BRZedit Fuel Trims, Closed to Open loop transiton

[mad_sb]

Originally Posted by Visconti

I certainly don't have any problems with that.

I'm able to set the fuel targets correctly all while having a correct looking MAF curve with very small trims if any across the board.

Looks like you just need to spend a little more time on it or brzedit doesn't offer the control that ecutek does.

John

Well, without knowing how you have your transition setup (stock timers, modified timers, zeroed timers) I don't under stand what areas of the map your talking about tuning. If you zero out the timers you can't set the closed loop areas any lower than the min closed loop afr or it will drop into open in that cell.

I completely agree that getting the maf scaling dialed in is the key to getting rid of big fuel trims.

It's also possible my stock maf housing had some issues.. once i went to the fa20club cai i was able to go back to the stock maf table (just modified the 3.2v and up area a tiny bit to line the actual afrs up with the targets) with much better fuel trims than i had stock.

[mad_sb]

Here is one more table that seems to have an affect on trims, all be it indirectly.

I'm not 100% sure what this table is, but I believe it is a load scalar table. So, whatever the ecu calculates the load to be based on maf is multiplied by the values in this table to determine the final load.. but that's just a guess.

It does not seem to be speed density fail-safe table just based on the fact that the higher vacuum (lower manifold pressure) area have larger values than the higher load (higher manifold pressure). If it were a fail safe table I would expect to see values more like 1.2 or 1.3 in the higher pressure area and .5 in the lower pressure areas. I don't know about anyone else but I am easily hitting 1.2 g/rev at wot between .933 and 1.0 bar MAP.

I have experimented with a few areas of the table and seem to have been able to affect fuel trims with it. For example in an area where i consistently had + fuel trims, the table values were less than 1.0 (like .984) and by raising them, the fuel trims were reduced.

One of these days I'm going to flat fill this table with 1.0's and the closed loop target tables with 14.7 and go out for a long logging drive and see if my hunch is correct.

If anyone has any suggestions or information about this table or any table really I'm always open to comments / discussion either openly on the forums or via PM

[arghx7]

The long period of closed loop (i.e. near stoichiometric) is for fuel economy and CO emissions. You'll see more and more of that as time goes on, especially on engines that have been designed to handle high exhaust temps (water-cooled exhaust manifold).

I have a comment about the seemingly large fuel trims on a stock car.

During engine development, the MAF scaling and a lot of other tables were set using what's called "mean spec" parts. You know how when you buy aftermarket fuel injectors from certain vendors, they give you "flow matched" sets? It's a similar idea. Take the tolerance band of the part, and find one that's in the middle of the range. These are special parts right from the supplier. It includes front and rear o2 sensor, MAF sensor, injectors, etc. So what happens is, on production parts things start to drift. As long as they are within some specification, and the various long term (engine dyno & vehicle) tests confirm, they're ok.

One more thing to consider is the level of friction on the vehicle. The vehicle itself isn't "broken in" until 4000 miles. What I mean by that is 4000 miles on the vehicle is required for certification testing. In that time you decrease friction all over the vehicle. Even on an engine dyno, as the number of hours increases, you can see an increase in output and might be able to measure a decrease in friction mean effective pressure.

One implication of break-in is in the tuning of the electronic throttle for idle behavior. Electronic throttle typically has various learning values, similar to fuel trims. The idle air trims change as the friction decreases on the engine. The feed-forward idle airflow lookup tables need to reflect this.

Just think about what you would do if you had to tune 10s of thousands of engines that you'll never physically see and haven't even been built yet.

[mad_sb]

Originally Posted by arghx7

The long period of closed loop (i.e. near stoichiometric) is for fuel economy and CO emissions. You'll see more and more of that as time goes on, especially on engines that have been designed to handle high exhaust temps (water-cooled exhaust manifold).

I have a comment about the seemingly large fuel trims on a stock car.

