| Author | Message | ||
     Xoptimizedrsx |
It’s All About the Sensors Electronic engine management is all about getting accurate information from the engine’s environment and then sending out accurate signals to the control mechanisms of the engine. In a typical system you have the following input sensors: 1. throttle position of the accelerator 2. air temperature in the intake manifold 3. air pressure (or weight of the air) in the intake manifold 4. a measure of the residual oxygen in the exhaust manifold (the O2 sensor) 5. engine rpm 6. camshaft position 7. crankshaft position 8. engine load, usually as a function of manifold vacuum pressure 9. engine coolant temperature 10. about a zillion other things, like “knock” sensing, battery voltage, ambient air temperature, and whether Hoobastank is still on the Billboard top 40. The computer simply takes all this data, compares it to data about optimum spark and fuel volumes for different conditions, and controls how much fuel to inject and when to fire the spark plugs. Modern Mixture and Maps: We know that there is an ideal air:fuel mixture ratio for each type of fuel - for gasoline this is 14.7 weight units of air to 1 weight unit of fuel, a so-called “stoichiometric” mixture. In principle, this is the mixture point at which all fuel injected will combust with all the gases in the ambient air. However the combustion is rarely perfect – an oxygen may never meet a hydrocarbon on the other side of a crowded cylinder of dancing vapours - so a somewhat leaner mixture (around 16:1) will ensure that excess air will gather up all of the hydrocarbons in the fuel and so will produce fewer emissions and better fuel efficiency. On the other hand a richer mixture will ensure that all the air that can be moved into the engine will combine with hydrocarbons in the fuel and so for maximum power you actually want this mixture to be richer - closer to 12:1. Since you’re more concerned with power under hard acceleration and less concerned on the overrun (when you lift the throttle), the computer can target leaner and richer mixture ratios depending, literally, on how hard you stomp on the throttle. Maintaining the ideal mixture is the function of your “fuel map” and it determines how much fuel the ECU will inject during each cycle of the engine. Essentially the map is a compromise to give you power when you want it, keep emissions as low as possible, and all the time keep the mixture within safe parameters to avoid damage to the engine. How much Air? Before it can determine how much fuel to squirt in, the main challenge for the ECU is to know what weight of air is in the intake manifold and getting sucked (normally aspirated) or pushed (turbo or supercharged) through the intake valve. Modern ECUs measure this in one of two ways: either by taking the pressure and temperature of the air in the manifold and calculating Manifold Absolute Pressure (MAP), or by measuring the weight of the air (Mass Air Flow or MAF) as it moves into the intake past a heated wire filament across which the ECU can measure electrical resistance. Older systems used other means, including various flaps and pulleys (I’m not kidding) to measure airflow. Now most production cars use MAF, although I use MAP in my rally car and some manufacturers, including Subaru, are moving back towards this. How Much Fuel? The main mixture job of the ECU is to control the injectors. Fuel injectors are little spray nozzles that can be snapped open and shut by an electrical charge: the amount of time they’re open, combined with the pressure of the fuel coming in and the size of the injector, determines how much fuel will be injected. In “multi-point” fuel injection systems there’s an injector for each cylinder, while in more basic systems (“Throttle Body Injection”) there’s just one for the whole intake. So what determines how long the injectors stay open? The ECU senses the weight of the air, the rpm of the engine, and compares this to the data “map” it contains about ideal mixture for each load-rpm condition (see “base fuel delivery” graph). It then decides how long to open the injectors for and sends an electrical signal to them for the right amount of time. At idle the injectors open for just a millisecond – long enough to keep the engine ticking and no more. But mash the throttle and they go towards the maximum “duty cycle” – theoretically they could stay open 100% of the time but generally are engineered to give full throttle by being open about 80% of the time – that is, 80% of the time that it takes the crankshaft to go around twice. Thus is mixture in the modern engine determined. When to Spark? As engine speed increases, you need to ignite the mixture in the combustion chamber progressively earlier to get most efficient ignition. Note that the “explosion” in each cylinder is really a very fast fire and that the “fire front” moves