Dynamic Fuel Tables

[Redline Motorsports]

The stomp impact factor table notes..."fraction of fuel that will remain on the intake port after the fuel has been injected".....the table is a "factor" table that is multiplied to what? I would assume that lowering the values would help a "rich" tip in of fueling and vice versa.

This also leads to the function of the Evaporation Factor table. How do we approach both of these tables properly?

Seems as if we can keep the engine at a steady state with a lambda on 1, we could do a throttle "tip in" and monitor the actual lambda and watch the recover time back to the target lambda to deal with the excess (or lack of) fuel during that transition.

Anyone have any conclusive insight to these tables?

Howard

[gmh308]

The stomp impact factor table notes..."fraction of fuel that will remain on the intake port after the fuel has been injected".....the table is a "factor" table that is multiplied to what? I would assume that lowering the values would help a "rich" tip in of fueling and vice versa.

This also leads to the function of the Evaporation Factor table. How do we approach both of these tables properly?

Seems as if we can keep the engine at a steady state with a lambda on 1, we could do a throttle "tip in" and monitor the actual lambda and watch the recover time back to the target lambda to deal with the excess (or lack of) fuel during that transition.

Anyone have any conclusive insight to these tables?

Howard

Simon is probably deeper on these than anyone I have seen.

[swingtan]

You know that line that shouldn't be crossed...... it's behind you.

Dynamics are a fairly tricky set up. The basic understanding is not that hard to grasp, but how the tables then work to maintain correct fueling is somewhat harder.

Specifically regarding "stomp", we have....


The stomp fueling is an "adder" to the normal calculated dynamic fuel load (Impact factor).
The value of the stomp fuel {B2005} is a "factor" because I believe that the value is added to the "Impact factor" {B2012}. So the "units" must remain the same.
Stomp fueling is only added once the calculated dynamic fuel load, exceeds a preset level {B2003}.
If stomp fueling is being added, the calculated dynamic fuel load needs to drop below {B2003} before stomp fueling is removed.
Stomp fueling removal is controlled by, {B2004} the rate of removal and {B2002} the hard cutoff limit.
Because of the cut off limits, I think stomp fueling is ignored on throttle close.


So stomp fueling is is controlled by the impact factor, which the stomp fueling adds to....

This should indicate that while the tables are worded like "the amount of fuel left on the intake port wall", it should be read as "how much additional fuel needs to be injected to compensate for the amount of fuel left on the intake wall and not delivered to the combustion chamber." So higher numbers result is more additional fuel being injected.

Here's the big "gotcha" though. If the Impact factor is set too low, then the calculated dynamic fuel load may never reach, or the levels needed to enable the stomp fuel tables. So increasing the stomp fueling may have little effect, until the required impact levels are met. Then you may have way too much fuel though, because previously the stomp tables were not being used.

The lesson is, that you can't just use the stomp tables alone. You need to look at both the Impact tables and the Stomp tables together to work out how much fuel will be added. Now for the kicker, as the stomp tables are activated by "calculated dynamic fuel load", the evaporation tables also come in to play. If the ECM thinks fuel is evaporating slowly, the "calculated dynamic fuel load" may increase until a point of equilibrium. So a slower predicted evaporation factor is more likely to give a higher calculated dynamic fuel load and so enable stomp fueling. A faster predicted evaporation factor will then be more likely to not trigger stomp fueling.

One of the biggest issues I've found when looking at dynamics, is the lack of any log data to accurately indicate the dynamic fuel load. Yes, you can look at IBPW and the WB signal, but it doesn't help you with differentiating between Impact fueling and stomp fueling. You also need a very stable and repeatable test bed to get accurate data, something that's possible on a dyno, but still difficult.

I'd love to have an "E38.DYNFUEL" PID that gave the ECM's predicted dynamic fuel load. Then you could use this with the wide-band to see if what is going on.

Simon.

[gmh308]

You know that line that shouldn't be crossed...... it's behind you.



Keh ? :)

[swingtan]

I'm thinking of doing a dynamics guide for the E38, but I really need to know a bit more. I'm still guessing on a few things because of the lack of data that can be logged. I'm down to trying to think of new ways to get a total view of the dynamic fuel load.

