Discussion
If you've only got a narrowband lambda then any values it shows up on screen are useless, particularly if you've had the ECU mapped whilst your car is on a RR.
Narrowband lambdas only read stoic values plus or minus a very very small percentage either side. After that they show up as either LEAN or RICH. Given that a well tuned sports car will have a lower AFR than stoic, the Lambda will always show RICH while your foot's on the gas, and LEAN on the overrun, and that's it I'm afraid.
As an aid to tuning, a narrowband Lambda is a proverbial chocolate teapot.
Narrowband lambdas only read stoic values plus or minus a very very small percentage either side. After that they show up as either LEAN or RICH. Given that a well tuned sports car will have a lower AFR than stoic, the Lambda will always show RICH while your foot's on the gas, and LEAN on the overrun, and that's it I'm afraid.
As an aid to tuning, a narrowband Lambda is a proverbial chocolate teapot.
My impression was that the MBE remap retains the use of the lambda sensors, although these are their for emissions purposes and trim the fuelling only under certain conditions. I don't believe the earlier Emerald remaps ran it, however Emerald has supported it for some time now.
Steve.
Steve.
What mars means (quite correctly) is that unless you have a mixture very close to lambda=1 a narrownband sensor is all-but useless for mixture reading. you can get more accurate narowband sensors which monitor the tip temperature and compensates the voltage output, these are reasonably accurate down to lambda0.85 and so are good for mapping normally aspirated cars, not good enough for some mega-rich turbo applications though. My Autronics AFR reader is of this type.
You can still keep closed loop on the remapped cars, though i sometimes disable much of the closed loop on problem cars because running the cerbera slightly richer then lambda=1 on part throttle smoothes it out considerably .. and then i just run closed loop where you have to. You need access to the mbe lambda maps though and i'm not sure anyone else has managed this yet. On the emerald, stuff like this is easy.
You can still keep closed loop on the remapped cars, though i sometimes disable much of the closed loop on problem cars because running the cerbera slightly richer then lambda=1 on part throttle smoothes it out considerably .. and then i just run closed loop where you have to. You need access to the mbe lambda maps though and i'm not sure anyone else has managed this yet. On the emerald, stuff like this is easy.
Sorry, I was being a bit less than specific. I was reading a large book on tuning which was talking about acceleration lamda values and the need to keep them within certain limits. I was most interested in the 0.8 level values.
The lamda, even in the Cerb continues to operate even on full acceleration but is ignored above 4000, and may even be ignored on acceleration.
My questions should have been does the lamda value show anything on full acceleration or is it on the stop so to speak. Although the Cerb is a narrowband sensor I didn't know how narrow 'narrowband is'.
I was stunned to read that lamda 1 isn't the point of greatest fuel efficiency but a compromise between fuel efficiency and emissions.
The lamda, even in the Cerb continues to operate even on full acceleration but is ignored above 4000, and may even be ignored on acceleration.
My questions should have been does the lamda value show anything on full acceleration or is it on the stop so to speak. Although the Cerb is a narrowband sensor I didn't know how narrow 'narrowband is'.
I was stunned to read that lamda 1 isn't the point of greatest fuel efficiency but a compromise between fuel efficiency and emissions.
You can pick up a lot of part throttle hp by running richer than lambda=1, this means that you then have to take your foot of the pedal compared to previously, so the net result is that by running richer ratios than lambda=1 you actually return better fuel economy. there are many very good reasons to take the cerb (and other cars) out of closed loop control, even catalysts last longer when not subjected to the weak mixture misfires that occur as the engine dances to the weak side of lambda=1 before the trimming catches up again ...
Julian64 said:
Sorry, I was being a bit less than specific. I was reading a large book on tuning which was talking about acceleration lamda values and the need to keep them within certain limits. I was most interested in the 0.8 level values.
The lamda, even in the Cerb continues to operate even on full acceleration but is ignored above 4000, and may even be ignored on acceleration.
My questions should have been does the lamda value show anything on full acceleration or is it on the stop so to speak. Although the Cerb is a narrowband sensor I didn't know how narrow 'narrowband is'.
I was stunned to read that lamda 1 isn't the point of greatest fuel efficiency but a compromise between fuel efficiency and emissions.
The lamda, even in the Cerb continues to operate even on full acceleration but is ignored above 4000, and may even be ignored on acceleration.