through the mixture until burning is complete. The earlier you can spark the mixture, the more power you’ll get as there will be more time to burn all the vapour before the exhaust valve opens. But if you fire the spark too early, you will end up with the explosion completing too soon and you get an uneven leading edge to the burn – effectively a separate explosion - called a “detonation.” This may allow a “hot spot” of the explosion to hit the top of the piston (normally it is protected by a barrier of other gases at the edge of the explosion), and this is what a “knock” in the engine is. All engines knock somewhat, actually, but if you don’t keep it within an acceptable level you will burn a hole right through the piston. Don’t ask me how I know. Higher octane fuel, by the way, burns more consistently at the leading edge, which is why for most race applications we use fuels with an octane above 100 while you might use 87 in your street car, and why we can advance the timing and run higher compression in the race cars. Lead additives in the fuel prevent detonation, but as you know this is mildly toxic and so we don’t use it any more. Of course blowing engines is mildly toxic, too. So for ignition, you have a second “map” that determines when to fire the spark plugs, based again on engine rpm and load (see “base ignition timing” map). In many four cylinder engines you have two coil packs and each has two plugs attached to it – when the ECU determines it’s the right time to fire one of the plugs it sends current to one of the coil packs and two plugs actually fire, although one is redundant and has no fuel to ignite (creating a “waste spark”). It’s just cheaper and more reliable than having four individual coils or (horror!) a distributor. Other Stuff: So the fuel delivery and spark timing maps are the real essence of what the ECU is doing for you all the time. Thousands of times per second the ECU is calculating how much fuel to deliver and when to spark. But that’s not all. For example, it knows when the engine is cold and it delivers more fuel and adjusts the spark for smooth running until the engine warms up. It also knows if the engine gets too hot and it can drop into “limp mode” to get you home by cooling the explosions with a richer mixture and later spark. It also has two other very important sensors that adjust the fuel and spark maps constantly. The first is the “oxygen sensor” in the exhaust manifold, which sends information about how much oxygen is left over after the explosions in the engine. The computer takes this data and constantly adjusts the fuel mixture to keep this near maximum efficiency (14.7:1) or maximum power (12:1) depending on settings. Running in this mode is known as “closed loop” as the ECU is constantly adjusting itself to get the desired exhaust. There are a couple of interesting consequences to this: it self-adjusts for different qualities of fuel, and (this is cool) it even learns to change the entire fuel map for the car as the engine wears over the life of the car. But the sensor itself is a rather specialized device and deteriorates over time, which is why the #1 emissions-test failure is a faulty oxygen sensor. Now you know. The second common and important feedback sensor is a little microphone attached to the engine block that can sense “knocking” or detonation from the cylinders, allowing the computer to adjust the timing backward (later) to prevent damage to the pistons. A few knocks per 100 cycles will be OK. More is bad. Note that there can be problems associated with this sensor: the Mitsubishi 4G63 engines especially in Galant VR4s and Talons can be afflicted with a “phantom knock” even when the engine is running properly but goes into limp mode anyway, sapping power. Basically the microphone is picking up noise from somewhere else in the engine and sending it to the computer as a knock; many Mitsu people find that the valve lifters are the source of the noise. Similarly, in a rally car we have constant barrage of noise from rocks on the chassis and this can be picked up. As a result, I don’t have a knock sensor. those running O2 off and locking the afv. beware over time you will have issues. I have been in the field way to long to let this madness keep going. there is plenty good info for free online. and seminars are available from various places on proper tuning and why sensors are needed In FI. don't fall trap to a poor flashed ecm. there is a reason. | ||
| Xoptimizedrsx |
This is especially necessary in the Rotax based 1125. do it right and you gain a lot more. do it 1/2 way you may gain now but you will loose in the long run. if any doubt it not true contact the builder of the Buell rotax engine. they will tell you why It very important to use all functions on. mike | ||
| Xb1200rick |
how is it going with the new spreadsheets?  | ||
| Barker |
