled some of the gases that we put into the container with a WB sensor, the sensor will use a destructive process called catalytic reaction to measure what ratio the reactants to Oxygen is. The WB sensor will read Lambda 1 for our perfect chamber mix and Lambda 0.95 for our rich mixture example. The wideband sensor can only measure the lambda of gases. There is no need for conventional combustion of our gases to occur in order for the WB sensor to work.
There is no need for air or anything else to confuse us. So thinking in terms of lambda enables us to think independently of fuel type or air quality.
How an AFR number is arrived at, is a simple multiplication of the Lambda value times the stoichiometric value of whatever fuel you are using. However finding out the true stoich value of a commercial gasoline is just about impossible so we just use 14.7. It doesn’t matter if you use Lambda or AFR in the concept of engine tuning, but Lambda has advantages for ECU tuning and jet changing.
If we want to alter the mixture based upon thinking in Lambda its easy because Lambda 0.95 is 5% rich and to alter it to lambda 1 requires 5% smaller jetting flow or injector timing etc. Whatever you’re using to tune, it is easy to calculate changes.
If we place the WB sensor in a container of only CO2 (Carbon Dioxide) gas it will read lambda 1. This is because CO2 is the completely Oxidized state of carbon. There are no molecules that can be reacted with oxygen.
If we fill the container with CO (carbon monoxide) the WB cell will transfer oxygen’s from the atmosphere onto the surface of the chamber as the catalyst reacts the Oxygen with the CO until all the CO molecules have turned into CO2. The number of Oxygen molecules to complete the process of CO to CO2 is counted by the electronics and a Lambda value is displayed. The Lambda value will be infinity rich as the oxygen transfer rate will be the maximum the cell can produce.
If we fill the container with Hydrogen, the same Oxygen ion transfer process occurs and the catalyst will do the same thing etc. The electronics will ‘count’ the number of Oxygen molecules transferring that are necessary to react the Hydrogen to its fully oxidized state (H2O--water) and display a Lambda value. This counting process is very dependent upon time required to react the gases.
Gasoline and other fuels are Hydrocarbons, which means they consist of Hydrogen and Carbon. If we fill our container with Hydrocarbon gas the WB will count the Oxygen’s needed to fully react the HC gas just the same as it did for the earlier examples.
The catalyst in the sensor needs to be heated to a constant temperature in order to achieve a fast enough reaction. WB sensors have a heater element in them to maintain the temperature to 850 Celsius for correct reaction rate etc of the catalyst. Basically when it’s hot, it happens fast enough to be of value as a measurement, so if the heater circuit can’t heat up the cell enough or the cell gets too hot it’s not accurate and you get an error.
So a power supply problem or an exhaust gas condition that is preventing the cell from being within its correct operating temperature range causes the system to be no longer accurate.
Now let’s fill our container with HC’s and Oxygen so that there is a 5% rich mixture in it and see what’s happened after some time to complete the reactions. The H’s have consumed the necessary Oxygen’s to turn into H20 (water) and some of the carbons have completed both stages of reaction to make CO2 and then we have run out of Oxygen within the container. That results in some of the carbons only making it to the CO stage. It may have even resulted in some of the Hydrogen’s not being combusted so there would be H’s left over in that case. In order to complete these reactions we have to supply oxygen from the atmosphere. So now the electronics measures the charge from the oxygen ion transfer that is supplying Oxygen’s to the catalyst to enable the completion of the reaction of the remaining CO’s and H’s to the final reacted state of CO2 and H2O. Once the reactions have been completed there will be stoichiometric conditions within our container. We had to supply some extra oxygen to the mixture by grabbing the Oxygen’s from the atmosphere to do it and that is how we calculated that we were 5% rich. The point is that technically the WB cell alters the air fuel ratio of the environment. This effect is insignificant in a moving environment like a exhaust pipe. However in an engine we don’t know what the actual Oxygen present in the exhaust is. Any available Oxygen’s in the exhaust gas gets used without being counted and that’s where the WB cell falls down, it can’t measure the combustion efficiency. Inefficient combustion has HC’s and H’s and O’s left over and therefore an inefficiently combustion engine is read as either richer or leaner than truth.