My questions should have been does the lamda value show anything on full acceleration or is it on the stop so to speak. Although the Cerb is a narrowband sensor I didn't know how narrow 'narrowband is'.
I was stunned to read that lamda 1 isn't the point of greatest fuel efficiency but a compromise between fuel efficiency and emissions.
On the cerb the lambda sensors are used to trim over the base fuel map at throttle site 10 and under and rev sites 4000 and under (at 4000 it's back to the base map and the 3750 number is an interpolation from the 3500 number).
The trim values are stored in the adaptive maps, so when a fuel site is visited at any trimmed stage the fuel number injected is the base map, with any lambda trim, with any coolant , air temp and air pressure trims also applied.
Under acceleration there is no accel enrichment applied unless the throtle passes through 4% or more of throttle change in a given time interval ( sadly i've forgotten what that time interval is at the mo!). When accel enrichment is being applied you will see the rate of change of throttle box (dThrottle) read "no control" or "disabled" (been so long since i used the software i can't remember what it says exactly). This is also useful for seeing when the throttle pot spiking is disrupting the idle and part throttle fuelling .. if the signal spikes enough to register on the dthrottle box then accel enrichment is being applied, this richens the mixture .. some over-fulling then stays in the ports after the signal stops spiking, this is read as a rich mixture then by the lambdas and the adaptive trim is weakened off .. this is a primary cause of poor running at low speeds ..
When under acceleration the engine revs and throtle change too quickly for accurate closed loop fuelling to occur, therefore lambda control is suspended (box reads "disabled". In this case the fuelling delivered is the base map, plus adaptive map table lambda trim, plus corrections, plus accelration enrichment. real-time lambda control and lambda trimming is suspended until revs and throtle become stable again.
The article I was reading was slightly seperate from enrichment. It was an analysis of fuel combustion. 14.7:1 being the conventional lambda value.
However peak fuel power came from a ratio of 12:1, which corresponds to a lambda of 0.8, and peak fuel efficiency/consumption came in at 16:1 corresponding to a lambda of 1.1.
I was wondering if the lambda (which agreed is ignored but still working to be seen on the graph) could be used as a tuning aid not just for closed loop emission control below 4000 but along the entire rev range for matching to 0.8 in the rev range when peak power is required.
Obviously according to mars this is not possible as the 0.8 is outside the narrowband spectrum of the Cerb, so a bit of a non starter. Hence the how narrow is narrowband. But I would be intereted in a post from a wideband user an acceleration graph.
However peak fuel power came from a ratio of 12:1, which corresponds to a lambda of 0.8, and peak fuel efficiency/consumption came in at 16:1 corresponding to a lambda of 1.1.
I was wondering if the lambda (which agreed is ignored but still working to be seen on the graph) could be used as a tuning aid not just for closed loop emission control below 4000 but along the entire rev range for matching to 0.8 in the rev range when peak power is required.
Obviously according to mars this is not possible as the 0.8 is outside the narrowband spectrum of the Cerb, so a bit of a non starter. Hence the how narrow is narrowband. But I would be intereted in a post from a wideband user an acceleration graph.
lambda=0.8 is very rich, and you certainly won't see max hp from a cerbera engine at 0.8 and someone told me the other week that RV8s generally like 0.85-0.88 for max hp so not sure what engines the article was refering to. It's not as easy as just saying "an engine needs lambda 0.8 for max hp" .. that's where the hours on the dyno come in, fine tweaking the mixture and timing to get the hp, in fact the fuelling curve for the cerbera changes through it's rev range as regards what the engine actually needs, as i'm sure most engines will.
If you ran lambda=1.1 on a cerbera youd find it very rough, it's hard enough trying to ignite a lambda=1 mixture in the cerbera combustion chamber never mind anything 10% or greater weaker, as you can feel from the stutery nature of the part throttle power delivery on most cerbies. Richen it up slightly and all smoothes out. In the emerald you can target the lambda you require across the whole rev/throttle range and ask the ecu to self-map to those values. In the std MBE you can do that too, but not on the TVR version although the facility to take it out of lambda control exists, but it's eiher target lambda=1 or no lambda at all (at the moment)
As regards the accel enrichment, I'm not sure you'd get a lambda to respond fast enough to register what was actually going on in real time. You can put a lambda in the exhaust right at the head face, but by the time it's reacted the engine is already at the next part of the rev range .. i guess you could have some kind of time-aware data logging with time-shift effect to put a registered lambda voltage into the correct part of the throtle/rev time frame but that's beyond anything most people get involved with. Maybe someone like 350matt who used to work for cosworth could give you some info on registering and logging lambda value under acceleration.
most people (like me) give the engine what it needs to accelerate smoothly (in terms of enrichment, how long that lasts for and how quickly that blends back into the base map) and leave it at that. For me (and the customer) the actual numbers are irrelevant and all that matters is that the car accelerates smoothly and quickly.
