If it was possible to provide fast accurate wide band O2 readings to an engine computer is there any reason why it couldn't tune itself? And continue to do so as conditions changed. In fact with a fast enough computer (which there are plenty of these days) why would there even need to be a fuel map? It could just be calculated on the fly. Perhaps?
I was having this idea some 6 months ago, on the fly tuning would be totally awesome, especially since conditions change, it's not like you can have separate fuel maps for different atmospheric pressure and temp.
You'll find that this is one of the most important factors in tuning for hillclimb events- the Pikes Peak Hillclimb in Colorado USA is the best example: 20km long, climbing 1,439 meters with an average grade of 7% on dirt, snow, and paved sections. So yeah, atmospheric pressure and temp are important considerations. But using different tuned maps for different barometric pressures would be a great start. I was talking to a mechanic recently about this- his hillclimb VW beetle was running poorly at bathurst (cold and high elevation) compared to its previous race at grafton just a week before.
Someone posted a link to SAU not too long ago where someone got their 34GTR setup for any fuel. Just pour in whatever and the car adjusts everything to suit. This is basicly auto tuning. Can't find the link now tho.
They may well do but that is not what I was really talking about. Those systems get put into 'learn' mode and then create an A/F map that can be tweaked if desired. I was wondering if it would be possible where the 'map' are generated instantly depending on what ever the current inputs where. Cheers, Evan.
It's definitely possible but would require the addition of a few more sensors, a microprocessor with many ADCs and quite a lot of boring sequential logic design.
I'm not serously thinking of doing it, just wondering if there is good reason why no one else ever has.
The scope for an aftermarket ECU to do this is un-feasible because ideal AFR's for one motor will produce less than ideal results in others. Timing advance has different results on power and reliability with different AFR's and in different engines as well, so is your ecu also going to dictate timing advance on the fly as well? This will need EGT sensors for every cylinder. And now we have both AFR's and Timing advance to be considered, how is the ecu going to measure the performance of it's tune? Some engines can handle higher EGT's than others, and some engines can handle much leaner low load running than others. Most modern factory ECU's adjust AFR's against a base map. Don't forget that the base factory fuel map is essentially the map designed by the factory as the most ideal map for a compromise between power, reliability and emissions. From there the ECU adjusts fuel delivery based on the O2 sensors. Some new cars use widebands and hence adjust the fuel delivery across the rev range rather than just in the open loop section of base map.
Adaptronic does everything you are asking for. You never going to get anything to run from scratch with nothing loaded. You need to give the engine "something" to start with and that something must fall between the upper and lower EGO ranges of whatever EGO sensor type you are using. This will be 9:1-19:1 for a wbo2 but realistically, anything above about 16:1 will barely keep running anyway. So you still need to enter into the VE tables a realistic figure (in VE tuning mode) and the ECU will do the rest adaptively along with adaptive ignition control if desired and on the fly. Again, ignitionwise, you will still need to enter some realistic timing figures to begin with and the adaptive ignition control will tune around that. Adaptive ignition control is a tricky affair to get right tho. However, in order for the ECU to make an informed decision whether to apply a correction or ignore it as a one off occurrence, requires a preset amount of time. Anbody who has watched a wbo2 output will realise that AFR's are never actually a nice constant figure and do fluctuate markedly depending on a variety of factors. So in order to see realistic figures and to prevent the ECU from making wild correections based on something funky like, say a sudden throttle rolloffoff which will invariably result in a very rich spike, the ECU will need to see a particular set of conditions for a specific amount of time and the time varies depending on the resolution, or the "fineness" of the correction. Many oem ecu's will eventually apply these correction permanently but its a fairly slow process as the engine parameters need to be held at a given level, say, load v's rpm for a significant length of time to actually make a correction that sticks. Also, its fairly difficlut to achieve enough load-rpm sites for long enough to actually make decent