Why is there all this talk about combustion efficiency? In an engine with poor combustion efficiency the gases present in the exhaust doesn’t follow the science books, you can have an engine that is rich but it can have Oxygen in the exhaust, or you can have an engine that is lean and it can have unburnt fuel in the exhaust. These engines give results opposite to the science books. The factors that control combustion efficiency are not always optimum in all areas of the chamber and you get fuels exhausting from one area and Oxygen’s from another.
If we had a mixture in our container that was 5% excess oxygen (lean) the sensor will complete all the oxidization of the fuel molecules and then discover that there are still Oxygen’s left over. So only after it’s completed the reactions will it remove the Oxygen’s from our container and tell us what it did by displaying lambda 1.05. In an engine with poor combustion efficiency a lean mixture can contain unburnt fuel so the catalyst has to react that first with whatever available oxygen present in the exhaust gas and then the electronics removes any excess Oxygen’s and counts them.
The combustion within the cylinder is far from consistent throughout the cylinder. If it was a perfect world and we input a perfectly homogenized stoichiometric ratio there would be lambda 1 mixture everywhere even viewed on a molecule by molecule basis. If we use a simple HC compound such as Methanol and that’s all that we put in there and somehow we stop engine oil etc contaminating it, we can surround that methanol molecule with the right number of oxygen molecules to completely combust it. If we did that with every molecule the energy release would be as per theory. The energy release would occur at the same rate throughout the mixture. In practice we can’t do that because we don’t have time to perfectly mix the fuel and air and we don’t have all the other factors necessary to achieve that perfect combustion.
Let’s just look at Methanol for a little bit, it’s a CO molecule surrounded by H’s. In fact the Carbon element has 3 Hydrogen’s attached to one side of it with 1 Oxygen attached on the other side of it. To that Oxygen is attached another Hydrogen. The way that this simple molecule reacts is not simple. There are many processes or methods of reactions, no theory explains it all. Mankind cant work it all out yet so don’t ask me. Suffice it to say for our purpose here, that a reaction sequence takes place with an unknown order or reason. But we know the end result of a perfect reaction with certainty. The 4 Hydrogen molecules of the methanol will create 2 water molecules and the Carbon and oxygen will eventually end up as CO2. If we don’t get this to happen in the chamber then the catalyst in the WB will finish the job for us and tell us some sort of answer as to how it did it. That answer is of course told to us in the form of Lambda.
The inconsistencies throughout the chamber mean that somewhere there is complete combustion of the Methanol and somewhere there isn’t. So the sensor has to receive the incompletely combusted molecules and complete it. That isn’t always going to happen is it? Any liquids present can’t be combusted until the liquid is turned into gas and we can only react the actual sample within the cell. Anything that bypasses we miss out on.
Why do you get soot in the tail pipe? It’s because Carbon hasn’t been reacted in the cylinder and the temperature conditions in the exhaust pipe are too cold to complete the carbon reaction that you missed in the cylinder. So soot affects the lambda readout.
Even if we forget about that, we then have to react whatever incomplete reactants there are in contact with the catalyst. Any Hydrogen entering the chamber that is not attached to another element is the easiest to react so they start the process first and demand the oxygen first. Any CO in contact gets reacted in a longer time period than the H’s. Any HC’s take longer again. Etc. The more complex the molecule the longer it takes to break it down and react it. In the mean time the environment is constantly changing, there is even oxygen hitting the catalyst. It’s a mess in there.
Any oxygen in the exhaust pipe gets on the surface of the catalyst and reacts with any molecule present that is not completely reacted and isn’t counted by the electronics because it’s not transferring in or out of the cell. So as Oxygen’s arrive they react within this constantly changing environment.
The reality of all this is that you can’t determine the state of the reaction completion within the engine cylinder, which is where you really want to know how completed it is. You can’t tell that from a reaction that is completed due to catalytic action. Assuming a perfect world all you can determine via catalytic reaction to completion is how far from stoichiometric the mixture in the chamber was.