Thanx for the info. What are some of these issues that one could experience with the locked AFVs? | ||
| Ponti1 |
Wouldn't it have been easier to just post a link to the article?  http://www.innovatemotorsports.com/resources/ecu-b asics.php | ||
| Xoptimizedrsx |
well it would have been eisier to do a link but a lot of people wont go to links so I copied the txt >never said it was my info just facts.< spreed sheet is almost done. I took a break from inputting a bazillion txt values. Had to get a few friends to help with a few txt value conversions. some were tricky to find. my fingers are crossed to meet my personal deadline on the open source formulas/functions sheet. with this sheet it can easily be converted to any ecm. I may not do all the ecms but I'm sure there are others who will use this sheet setup to keep it going. fingers are crossed its free. I will load mine to microsoft to be found there as a shared sheet. mike | ||
| Ponti1 |
Sorry, Mike...Wasn't trying to say you were claiming it as your own. Guess it did sorta sound like that, huh? I had just read the article previously and saved the link, so I recognized it pretty quickly. | ||
| Xoptimizedrsx |
No problem. I can see how it could be assumed I was claiming it as mine. Anyone whom/who (which one arghh) has followed my post would have clearly noticed there was no way I wrote that. lol mike | ||
| D_adams |
So tell me something. How did all the internal combustion engines from the mid 1800's to maybe 1970 run without O2 sensors? Seems to me they ran fine without them. Maybe not quite as efficiently, but they ran fairly well none the less. You can tune a carbureted engine to run pretty damn good without any sensors at all. http://en.wikipedia.org/wiki/Four-stroke_engine
Disabling the O2 sensors only takes away the functionality of the closed loop mode. If you've tested and analyzed with a 4 gas EGA machine, you could probably do away with them, provided you got the fuel map set right. The ECM runs in open loop mode when cold or at WOT, so setting it to run that way all the time with the right map shouldn't hurt much. | ||
| D_adams |
Hmmm, this looks interesting. http://www.powercommander.com/powercommander/power commander_v.aspx?mk=21&mdl=318&yr=2009 | ||
| Slaughter |
Yep - and before carbs there were needle valves... simpler still. If you've ever raced a bike and carried a box of jets, you'd appreciate just ONE aspect of the FI system operating in closed-loop. If you have ever re-jetted for altitude and then had a hot day the next day... or gone from a cold morning to hot afternoon (we can see easily 50 degree temperature changes and as much as 1000 feet changes in density altitudes in just 24 hours) Yeah, you can re-jet or just accept a carb's slightly lower efficiency. FI is not the answer to every question - but it DOES work well. Kinda nice being able to switch from 91 octane (winter) and 100 octane (mid-summer) and just change a couple parameters in a 2 minute laptop hook-up on a hot engine. I've never liked changing jets. | ||
| Tbowdre |
How many tuners/available maps have locked AFV, or manual adjusted AFV? I only know of one.. in the midwest. How much do the AFV move anyway? I have 2100 miles with variable riding conditions and mine have been in the same spot for the past month... about 700 miles I am saving up for the O S B map, since this seems like the best option right now and want this to be no big deal! | ||
| Anonymous |
>>How many tuners/available maps have locked AFV, or manual adjusted AFV? As the final map delivered to a customer for street use? If so, none that should be allowed to play with anything more complicated than a paperclip. | ||
| Highscore |
Mike, your intuition is absolutely right: modern FI systems are a sophistic piece of technique and needs each of its sensor to work properly. So removing the O2-sensor reduces system "intelligence" and reduces its operational performance: The ECU uses the information gathered and "learned" by the lambda probes during "closed loop" , to calibrate the fuel injectors regarding actual flow. Here is an example of closed loop - operation:  The blue line in this picture shows the probe´s signal, the other the injector pulse for the corresponding cylinder at idle. The graphs belong to the closed loop of a Keihin ECU working in KTM Superduke, featuring like the 1125 two lambda probes for individual cylinder control. I do this test at each new model when applying a "Kastl" (my personal piggy back fuel nanny) to test the boundaries of closed loop range. I have performed this test for the 1125 too, but obviously not stored. But the Keihin ECU of the KTM shows a very similar pattern. The upper Limit of the probe current is 0,8V representing the "rich" side and limit of the lambda window, consequently 0,1V represent its "lean" limit. At the famous 14,7:1 or Lambda=1 the current is in the middle between these