[GMPX]

I'm still guessing on a few things because of the lack of data that can be logged.
To log such things obviously requires DMA PID's. Defining DMA PID's is one of the more time consuming tasks we have to do, made worse on the E38 & E67 because there is over 70 operating systems.

[swingtan]

Yep, I understand that, hence why I'm trying to think outside the box on this. I'm toying with importing all the details into a spread sheet to see if it starts making sense.

Simon

[5.7ute]

Yep, I understand that, hence why I'm trying to think outside the box on this. I'm toying with importing all the details into a spread sheet to see if it starts making sense.

Simon

Have fun with that.:grin:
When I was playing around with the LS1 stuff there was a lot of steps I tried to get things lined up. One problem was precision which I got around with calc pids. The other was working out the scaling factors for all the pids so I could use the PCM,s values. Eventually my brain blew a fuse & I never got back to it.

[swingtan]

That looks like a challenge......

[hymey]

I have had some speed density cars have some issues with throttle movement at low rpm in N, With MAF tuned engines the engine rpm may rise from 1000-3000 rpm say quite fast but the overall change in MAF frequency is not that great, . I always thought speed density tuning to be a problem with dynamics, but overall I do not believe that it should be. If you look at the HSV calibration the impact factor multpiler is set to 1.00 across the board, and the Holden calibration goes from 0.6 to 0.8 as g/s increases.

The vacuum in the engine changes very abruptly and rapidly from say 40 kpa(idle) to 100 kpa, Everytime there is a change of load dynamics takes effect.

So if I blip the throttle from 1000-2500 there will be a very fast rapid change and hence dynamics will take play in this, with the MAF connected there is a rapid change but it is very gradual, change of airflow, not one abrupt change of vacuum.

I always added more fuel to compensate this, the stomp tables do add to impact but they work from sudden throttle movement. I found I could start an engine hot and for a period of around 30 seconds there would be no change to stomp. Even if I increased it very high. It would take around 30 seconds of engine running to kick in. From this I learnt that because of the rapid change of map with fast throttle movement, that the extra fuel from stomp compensation was not required, dynamics(impact) requires a lot of work but stomp itself I reduced a lot over standard settings. Basically I removed the fast TPS movement out of the equation and just focused on impact alone and this has been a big improvement. To be honest the factory HSV calibration is now used but I make a few small changes.

I also found that if I increase impact by say 30% then fuel economy increases by nearly that amount. If I go back to stock impact the economy improves, and by going much less on stomp compensation and getting impact right aswell as PE settings I have the best fuel economy figures to date. The MAF tune requires stomp more so as previously explained the increase in airflow is more smoother gradual and less abrupt then MAP so it requires throttle movement to add extra fuel, The MAP alone does a good enough job of this.

I learnt this tuning import Nissans ten years ago, when a lot of cars guys didn't wire up TPS sensors and just MAP sensors. They ran dirty big aftermarket throttle bodies and MAP only tunes, there were no accelerator pump settings. They were very very responsive with no accel pump compensation via a tps input. And cars that had TPS input with this feature like a microtech, well I had to reduce the amount of fuel quite substantially.

The accelerator pump idea comes from carburettors. Carbs use a cam which operates a squirter, which you can alter the stroke of the pump via the cam and the duration of the shot of fuel via the diameter of the squirters. Carburettors require these as the engine requires air speed and vacuum drop in the venturi to pull fuel from the boosters, So all the pump does is allow the engine to rpm and give it fuel mechanically so the boosters have time to atomise the fuel in the venturi then make it to the runners, then the pump is tuned to allow smooth transition.

I can see how the thinking come across to efi but in general it does not require no where as much accel pump simply because we program injector pulsewidth to supply fuel at every correct load point, it doesnt require fuel to be drawn out of the booster and atomised, it is programmed to do it.

[swingtan]

I think the biggest difference between MAF and SD, is that the MAF will always indicate air flow and the MAP sensor will always indicate pressure. While they can be used to indicate the amount of air the motor is consuming, they do it in different ways. If you plot both MAF airflow and calculated MAP airflow (see here... http://forum.efilive.com/showthread.php?16413-Calculating-MAF-Airflow-From-VE-Table..CALC-VET-In-Reverse ) then you will see that the MAP based airflow always lags the MAF on transients. This is true for both throttle opening and closing.