During engine development, the MAF scaling and a lot of other tables were set using what's called "mean spec" parts. You know how when you buy aftermarket fuel injectors from certain vendors, they give you "flow matched" sets? It's a similar idea. Take the tolerance band of the part, and find one that's in the middle of the range. These are special parts right from the supplier. It includes front and rear o2 sensor, MAF sensor, injectors, etc. So what happens is, on production parts things start to drift. As long as they are within some specification, and the various long term (engine dyno & vehicle) tests confirm, they're ok.

One more thing to consider is the level of friction on the vehicle. The vehicle itself isn't "broken in" until 4000 miles. What I mean by that is 4000 miles on the vehicle is required for certification testing. In that time you decrease friction all over the vehicle. Even on an engine dyno, as the number of hours increases, you can see an increase in output and might be able to measure a decrease in friction mean effective pressure.

One implication of break-in is in the tuning of the electronic throttle for idle behavior. Electronic throttle typically has various learning values, similar to fuel trims. The idle air trims change as the friction decreases on the engine. The feed-forward idle airflow lookup tables need to reflect this.

Just think about what you would do if you had to tune 10s of thousands of engines that you'll never physically see and haven't even been built yet.

I completely understand the reason for the long closed loop. I have also seen cars holding closed loop under full boost and audibly deting on the dyno with stock tunes, subaru's and others. While I'm not boosted and hence not too worried about detonation at this point, repeatable performance is my #1 priority not emissions or fuel economy. I would much rather control closed to open loop transition via the fuel map and keep the long term trims to a point that does not affect my tuned afr when wot.

Your point about the mean spec parts certainly make sense and that is sort of what i was getting at when i said the stock scaling didn't work well for my car. My sample size is exactly one Though i have spoken with at least one tuner who said he has seen the same thing on other cars.

I had hoped to have others chime in with what they have seen / are seeing with their long and short term trims both un-tuned and tuned.

Ideally, I would like to run closed loop all the time so the ecu is constantly adjusting trims to hit my target afr at every rpm and load combination. With enough people trying various things out we may eventually get to the point that we can do that without issue. In Fact, I may already have enough information to do it if i get motivated enough to sit down and do it.

[arghx7]

I just saw your update.

To make a long story short, there's a very good chance those A & B closed loop targets are there for emissions purposes. Basically, you have to offset the target AFR based on feedback from the rear O2 sensor. The two maps may be a feedforward portion of the calculation, and then the rear o2 compensation is on top of that. A lot of it has to do with the fact that the front O2 sensor is a wideband and the rear o2 sensor is a narrowband.

So the two maps are probably for the catalyst oxygen storage capacity control and NOx emissions countermeasure. The maps are set You probably don't have access to the oxygen storage capacity controls with BRZedit (or ECUTek for that matter), but trust me they're in the ECU somewhere. If the target AFR isn't set right, the engine will fail NOx on the LA4/FTP test cycle.

[mad_sb]

Originally Posted by arghx7

I just saw your update.

To make a long story short, there's a very good chance those A & B closed loop targets are there for emissions purposes. Basically, you have to offset the target AFR based on feedback from the rear O2 sensor. The two maps may be a feedforward portion of the calculation, and then the rear o2 compensation is on top of that. A lot of it has to do with the fact that the front O2 sensor is a wideband and the rear o2 sensor is a narrowband.

So the two maps are probably for the catalyst oxygen storage capacity control and NOx emissions countermeasure. The maps are set You probably don't have access to the oxygen storage capacity controls with BRZedit (or ECUTek for that matter), but trust me they're in the ECU somewhere. If the target AFR isn't set right, the engine will fail NOx on the LA4/FTP test cycle.

Yeah, that makes a lot of sense. I remember the 32bit wrx ecu's would go a couple decimal points richer if you removed the rear 02 even if you patched the rom to turn off the CEL.

I know you can log catalyst temperature, even though i assume it is a calculated value. I assume they are looking at the out put of both 02 sensors to determine cat temp and then adjusting target afr based on the desired temp. I guess thats my version of what your saying lol.