two extrema at aprox.0,45V. Two things are remarkable: (1) the complex shape of this control, the mixture changes the whole time up to 10% - this is not less. There is no equilibrium, no constant fuel timing, just continuous dynamic change. (2) The ECU holds the mixture preferably at the "rich" side of the window, This is typical feature of motorcycles. But this complex interaction of keeping the mixture inside of the lambda window is just the "short time adjustment" performed by the ECU. As a side effect of this operation the ECU senses additionally the efficiency of injector flow. For this purpose it compares the effective length of the injector pulse width, it needs to keep the mixture within the lambda window, with an internally stored reference value. The offset between gives the ECU the "idea", how far the actual injector flow is off. This procedure is called the "long time adjustment". The correction value, gathered by this operation, is valid for the whole engine range up to WOT, to compensate differences in fuel flow caused for example by a clogged fuel filter. I have tested this ability unintended on my car. My car is - of course - dyno-tuned. During two test series I changed the fuel filter. Happily the ambient conditions stayed the same between these two series. Also the A/F-ratio as the power output were rather identical. The only thing, which has changed due the filter swap, was the injector pulse width, which became suddenly significant shorter. The terms "short and long time"adjustment belong the the BOSCH-language. This company invented this procedure. And the control strategy of the 1125 ECU works obviously in the same manner. In closed loop the ECU controls the mixture beside of the lambda cycle in fact by sensing the air mass directly. Here the butterfly of the throttle body is just open for a small percentage and angle. This angle is known by the ECU by the TPS-sensor. So the butterfly acts like a bleed of known surface. This enables the ECU using the MAP sensor signal to calculate directly the amount of air passing the gap at the butterfly in the "speed density-style". Unfortunately the signal of the MAP sensor breaks down, when the butterfly is opened for a larger extent. This is caused simply by the fact, that the throttle diameter is so large. Therefore the area of the gap becomes really wide too, so wide, that beyond 20-25% fly-opening the pressure behind breaks down to air box level. This causes that the MAP sensor shows from this point upwards until WOT no change of its signal and is therefore unable to sense engine load anymore. Here in this area up to WOT the engine is controlled in the Alpha/N-manner, using only engine speed and TPS sensor as input for the ECU. Here at open loop the engine is controlled solely by a fix stored table which correlates certain engine and load points to a certain ECU-output, for example an injector pulse width. But there is no "thinking and calculating" in this action involved, this is just a plain "when"-"then" correlation. There are two of this correlation tables, one for injection and one for ignition. These two tables form the "core map" of an ECU. The correction of that correlation values, stored within the core map, is then the job a a different map, called the "trim map". This map merges the input of the ambient sensors with the "long time adjustment" value to a single common correction factor applied to the correlations inside of the "core map". "Core" and "trim-map" is again Bosch-terminology. So what happens now if we remove the o2-sensors (Sorry for the length of this description, but the matter is subtle)? One disadvantage of this procedure I have named above: no matching of the injector flow anymore. But in fact without lambda signal the trim-operation is switched of as a whole! This sounds curious because temperature and ambient sensors are still in place. But the "brain" of an ECU, its microprocessor, is anything but "flexible" in its job. If one of the trim variables as the 02-fuel-adaption is missing, before combining the remaining variables to something new, the common logic here is: If one variable is missing - fall back. The fall back map is the "core map" without any trim. This is the map "awaking" the engine at its very first run. So without o2-sensor the sophisticated operation of a modern FI system boils down to the functionality of an old fashioned carburetor. Highscore | ||
| 1_mike |