Now this is the important point to understand; a WB tells you nothing about the burn within the cylinder. It actually doesn’t even need a burn. When you get a misfire it reads it as lean but that’s because it hasn’t had time to react to completion. In reality there is a usable AFR in a misfiring cylinder, as an example say its EFI, if the injector opened correctly and the other factors were correct then fuel was given correctly to the cylinder but if the arc failed to ignite the cylinder with sufficient energy the burn will be far from complete or not even a burn at all in the case of a electrical short and yet the WB reads LEAN. No its not, the cylinders not lean, it just didn’t ignite that’s all. The reason an ignition miss shows up as a lean spike is that the HC chains are harder to oxidize. They go unnoticed, and that is what causes the ‘lean’ indication. Remember that the wideband uses the O2 that’s present in the exhaust gas firstly. If the wideband adequately completed the HC reaction at the sensor interface – most of the lean error would go away. Consider that a misfire is not a change of AFR it’s just a non combustion within the cylinder and the wideband is a combustion device so if the wideband was able to complete the misfired gas output it wouldn’t read a lambda change. Hopefully this may make it clear as to the limitations of the WB sensor and give you some concept of its time limitations too. A miss fire is a time fault.
-Testing WB accuracy.
If the combustion efficiency is poor then altering the ignition timing can alter the Lambda readout of a WB. If the cylinder is insufficiently vaporized the combustion has to progress through more stages to get to completion. Altering ignition timing alters the completion of the combustion within the cylinder. Sometimes advancing the ignition timing improves completion sometimes it reduces completion. So when you do a test and alter the timing and the WB reads the Mixture has changed and you know you haven’t changed the mixture, what you are really seeing is a change in the reaction completion loading of the cell. You have no way of knowing what gas changed, to do that you have to use a gas bench or better still, a gas spectrometer.
Now we put together a hotrod engine and start it up. It’s got more power than the stocker we started with, as it has a higher fuel and air flow rate. We install a WB and measure the Lambda and we decide we don’t want it so rich so we change the fueling of the engine. Wrong--- we didn’t check the combustion efficiency first, we didn’t find out how complete the combustion was. That’s why you can sometimes change jets, etc. and things don’t change as they should. You can spend a long time tuning with a WB and never see what is really happening inside the cylinder.
The main factors that control combustion efficiency within the cylinder are vaporization, homogenization and distribution. The lambda measurement of the catalytic reaction is only based upon gases present and time available.
How different gases cause problems with the WB sensor
Hydrogen (H) makes the sensor think the mixture is richer than it is The Hydrogen molecule is small and fast to react and requires 2 Oxygen’s to finish the reaction this gas creates high oxygen demand. Free Hydrogen’s are present in exhaust gas.
Hydrocarbons (HC) make the sensor think the mixture is not as rich as it is. HC creates low oxygen demand in a given time frame. HC is hard to react so that’s seen by the cell as low oxygen demand. The cell simply can’t react the HC before its moved on.
Carbon Monoxide (CO) only takes 1 Oxygen to fully combust and is faster than HC but not as fast as H. The CO content is a very good indicator of AFR, but it’s like the other gases in that some engine combinations can produce unrelated amounts.
In summary it is the balance of the gases and their particular response that creates the signal and when people modify engines to various performance levels the combination is not always going to arrive at a normal; whatever that is, burn. It’s when this happens that the WB falls a long way short of a 4 or 5 gas exhaust analyser. The gas bench displays the gases and that enables the tuner to identify specific areas of burn sequence, you can’t do that from AFR information only. WB's are great time savers but they can equally be great performance by-passers. It’s not the WB's fault as such; its peoples lack of understand of their limitations. A wise move is to tune with a gas bench and a WB, when the two are in agreement you can be confident that the WB will serve you well for future atmospheric adjustments.
-Using a WB on Alcohol
The sensor works the same way – in a basic way it is independent of the fuel - except there are differences in the way it responds to different fuels. This is because it has to first catalyze the unburned gases in the exhaust, and then make the measurement of oxygen demand. If you change the fuel of the engine you change the engines ability to condition the fuel prior to ignition, therefore altering the combustion efficiency of the engine - you change the catalytic loading requirement of the lambda sensor. Basically unless the sensor can complete the oxidation reactions first, it will be biased leaner than truth. This is fuel independent - strictly an issue of how much unburned gases (H2, CO, HC) there are in the exhaust gases.
-Wideband sensor verses a Gas bench and how that relates to combustion efficiency improvement.
The wideband (WB) sensor is fast and capable of determining a result for an individual combustion event of an engine even when mounted in the exhaust collector. But its limitation is that the result is only the Lambda of the gases. This lambda is converted to an Air fuel ratio (AFR) by multiplying the Lambda by the stoichiometric value of the fuel used. Some people prefer to use Lambda others use AFR it doesn’t matter, whatever you’re happy with will do.