The MAF also shows a big "burst" of airflow on sudden WOT events, which the MAP doesn't. This occurs as the manifold fills with air and is not actually entering the motor. But the ECM still seems to not over-fuel for this event, which I think is the wall wetting delay coming into effect.

Here's a quick image of how the MAP airflow lags the MAF.
10920

You can see that the MAP based airflow at times misses some throttle blips, even though there is a change in the measured MAP pressure and gm/Cyl. I think what it really points out is that dynamics settings that work with a MAF, may not even come close when working in SD.

Simon

[hymey]

I think the biggest difference between MAF and SD, is that the MAF will always indicate air flow and the MAP sensor will always indicate pressure. While they can be used to indicate the amount of air the motor is consuming, they do it in different ways. If you plot both MAF airflow and calculated MAP airflow (see here... http://forum.efilive.com/showthread.php?16413-Calculating-MAF-Airflow-From-VE-Table..CALC-VET-In-Reverse ) then you will see that the MAP based airflow always lags the MAF on transients. This is true for both throttle opening and closing.

The MAF also shows a big "burst" of airflow on sudden WOT events, which the MAP doesn't. This occurs as the manifold fills with air and is not actually entering the motor. But the ECM still seems to not over-fuel for this event, which I think is the wall wetting delay coming into effect.

Here's a quick image of how the MAP airflow lags the MAF.
10920

You can see that the MAP based airflow at times misses some throttle blips, even though there is a change in the measured MAP pressure and gm/Cyl. I think what it really points out is that dynamics settings that work with a MAF, may not even come close when working in SD.

Simon

Simon,

I found quite the opposite when tuning a LS2 with a smaller injector and cathedral port runner with the e38. I had to pull fuel away from stomp compensation. It is not the fact that the dynamics is slow at all that is not the issue. It is because the engine goes from say 51 kpa to 88 kpa at such a rapid rate which causes VE to change instantly. Dynamics here especially ETC and g/cyl is a little out of the equation. Airflow is gradual and climbs slowly and gradually in g/cyl, But with MAP the output changes instantly. VE goes from say 1200 to 2200 in the blink of an eye and all of a sudden the pcm is commanding more pulsewidth, So I have had to 0 all stomp compensation to stop the engine overfueling. This was very prominent with larged cam engines. The g/cyl vs ETC airflow was out of the equation but it is relying on VE load values. So I have found. More to the point, The engine gets a big burst of fuel then has to recover. It requires less stomp compensation, In the the event you blip the throttle with MAF it does so fill the cylinder, but with the dynamics compensated for it requires more stomp compensation as the increase in the MAF scale is more gradual, ie it is not going rapidly from 30kpa-100kpa. I basically put all the logs out of the equation in the end and used old school methods of how to set up an accelerator pump. The things I spoke with you the other day were solved by some VE settings but mostly with setting up the stomp values to zero. But every engine is different maybe the next one will throw a different curve ball. I think dynamics is one area that needs a little work. It is not fully understood by most yet. How fast PE enters helps the WOT side of things dramatically and removing stomp compensation all together helps to, so overall I am using much less dynamics then I initially required. If you set impact and stomp to high and you race at the drags, You will see the car leave the hole and even though a gear change wasnt performed the car will load up(fuel up a little) when it changes from say 5000 rpm and 1000 kpa on the launch to 3500 when it hooks up. Even though the engine has never left power enrichment. The change in VE along with to much impact and stomp has created some overfueling. On a dyno pull the airfuel ratio is fine and even looks good on the launch but with fuel not atomising and running down the cylinder wall wet it creates a poor burn. Reducing the stomp prevents overfueling when the throttle is depressed quickly on the launch then when it bogs is because impact is to high. Values of around 0.15 seam to work very well in impact, Above 0.2 we found the car bogging on the launch and lost et. The car picked up with factory dynamics and the MAF in place. It didn't bog but with lost mph. At WOT shifts with big airflow changes dynamics works wonders, to overcome this with SD I increase how fast it enters into power enrichment which solves it, but once it is in PE dynamics is still working and any change in load will cause excessive wall wetting on change of injection event. This is where you notice SD on WOT shifts causing problems, but keeping it around the 0.12-0.15 setting works here(with multiplier of 1) and stomp values to zero. But again a guide only. Dynamics is a sensitive thing that needs as much alterations is some cases as MAF or VE tuning does alone.