It ain't all that big a deal. Yea, by disabling the 02 sensors, you do cut out some functionality of the original system. But as others have said, carburetors fuel'd engines just fine, for many, many years, the original electronic fuel injector systems worked just fine without 02 sensors, for a few years...my 03 Yamaha R1 runs just fine without any 02 sensors..and funny thing...it didn't come with any.. And as for the milage add-up on a no sensor engine...my R1 has almost 980000 miles on it. Any change in tuenup because of its wear, will be negligible. As for the current crop of ECM's designed to work with them...there's MANY both in cars AND motorcycles running without them working..and running just fine. Is this basterdizing the original design...yes. But with what Buell (in this case) has given the general public...someone needs to do something to make these things run SOMEWHAT correctly, even if it never will be perfect. Beside...there is NO single perfect...for every situation...even with the 02 sensors working with the way this system is designed. Mike | ||
| Xoptimizedrsx |
one cool factor of FI. look at the old carb 2.0 liters running aprox 87 to 125 hp max. now look at the new 1.6 liters 160 to 250 hp. both have the same cranks and compression dimensions internally. its efficiency that changes. carbs need needles and a lot more work at altitude and pressure changes in the atmosphere. where FI has the sensors to make the adjustment for you on the fly based off the return feedback. IE self tuning within parameters. taking the sensors out of the fuel injection is like removing the vacuum lines off the carb. It will work still but not so well. thanks for a great post highscore/slauter. that was some key info. mike | ||
| Aeholton |
my R1 has almost 980000 miles on it. Is that like a record for an R1? Only 20K short of a million miles! I think you put an extra zero in there.  | ||
| Buelldyno_guy |
To me it's sort of simple. The engineers at BMC did some good work developing and then updating the different ECM's and FI systems over the year. I also believe the algorithms in those ECM do some serious math to keep the bike running as the sensors see different variations during a days ride. Out here we can see a 3000 foot elevation change as well as a 30 deg difference in temperature all within 150 miles. But then I also feel that when I change the engines volumetric efficiency number, by changing the intake track, combustion chamber size, displacement and or exhaust flow, then the fuel and timing tables need to be revised to supply the correct AFR. When you really get a chance to see what goes on with the cylinder splits on the Dyno during the same heat cycle/ temperature where the hot rear cylinder is controlling the cooler front cylinder, you will see how well it was designed. So I guess what I am saying is I like to leave almost everything inside the ECM, like temp offsets alone and letting a Buell be a Buell. As a rule our bikes run in "Closed Loop" learn at 100% AFV. So far that seems to work well for us here. terry@jtsperformance.com | ||
| Dave_bannister |
All very interesting and all! And doesn't seem that anyone has a better option just being negative with a closed loop.. But I love the difference with an O.S.B tune compared to factory tune,and it was so easy.. My bike is runs so smooth and cost is very low considering the instant results! I can flash it back to factory in seconds enabling those sensors... Disabling just O2 sensors and making it a close loop with adjusting AFV works well for me with stock exhaust. I love it and will be out riding with a big smile! | ||
| Strongbad |
If you are just cruising on the street then the O2 sensors do offer some benefit. However, none of the Buell factory race bikes used closed loop fueling. Purely relying on O2 sensor feedback to ensure proper fueling is far too slow for the transients that occur during racing. Each motor started with the same base map and then was modified depending on the different characteristics of each motor. The O2 sensors seen in photos were wideband sensors used purely for data acquisition. | ||
| Buelldyno_guy |
Disabling just O2 sensors and making it a close loop??? You just made it a open loop only system and turned off the sensor that controls some very important engine safety guards, best of luck. | ||
| Crackhead |
disabling the O2 sensor just turned the EFI system into a over weight and overly complex carb. | ||
| Easyrider |
In both situations, O2 turned on-off, It will depend on the quality of the current parameters inside the ECM how the bike will run. (this will depend on how much time somebody puts in the basemap) Making a good new basemap takes about 5 days. You can run in both situations a Beautiful bike, but in both situations you need a better basemap. | ||
| Highscore |