A gas bench is actually a gas species (species meaning type of) measurement tool. There are different ways of measuring gas concentrations and species etc but the method used for automotive exhaust consists of 2 processes. A gas bench for auto use has an infra red sensor that measures the Carbon dioxide (CO2) the Carbon monoxide (CO) and the Hydrocarbons (HC) present in the exhaust. There is another device called a chemical cell that measures the Oxygen (O) and a different chemical cell that measures the level of Oxides of Nitrogen (NOx). The gas from the exhaust is firstly passed through a water trap (to prevent condensation occurring inside the unit) and then filtered to remove carbon so the sensors are protected from soot buildup. There is a pump to draw exhaust gases into the unit. The gas transport time to achieve this slows down the reaction time of a gas bench. So the hose length etc all affects the sampling time of a gas bench. In addition to that there is the actual sampling time inside the infra red chamber and the chemical cells. A typical gas bench can sample HC, CO and CO2 in less than 2 seconds and NOx and Oxygen in around 4 to 5 seconds. If the gas concentrations are not moving around much it is faster than that. You have to add the gas transport time to those specifications. So they are pretty good but not capable of measurement on a per combustion event basis like a WB is. Not all WB controllers are capable of that performance either but one brand is.
So why am I talking about a WB verses a gas bench? It’s because there is a big difference in the usefulness of the information given by these machines. The WB only reads out the AFR, the gas bench reads AFR as well, but it also shows you what that AFR is made up of.
Knowing the AFR in the engine is only one aspect of tuning. It is common place to read in magazines and even some tuning or engineering sites that setting the AFR of the engine is the only thing you need to do. What is important is to improve the combustion efficiency of the engine not just set an AFR. By reading the individual gases present in the exhaust, the gas bench information can be used to determine combustion efficiency whereas the WB can’t tell you that. AFR is not an indicator of combustion efficiency.
-What influences combustion efficiency?
In order to answer that question its best to explain how the preparation for combustion is important. If the preparation is poorly done then the combustion efficiency suffers.
When an engine is running, the fuel is delivered to the intake as a liquid. Liquids do not burn so the liquid fuel has to be converted into a gas in order to burn. However that is not the only consideration, the ratio of gas to oxygen has to be within the combustible range for the fuel gas present as well. To understand what combustion efficiency is its best to view the engine as an energy transmission machine. Essentially that’s what engines are, some people say they are air pumps but that’s never going to explain how they release energy from fuel. You can pump air all day but its no use unless there is fuel combusted so really its best to understand how energy is moved etc.
I will start from the beginning of the process; this is hard to do in a simple document so bear with me. You may need to read and absorb information by a paragraph at a time etc.
Gasoline or petroleum contains potential energy in the form of chemical energy. It is stored as a liquid and starts to vaporize when subjected to approximately 39C; it will be completely converted to gas form at approx 220C. So why is something that is a liquid requiring such a wide range of temperature to vaporize it? It’s because petrol is not just one chemical, its around 480 to 520 different chemicals. The exact number of chemicals contained in petrol is not even known by the petroleum refiners. Approximately 20 of them are unknown Hydrocarbons that scientists cannot isolate yet. The first chemicals to boil into a gas are the ones that boil at low temperature, these are called light Hydrocarbons. The ones that boil at high temperature are called heavy Hydrocarbons. There are no strict rules of what temperature determines what classification. If you just conceptualise that light HC’s turn into a gas first you will be on the way to understanding a very important point about combustion. This will be revealed later on.
The engine has to turn the liquid fuel into gas in order to burn it; there are a few things that determine the energy necessary to achieve gas.
• Different chemicals require different amounts of energy, and there are many fuels to choose from, they are all different but street fuels are regulated by government to have vaporization characteristics within a narrow window. Race fuels are not regulated. You can inquire with race fuel companies via the net and download data of their various fuel mixture distillation points. This data is given as the temperature necessary to boil away a mass fraction of the fuel. Here is an example and a link to VP fuels. http://www.vpracingfuels.com/spec_sheets.asp
Distillation: for VP fuels C12 (C12 means the fuel has an average carbon molecule length of 12 carbon atoms linked together)
10% evap @131.0°F (this means if the fuel is held at 131F eventually 10% of it will vaporize into a gas but the rest of it will stay there. When you try to start an engine and its cold, these HC’s help a lot, without them starting is difficult, as methanol is )
50% evap @194.0°F
90% evap @228.0°F
E.P. @233.3°F this is the temperature needed to fully vaporize all the fuel in the test tube. What is interesting is that most race fuels vaporize completely at around this temperature but unleaded street fuel requires 420F. Makes you think doesn’t it. It’s much harder to vaporize street unleaded than it is racing fuel.