tuning decisions whilst out on the open road. Also, Adaptronic, runs in adaptive mode FULL TIME referencing to either WB or NB and unlike most others, does not simply apply corrections over a tuning table table but rewrites corrections to ignition and VE tables on the fly and permanently to the fuel and ignition tables ALl the time. I can make some fairly major changes to the VE tables then go for a drive and log the run. I display it in an analogue viewer ( I use Megalog Viewer) and you can see exactly when the engine moves into those dodgey areas and the ecu makes fairly savage changes in closed loop mode to compensate. However, you still need a "base" map which the ecu can default back to in the event of an EGO failure or other issue or perhaps for cold running before the EGO sensor reaches ideal operating temps, or say the EGO heater fails. And that base map needs to be a pretty good one, as close as possible to fairly "right" to prevent the engine from grenading itself if the map was too lean. Adaptronic does it all pretty well at pretty much a fraction of the cost of the big name stuff. No i dont work for or sell them, I just use one in the "Ute Of Death" is all and seriously happy with it! have fun. E
Its just not practical - thats why Yeh agree with you -and as you more or less point too in the end, it would be best to seperate each intake and outake if your that inclined about optimal tuning. I dont think that the difference of a recent wideband preset map and an on the fly tune would be that great, unless as mentioned, your doing a professional hill climbing competition. But for your typical street car, it would be an overkill in the sense that it would invariably be more costly and more prone to failing, from the considerable extra number of sensors required in the longer term. Jamie - (Misszens other half)
WRONG. Try reading my post above again. Its not costly at all, it does not fail and does not require additional sensors. I suspect you have NO experience with such stuff and are merely speculating. In which case, how can you "agree' with anything to do with autotuning and full time adaptive tuning. l8r E
I don't think the elevation would make that much of a difference, even for a hill climb. If we use mount panorama circuit as an example, going by the weather there right now it's 1014 hPa, or (to put it in car powahhh talk for everyone, 14.71psi)... compared to 1021 hPa (14.81psi) right here in Perth at the moment, or the current 1012 hPa (14.70psi) in Sydney.
Easy fix, have three no four ecu's, one under the seat another in the glove box one in the spare tire spot ect, have them all set up for different altitudes temptures so on. told ya easy.
Elevation does make a reasonable difference. When you quote the hPa at certain places, you use that pressure as a variance of sea level norm to get the altitude. So the actual pressure you mentioned at Mt Panorama is 1014hPa relative to sea level 1013. So the pressure height is only 30 feet less than the actual height on this day of 2830, making it 2800ft pressure height or so. So yeah.. the your actual pressure at the mountain top is not 1014... it indicating your pressure height is 30ft (30ft per 1hPa) more pressure than normal. So back to the numbers you mentioned... you dont have more pressure on the mountain top.. ofcourse you have a little less.. around 13.2psi relative to a sea level norm of around 14.7. (to put it in car powahhh talk for everyone, about 13.2psi) If you drive in mountains around 10,000ft (not in Aust ofcourse) or so (I mean even the pressure there is quoted at 1013hPa on average)... it means your pressure height of around 10,000ft is about 10.1psi. So about 1/3 or 1/4 less air available. Happy days.
Ah I see! Now, while I certainly understand pressure makes a difference (the top gear episode mentioned is a great example), at the heights mentioned here (mainly hill climbs in Australia) is it really going to be that significant? Or anything to mention, when compared to factors such as the temperature of the track? Cheers for clearing up my confusion about weather readings too. That "relative to sea level" pressure standardising is a bit annoying though. Shows how much I know about meteorology. :rofl:
Well im just guessing about the next bit... but a sea level of 1bar (14.7psi) compared to the Mt Panorama of .89bar (13.2psi)... About 10% difference in air pressure, i would guess a close enough to 10% difference in power available too... Any NA motors really struggle with real altitudes. Once you get above 10,000ft, they get really guttless. Power available is bugger all. Superchargers were invented for aircraft engines to fly higher. But yeah the heights in australia.. about 3000ft is about the highest track then.. is not signifigant compared to the rest of the world. I mean look at pikes peak!