Sorry, my explanation was not really clear: The negative effect of disabling all corrections by an ECU regarding changes in ambient conditions occurs only, when the lambda probes are removed or replaced by "eliminators". To build your own eliminator, you just need the proper resistor. This is anything but a subtle high tech item. But those of yours, who are able to "talk" to the ECU and change its internal mapping directly, have the opportunity to disable the "closed loop" operation by the ECU in a way, which avoids this shortcoming: Keep the A/F target ratio at idle at stock and "lambda", just enrich the ranges above. In this condition the ECU still has the ability, to perform its "long time adjustment" and matching of the injectors by this procedure. BTW: Most of this matching happens during idle. For this reason KTM makes an "initial run" of 15 minutes on idle obligatory, if a new map has been loaded to the ECU or a reset has been done. Highscore P.S. Question to the ECU-hackers: You have the opportunity to change the adjustment of the rev-limiter easily (in my opinion the only real advantage of this technique). What is the maximum of engine speed, a stock engine withstands for a longer period without breaking? | ||
| Strongbad |
With the Buell DDFI system, you can remove the O2 sensors from the system without disabling any other features. The system will continue to use all other sensors to determine injector pulsewidth, it just doesn't have the feedback from the O2 sensors. I think it would be incredibly short sighted if any system discontinued looking at all of the sensors if only one becomes disabled. The stock motors should not be spun much faster than 11,000RPM. The rod and rod bolts are the weak point. | ||
| Fast1075 |
Call Carillo...send them a stock rod as a model and say "Titanium". | ||
| Highscore |
When you remove the O2-sensors AND replace them by eliminators, than the ECU runs into trouble (sorry my last post was imprecise here). Furthermore my thoughts concern only a stock, that means emission-controlled, ECU, which is sedated by a usual lambda eliminator to avoid CEL flashing Such an eliminator, featuring just some resistors, is a perfect tool to lock an ECU inside of a single operational state: the period after engine start. The resistor simulates the presence of a O2-probe, but sadly not is function - no resistor produces its own current as the probe does. So the ECU "senses" now: "yes, O2-probe installed, but sensor signal (=voltage) is missing". This is only "true" for the ECU, if the O2-probe is to cold for generating its current, what happens only shortly after engine start. A modern ECU tries in this situation its best to bring some heat not only into the O2-probe, but especially into the catalytic converter downstream the exhaust, which needs a similar temperature to "lift off" and start its activity. During this period the ambient conditions are not the ruling factor in control, the only "intention" of the ECU is to shorten this period as much as possible. Unfortunately with eliminators this period never ends. The ECU runs continuously a rather rich mixture. This would not be the problem, would not be there some bikes in trouble when riding in the mountains. Re-installation of the O2-probes cures this trouble instantly. Highscore (Message edited by Highscore on January 05, 2010) | ||
| Strongbad |
The ECM will not run in an over rich condition to attempt to heat the O2 sensor. After a set amount of time if the O2 sensor does not transition it will throw check engine code and will ignore the input. All other functions will work as usual. | ||
| Highscore |
Actual emission tests start with an engine of standard ambient temperature, the known standard temp for power-testing. When cold, any engine needs a rather rich mixture to run smooth. The test last something around 15 minutes. The first 2 minutes produce 75% of harmful substances, the engine expels in total. Unless the core of a catalytic converter reaches temperature at its cells of more than 500`F (are this metric 270?) the exhaust gases are "unfiltered". In this situation a modern ECU operates the engine in a manner, that provides a sufficient concentration of (residual) oxygen in the exhaust, to "ignite" the cat as soon as possible Any improvement here means a lot for the result of an emission test. Therefore we may aspect the 1125 shows similar features. But again, this thoughts concern only the emission-controlled ECU, an open "race ECU" does not need such mechanism. An ECU throws a "lambda-conversion-error", when the current by the O2-probe does not change, although the ECU has altered the A/F-ratio. Replacing a probe now by a plain resistor gives the ECU no probe current, at no time. This avoids the CEL, but there is price to pay for this simple fake. Highscore | ||
| Strongbad |
Highscore, you are mostly right. However, you will still get a CEL with a resistor. The controller will see that after so much time, there is no transition from rich to lean or vise versa and it will throw a failure to transition code and will ignore the O2 input. At that point the closed loop correction will default to no correction but the learned fuel values will remain unchanged. All startup enrichment features are on timers and are temperature based. Regardless of whether the O2 sensor is functional or not, the startup functions run as normal. |
