Tuning Calc. Load max on NA engine
#156
Characteristics of LOAD_PCT are:
Reaches 1.0 at WOT at any altitude, temperature or rpm for both naturally aspirated and boosted engines.
Reaches 1.0 at WOT at any altitude, temperature or rpm for both naturally aspirated and boosted engines.
So, that is the basic problem here. What is called calculated load, the SAE calls absolute load, which is really just volumetric efficiency.
Here is a formula that comes pretty close to calculating the correct values for us.
Code:
VE = (MAF*(IAT+273.15)/(80*RPM*100)*212544)/2
edit:
This patent describes how load is calculated.
Importantly, the purpose of the "calc load max" table:
With the engine control system in accordance with one preferred embodiment of the present invention in the high pulsation range such as the range of high engine loads the estimate of a required amount of intake air which is made based on a factual amount of intake air is restrained or otherwise prohibited so as to eliminate a large error in measurement of the required amount of intake air due to pulsation of intake air Further when making a utilization of an estimate coefficient which is obtained based on a change in the factual amount of intake air the estimate coefficient is corrected less in the high pulsation range than in a range where pulsation of intake air is weak providing an accurate estimate of the required amount of intake air because of lenient reflection of intake air pulsation on the required amount of intake air
Last edited by oltmann; 12-31-2011 at 02:10 PM.
#158
Boosted Kiwi
iTrader: (2)
edit:
This patent describes how load is calculated.
Importantly, the purpose of the "calc load max" table:
wow - was that translated from Klingon or something ?
anyway the crux of it seems to be explained best by this :
so as to eliminate a large error in measurement of the required amount of intake air
#162
It pulls a value from each table, load, baro, IAT, and multiplies them together. Call the product X.
X is then used in two places. In one place it is used as a hard limit, if the load value, as calculated up to that point, is higher than X, then load is set to X.
In another place, the present load value is divided by X. This is what gives you sort of a soft load limit at times. This part can get branched over under some circumstances.
There is more that goes in to how the load value is calculated, but that is the role of the max tables, and the punchline is that if you set them all to 2, they won't actually act as a hard limit until load goes over 8. Good luck with that.
X is then used in two places. In one place it is used as a hard limit, if the load value, as calculated up to that point, is higher than X, then load is set to X.
In another place, the present load value is divided by X. This is what gives you sort of a soft load limit at times. This part can get branched over under some circumstances.
There is more that goes in to how the load value is calculated, but that is the role of the max tables, and the punchline is that if you set them all to 2, they won't actually act as a hard limit until load goes over 8. Good luck with that.
#163
that is exactly what I saw on a tune every load table was set to 2, and the fuel tables went to 3.50
and the logs happily went that high....
and the logs happily went that high....
Last edited by FazdaRX_8; 02-10-2012 at 01:50 AM.
#164
The max load table is not an absolute max. The product of all three max load tables is an absolute max. Just A x B x C. So, to get over 2.0 you would need to increase the IAT and baro max tables as well. "Calc. Loax Max" should be named "Calc. Load Max RPM Compensation."
Under normal conditions, I never noticed the effect of the IAT and baro tables, because they always give a value very near to 1.0.
But, let's say you are racing across Death Valley in a freak blizzard. Barometric pressure is 110kPa, and air temp is -40. In principle, you could see 1.41 load even on the stock tune.
Ideally, you wouldn't even need to expand the fuel tables, as they only determine the lambda target. If everything was set up correctly and you had 0.8 in the 1.25 load column, you would still get 0.8 even at 2.5 load.
Of course, "everything" includes a bunch of stuff that nobody understands, so it is expedient to overcompensate with the fuel tables.
Under normal conditions, I never noticed the effect of the IAT and baro tables, because they always give a value very near to 1.0.
But, let's say you are racing across Death Valley in a freak blizzard. Barometric pressure is 110kPa, and air temp is -40. In principle, you could see 1.41 load even on the stock tune.
Ideally, you wouldn't even need to expand the fuel tables, as they only determine the lambda target. If everything was set up correctly and you had 0.8 in the 1.25 load column, you would still get 0.8 even at 2.5 load.
Of course, "everything" includes a bunch of stuff that nobody understands, so it is expedient to overcompensate with the fuel tables.