[mad_sb]

Ideally, i would like to adjust the long term fuel trim rages (to cap them to say + / - 5%) and the long term trim maf bin's (the gram/sec break points for idle, low, medium, and high trims) so that i can leave the trims enabled and not muck around with the emissions.

Sadly, the bins are not defined yet for BRZedit, not sure about ecutek, and the currently defined long term trim man and min value does not seem to have any effect for my rom.

[arghx7]

Originally Posted by mad_sb

Yeah, that makes a lot of sense. I remember the 32bit wrx ecu's would go a couple decimal points richer if you removed the rear 02 even if you patched the rom to turn off the CEL.

I know you can log catalyst temperature, even though i assume it is a calculated value. I assume they are looking at the out put of both 02 sensors to determine cat temp and then adjusting target afr based on the desired temp.

The catalyst temperature estimation is first for OBD/CARB purposes. It could be used for detecting and diagnosing catalyst light off.

While it may be accurate enough for that, I'd be surprised if catalyst temperature were being used for normal heavy load enrichment without an actual EGT sensor. I'm talking about a sensor similar to what you'd find on catted uppipe turbo Subarus.

Usually, for heavy load enrichment control you have the sensor in front of the cat. Then a model estimates the actual cat temperatures, which is pretty complicated. The model has to be tuned according to time delays and gains and such. The cat itself (middle of the bed) is colder than the engine-out exhaust temp during cold start. It's basically like a heat sink. After the cat has warmed up though, it's the hottest part of the exhaust (except maybe preturbo if the car has one). The cat temperature also drops as mileage accumulates on it.

[ft_sjo]

I wonder what happens with that calculated cat temp model as you change other parts of the calibration, or perhaps it doesn't? I wonder what affect that'd have either way.

[Guff]

[mad_sb]

Originally Posted by ft_sjo

I wonder what happens with that calculated cat temp model as you change other parts of the calibration, or perhaps it doesn't? I wonder what affect that'd have either way.

I'll have to check and see what the logging penalty is for enabling it, if there is no hit with the other parameters i have enabled i"ll turn it on next time.

[Wonderbar]

I'm def digging out my tactrix cable.

[shadow]

Originally Posted by mad_sb

Here is one more table that seems to have an affect on trims, all be it indirectly.

I'm not 100% sure what this table is, but I believe it is a load scalar table. So, whatever the ecu calculates the load to be based on maf is multiplied by the values in this table to determine the final load.. but that's just a guess.

It does not seem to be speed density fail-safe table just based on the fact that the higher vacuum (lower manifold pressure) area have larger values than the higher load (higher manifold pressure). If it were a fail safe table I would expect to see values more like 1.2 or 1.3 in the higher pressure area and .5 in the lower pressure areas. I don't know about anyone else but I am easily hitting 1.2 g/rev at wot between .933 and 1.0 bar MAP.

I have experimented with a few areas of the table and seem to have been able to affect fuel trims with it. For example in an area where i consistently had + fuel trims, the table values were less than 1.0 (like .984) and by raising them, the fuel trims were reduced.

One of these days I'm going to flat fill this table with 1.0's and the closed loop target tables with 14.7 and go out for a long logging drive and see if my hunch is correct.

If anyone has any suggestions or information about this table or any table really I'm always open to comments / discussion either openly on the forums or via PM

That table looks very close to the STI's Engine Load Compensation Table. EL Comp uses rpm and manifold pressure to modify the amount of fuel that the ECU commands based on MAF sensor readings. So you are right in your assumption of how it works.

On the STI, this table is used to fine tune for intake manifold resonance and fuel rail resonance. More about the topic on iwsti.com and romraider.com forums. The latter has utilities to help tune a CSV log.

Again on an STI, I try to smooth the MAF sensor scaling, then use EL Comp to correct the mixture and make driving smooth.

[Sportsguy83]

Originally Posted by mad_sb

Reserved for Maf Scale How To

I would like to see this