• Fuel mass; as the fuel mass in the cylinder increases (like when you add the extra nitrous fuel load) more energy is required to vaporise it. If this energy is not increased then less fuel mass will be gas at any equivalent point in time. This is just like it takes longer to boil a kettle that’s full compared to half full.
• Surface area of liquid exposed to energy input. If the fuel is in large droplet form it has a higher mass to surface area ratio and it takes longer to vaporize the fuel. The liquid droplet absorbs heat via its surface area and the surface area of a sphere compared to the volume of a sphere gets better for heating the volume of the sphere if the sphere is smaller. A really tiny drop has a high surface area and hardly any mass so it increases temperature faster.
There are 2 ways you can change the state of the liquid fuel, one is to heat it up and boil it away, and the other is to place it under less pressure. Heating it up is obvious, that’s what compression and water temperature and exhaust gas retention etc does, but the inlet manifold vacuum is less understood. The boiling point of a liquid is affected by the pressure exerted upon it. For fuels etc this is measured as the vapour pressure or Reid index. The higher the vapour pressure the more atmospheric pressure is required to maintain a liquid state. When you run along at part throttle the manifold vacuum is high meaning the absolute pressure in the manifold is low. This low absolute pressure lowers the boiling temperature of the fuel chemicals and more of the fuel is vaporised within the intake manifold. This is important to get right because there is less compression heat in a part throttle engine. The fuel chemists do that blending for us. For a full throttle engine its still important, if the carby or injector throttle body is too large not enough vacuum is generated in the inlet manifold and virtually no vaporisation occurs there so all of it has to be done in the compression. Of course there is radiant heat etc and heat picked up by contact with the engine surfaces and left over exhaust gases in the chamber, it all helps. That’s what an engine is; it’s an energy input balance machine.
When thinking about your engine you must realise that it’s just a machine that is inputting energy into liquid fuel in order to create the most effective amount of fuel gas possible at ignition time. The AFR is delivered to the engine in liquid form but the engine has to convert it to use it. Here is an important point, at the point of fuel ignition the gas concentration is not always 100% of the delivered AFR.
By this point you should be thinking about how to find out what the gas concentration at ignition is. It’s easy, you pull the sparkplugs out. If you use a WB or a gas bench and either tells you that the AFR is in the 12’s for instance then the plug should have tan coloured deposits on it. But many engines will actually have fairly clean porcelain even though the overall cylinder is rich. (Unleaded street legal fuel with a rich AFR actually tans the same as the old days when you vaporise it correctly around the spark plug. There is no difference in sparkplug reading with unleaded fuel, the only difference is tuners don’t generally get it to vaporise as early in the time period of the engine).
Now this is where the WB fails and the gas bench shines. What is happening is the WB is reading the end of the burn just as the gas bench is, they are after all both reading the exhaust gas but the WB just says its running at 12.8 AFR for instance. The gas bench will calculate the same answer (usually) because it’s sensing the same gases (remember the article about WB sensors and how they work) but you can look at the screen and see that there is maybe oxygen present. Hang on a minute here, why is there oxygen present when the chamber is rich? There is oxygen present because the environment around the sparkplug is not fully vaporised at ignition time. That is why you get white plugs on rich engines. Its actually lean at ignition time, its not a lean cylinder as an overall package but what burns is only the gas and if there is lean gas mixture it will burn lean then use the heat from that burn to gas the remaining fuel and then it burns rich at the end of the burn sequence. Doing that is not the way to maximum power.
IF it was fully vaporised at ignition time there would be insufficient oxygen present around the sparkplug and the initial flame kernel would be rich resulting in unburnt carbon deposits.
CO is partly combusted carbon, the final stage of carbon combustion is CO2 but it requires sufficient heat and also oxygen supply to achieve CO2. If the heat is low or there is not enough oxygen to go around then CO is the result.