#166
Registered
iTrader: (25)
you will theoretically get 0.8, but in reality you will likely see something different at a variety of load factors in between which is why you do it that way
I give credit where credit is due though and that is the best explanation of it that I have ever seen on this forum, so congrats.
and btw my previous patent comment comes from having done a stint as a patent examiner, my first job out of engineering school. You often cannot always rely on verbage from them as being correct, let alone as being absolutes. They want to provide enough for legal protection without revealing any of their secrets. Otherwise all you need to get on the same playing field as your competition is to secure all their patents and then expand upon them to a next newer level without having done the hard work to get there, which is obviously not in the patent holders best interest. That's not to say people don't learn anything from patents, but rather to say that words in a patent align with a theory, which then makes it true, is rubbish. Until you have put it in practice and get out the expected results then you only have an unproven guess.
I give credit where credit is due though and that is the best explanation of it that I have ever seen on this forum, so congrats.
and btw my previous patent comment comes from having done a stint as a patent examiner, my first job out of engineering school. You often cannot always rely on verbage from them as being correct, let alone as being absolutes. They want to provide enough for legal protection without revealing any of their secrets. Otherwise all you need to get on the same playing field as your competition is to secure all their patents and then expand upon them to a next newer level without having done the hard work to get there, which is obviously not in the patent holders best interest. That's not to say people don't learn anything from patents, but rather to say that words in a patent align with a theory, which then makes it true, is rubbish. Until you have put it in practice and get out the expected results then you only have an unproven guess.
#168
The load value is meant to represent the ratio between the current air mass entering the chamber and the theoretical maximum air mass at wide-open throttle, standard atmospheric pressure and temperature, 100% volumetric efficiency.
The goal of these corrections is to figure out how much air is actually going in to the chamber, which is not usually the same as what the MAF is currently sensing.
If this works perfectly, then it will also be able to calculate exactly how much fuel is needed to achieve the equivalence ratio determined from the fuel tables. You could reduce all the fuel tables down to a single table with a single cell set to 0.9 lambda, and you would get that under all conditions.
Ideally, the fuel tables would be set to your actual desired equivalence ratio. However this would require that the load value be correct at all times, and likewise for the ecus model of the fuel injectors behavior.
I get what you're saying about patents, Team.
The goal of these corrections is to figure out how much air is actually going in to the chamber, which is not usually the same as what the MAF is currently sensing.
If this works perfectly, then it will also be able to calculate exactly how much fuel is needed to achieve the equivalence ratio determined from the fuel tables. You could reduce all the fuel tables down to a single table with a single cell set to 0.9 lambda, and you would get that under all conditions.
Ideally, the fuel tables would be set to your actual desired equivalence ratio. However this would require that the load value be correct at all times, and likewise for the ecus model of the fuel injectors behavior.
I get what you're saying about patents, Team.
#169
Still working on digesting this thread, but missing something. When you are adjusting max calc load for FI applications are you also adjusting baro and iat or leaving those alone? I haven't quite grasped the inter-relationship.
#172
I agree with Kane. I've never tuned FI, but you should leave those tables alone if you can. If you need load to exceed 1.99, you can scale up the iat and baro tables, but it will take some algebra. Even in that case it would be better to change the scaling value for the load max table, but that isn't possible currently.
Load is meant to be the ratio between trapped air mass and the mass of 1.3 liters of air at standard pressure and temperature. 101.72kPa and 25C in this case.
I have found it useful to estimate this using a WOT log data of MAF g/s and this formula:
Basically, MAF g/s over theoretical mass of air for 1.3l at current pressure and temperature. By using the current ambient air density, I exclude their influence and try to get at the rpm-dependent part of volumetric efficiency.
If you can manage to get this dead on, it does a good job of compensating for errors that the MAF, even when perfectly calibrated, cannot, due to weird things that happen with a variable intake (backflow, resonance, etc.)
Load is meant to be the ratio between trapped air mass and the mass of 1.3 liters of air at standard pressure and temperature. 101.72kPa and 25C in this case.
I have found it useful to estimate this using a WOT log data of MAF g/s and this formula:
Code:
MAF/(BARO kPa/(287.058*(IAT C+273.15))*(1308*RPM/60))
If you can manage to get this dead on, it does a good job of compensating for errors that the MAF, even when perfectly calibrated, cannot, due to weird things that happen with a variable intake (backflow, resonance, etc.)
#174
no agenda
iTrader: (2)
FWIW I flattened these tables during my initial tuning to eliminate the variables induced by these tables.
I put them back to the OEM values once I was happy with the Maf scaling and the fuel tables/VE table.
Not saying this is good/bad/best ... its just the way I did it.
I put them back to the OEM values once I was happy with the Maf scaling and the fuel tables/VE table.
Not saying this is good/bad/best ... its just the way I did it.