If you ran lambda=1.1 on a cerbera youd find it very rough, it's hard enough trying to ignite a lambda=1 mixture in the cerbera combustion chamber never mind anything 10% or greater weaker, as you can feel from the stutery nature of the part throttle power delivery on most cerbies. Richen it up slightly and all smoothes out. In the emerald you can target the lambda you require across the whole rev/throttle range and ask the ecu to self-map to those values. In the std MBE you can do that too, but not on the TVR version although the facility to take it out of lambda control exists, but it's eiher target lambda=1 or no lambda at all (at the moment)
As regards the accel enrichment, I'm not sure you'd get a lambda to respond fast enough to register what was actually going on in real time. You can put a lambda in the exhaust right at the head face, but by the time it's reacted the engine is already at the next part of the rev range .. i guess you could have some kind of time-aware data logging with time-shift effect to put a registered lambda voltage into the correct part of the throtle/rev time frame but that's beyond anything most people get involved with. Maybe someone like 350matt who used to work for cosworth could give you some info on registering and logging lambda value under acceleration.
most people (like me) give the engine what it needs to accelerate smoothly (in terms of enrichment, how long that lasts for and how quickly that blends back into the base map) and leave it at that. For me (and the customer) the actual numbers are irrelevant and all that matters is that the car accelerates smoothly and quickly.
Interesting thread this
.If you used an innovate wideband lambda LC1,you could datalog the lambda values on different runs.If you then used a gps performance meter,you could then datalog the power runs.Then you could cross correlate and find out which mixture is best for what your looking for.Dispite the fact that all the technology is easily available these days,it would be far easier,and cheaper,to visit your freindly rolling road and have some fun on the dyno finding out
.If you used an innovate wideband lambda LC1,you could datalog the lambda values on different runs.If you then used a gps performance meter,you could then datalog the power runs.Then you could cross correlate and find out which mixture is best for what your looking for.Dispite the fact that all the technology is easily available these days,it would be far easier,and cheaper,to visit your freindly rolling road and have some fun on the dyno finding out Edited by The AJP Griff on Tuesday 27th February 10:13
Sound like a dog with a bone here, and realise that reality and theory are not good bedfellows.
But as far as I understand it major car manufactors design the car around the lamda 1. Peak acceleration ignoring acceleration enrichment which is only a few seconds dependant on sudden thottle changes.
If you are a major car manufactor you have to stick to the lambda 1 throughtout the rev range.
If we don't have to stick to 1 cos we don't care much about emissions and we have a car accelerating hard and the trick is to achieve maximum energy out of the fuel in a completely testube way which relies on a a lambda of 0.8. The reason the Cerb has trouble achieving that is probably for a mass variety of reasons from airflow to flame propogation, to temperature of combustion and back.
Is the trick not to find out how to make the Cerb burn a lambda of 0.8 rather than accept the limitations that may be being placed on the burn from other bits we could change like the temp of the engine or the inlets?
But as far as I understand it major car manufactors design the car around the lamda 1. Peak acceleration ignoring acceleration enrichment which is only a few seconds dependant on sudden thottle changes.
If you are a major car manufactor you have to stick to the lambda 1 throughtout the rev range.
If we don't have to stick to 1 cos we don't care much about emissions and we have a car accelerating hard and the trick is to achieve maximum energy out of the fuel in a completely testube way which relies on a a lambda of 0.8. The reason the Cerb has trouble achieving that is probably for a mass variety of reasons from airflow to flame propogation, to temperature of combustion and back.
Is the trick not to find out how to make the Cerb burn a lambda of 0.8 rather than accept the limitations that may be being placed on the burn from other bits we could change like the temp of the engine or the inlets?
trackcar said:
i guess you could have some kind of time-aware data logging with time-shift effect to put a registered lambda voltage into the correct part of the throtle/rev time frame but that's beyond anything most people get involved with.