HC is a general term for any molecule with Hydrogen attached still to a Carbon. They can be any type of species. HC is unused fuel gas molecules.
O2 is unused Oxygen. Low O2 levels indicates good usage of the available Oxygen.
NOx is Oxides of Nitrogen a general term of 4 different combinations of Nitrogen Oxide that are possible. NOx is formed at very high combustion temperatures and can reliably used to indicate peak combustion temperature. Things melt and detonate when NOx goes too high (like over 1000ppm as a general guide).
Here are some examples of different gas readouts but they are all the same AFR.
(ppm means parts per million.)
CO% CO2% HC ppm O2% NOx ppm AFR
6.7 7.0 3000 6.5 1900 12.8
6.7 9 1000 3.1 2500 12.8
4.3 11.75 150 0.1 90 12.8
Note that the first burn has a lot of HC unused and there is also unused oxygen. These gases are a real world EFI engine I actually measured. The WB read that mix as 11.8 so what use it that. If you based your tuning decision on the WB data you would lean the engine off but looking at the gas bench data you can see that there is significant unused oxygen and lots of HC as well so you can tell that the engine has a vaporisation problem. Investigating what is causing that will fix the engine and it will be much more powerful. There is no way you could run Nitrous on that engine tune, it would melt everything. It’s obvious there are lean places in the cylinder, that’s why there is all that oxygen. There is plenty of fuel in there so what’s adding more going to do? Yet the WB said alls OK go and add the Nitrous. Makes you think doesn’t it.
Note the second set of gas readings, there is more CO2 and less HC and its using more of the oxygen. It’s still the same AFR and the WB read that as 12.8. It read the same as the gas bench in that example. It was the same engine, same injector duration, just more advanced timing. But the timing didn’t fix the entire problem so you have to dig deeper. Incidentally there was only 1 KW difference in power output between gas set one and two. So the dyno was no help.
Note the third set of gases, there is low CO high CO2 hardly any HC and just about all the oxygen is used and there isn’t much NOx. This is a good rich burn of a good efficient engine.
Here is another set of examples of 14.7 AFR
CO% CO2% HC ppm O2% NOx ppm AFR
0.18 13.85 0 0.13 90 14.7
2.0 12.0 500 2.0 3000 14.7
2.37 11.5 2600 6 0 14.7
The first set of gases is a near perfect 14.7 burn, very hard to achieve
The second set is a poorly vaporised 14.7 burn. Note the HC is up and the CO is up and the oxygen is not being used and there are high levels of NOx. The NOx indicates its high combustion temperature somewhere in the chamber and the CO is indicating there is a cold area somewhere in the chamber. So the gas bench allows us to deduce that the heads need to come off and get fixed. A WB would just tell us the AFR is 14.7 and you should be happy mate.
The third set of gases tells us that its poorly vaporised incorrectly distributed and too retarded. There is CO because it’s retarded and even more HC because it’s retarded and lots of wasted oxygen because it’s retarded and no NOx because it was a very cold burn. We can even perform an oxygen balance on the gases and discover that there is a lean misfire so it’s poorly distributed. It’s even possible the engine has the wrong cams in it. You can’t do any of that with just a WB.
Hopefully you should be realising that you need to see the gas relationships rather than just the AFR. But you must sample the gases before a catalytic converter, not after it as the cat alters the efficiency of the combustion. That’s what cats are for to fix the poorly combusting cylinder so that the government is happy. Cats do not alter the AFR readout of either a gas bench or a WB but they alter the combustion efficiency of the engine out gases.
It’s common for tuners to think about the engine only from the power stroke view but the power stroke is dependent upon the conditioning done on the fuel prior to ignition. It is better to correct the distribution and vaporisation of the fuel prior to ignition than it is to try and mix it after it has ignited.
So for Nitrous tuning it is important to add the additional fuel in a way that promotes distribution and vaporisation within the cylinder. It’s hard enough getting a rich power mixture vaporised on some engine combinations let alone having to deal with the additional fuel for nitrous as well.
A good rule to observe is that adding Nitrous or supercharging a bad engine combination will not fix the engine’s problems. It only makes it worse and at best limits the power addition without blowing the engine. Its not all doom and gloom as there are many successful tunings of nitrous and supercharging etc being done but it’s when racers decide to push the boundaries and go one better than the other guy that you have to start investigating engines in this way.
Bruce Robertson