It sounded hideously complicated, but as far as I could understand it that's something that the MegaSquirt people have been working on. If I understood it correctly they were modelling the propagation delay between the injection events and the lambda events. The only reason I can see for this is to provide very fast real-time feedback under changing conditions, which I think is what you're describing.
It seems far simpler to me simply to remember the error and get it a bit better next time you encounter similar conditions.
GreenV8S said:
trackcar said:
i guess you could have some kind of time-aware data logging with time-shift effect to put a registered lambda voltage into the correct part of the throtle/rev time frame but that's beyond anything most people get involved with.
It sounded hideously complicated, but as far as I could understand it that's something that the MegaSquirt people have been working on. If I understood it correctly they were modelling the propagation delay between the injection events and the lambda events. The only reason I can see for this is to provide very fast real-time feedback under changing conditions, which I think is what you're describing.
It seems far simpler to me simply to remember the error and get it a bit better next time you encounter similar conditions.
Sounds just like what we're talking about to me .. there's a lot to influence the response time, from physical restrictions like how close you get the sensor to the chamber to the actual change of response time of the sensor with temperature. not sure how they're going to corelate all that but it'll be fun watching it evolve.
trackcar said:
Are they saying the flame front travels fastest on lambda=0.8 then? in which case you'd want a chamber as close to spherical as possible to make best use of it? A bit tricky from the angled bath tub chamber in the cerbera.
Just to add a little more confusion according to the book I'm looking at
(Bosch Automotive Handbook 4th edition) it say that maximum flame propagation /combustion speeds are "obtained with approx. 10% excess fuel (lambda=0.9)" Julian64 - I really don't think OEMs tuning only to lambda=1 throughout the whole rev range and they do employ accel enrichment. They take much more into account than throttle position and load. Modern engines employ 'torque management' which adjusts fuel and timing based on every factor you can think of in every condition and a few you wouldn't think of as well. but yes I agree they do need to be more concerned about emissions than some, but that does mean its as simple as lambda=1 the whole time..
Edited by griffsmith on Tuesday 27th February 12:50
You may well be right, but I'm still trying to grab the simple stuff, and regurgitating the stuff I have read to have trackcar and the like tell me where the theory I'm reading doesn't mate with reality, helps. A lot of people talk about common engine stuff lambdas injectors etc but I think there are actual farily few people who actual understand them.
For instance could anyone here actually tell me exactly what the calculation is by which an adaptive is calculated. In other words if I have a calculator now, and freeze the engine in mid stroke, can anyone tell me what the adaptive about to appear will be.
In truth theory is all I have as practicality is difficult stuff to actually get your hands on. Specially when you car has not put a foot wrong in the last four years engine wise so no real excuses to go playing. From my theory I have a thousand ways to improve an engine, but I bet not one of them is practical.
The lambda is a completely non sensical ratio, because it seems to be an average of the stiochiometric calculations for three different gas chemical reactions. Only one of which actually produces a significant amount of volume and therefore power. The race to get complete combustion to C02 and H20 is probably an erroneous one unless you are interested in the environment. The trick seems to be just concentrate on the one reaction, and forget about the catalyst.
I really want a set of dirt cheap hub dynos and stop reading the books is the point of this thread in case anyone started wondering
For instance could anyone here actually tell me exactly what the calculation is by which an adaptive is calculated. In other words if I have a calculator now, and freeze the engine in mid stroke, can anyone tell me what the adaptive about to appear will be.
In truth theory is all I have as practicality is difficult stuff to actually get your hands on. Specially when you car has not put a foot wrong in the last four years engine wise so no real excuses to go playing. From my theory I have a thousand ways to improve an engine, but I bet not one of them is practical.
The lambda is a completely non sensical ratio, because it seems to be an average of the stiochiometric calculations for three different gas chemical reactions. Only one of which actually produces a significant amount of volume and therefore power. The race to get complete combustion to C02 and H20 is probably an erroneous one unless you are interested in the environment. The trick seems to be just concentrate on the one reaction, and forget about the catalyst.
I really want a set of dirt cheap hub dynos and stop reading the books is the point of this thread in case anyone started wondering
Gassing Station | Cerbera | Top of Page | What's New | My Stuff