Darkshot

iTrader: (4)

I'm currently trying to teach myself how to tune AE. I understand the modifications to the AE-TPS PW and AE-MAP PW as simply modificitions to the base pulse width (in milliseconds) for any given delta MAP or delta TPS. However, I do not understand the TPS and MAP filter vs Coolant temp. What exactly does this modify and for what reason? I'm just trying to visualize the physical effect it has on the engine so I can dial it in. How can I accuratly tune these two tables?

RBob

iTrader: (1)

The AE-TPS & AE-MAP PW tables are based on a delta value. The change in the TPS & MAP over time. This change, or delta value is used to define the row from the AE-TPS & AE-MAP table.

To get this delta TPS & MAP each is filtered with a 1st order lag filter. This 1st order lag filter uses a coefficient that defines the level or amount of filtering. A large coefficient provides for a fast filter. The output of the filter closely follows the input of the filter.

A small coefficient value has the opposite affect. The output of the filter lags the input value. With the smaller the coefficient, the greater the lag.

The TPS and MAP filter vs Coolant temp tables provide the filter coefficient.

A key part of the lag filter is that it takes place on a time based schedule. On regular intervals a new output value is calculated. So if the input value holds steady (the same value), the output value will eventually be the same as the input.

With the lagged output value, and the current value, subtraction provides the delta between them. This delta is then used for the AE PW lookup (it defines which row in the table to use).

For a TPS example, say we are sitting at a stop light, and the last yellow blinks, the throttle is mashed and the car starts moving. While sitting still the throttle was at 25% TPS, then quickly went to 100% TPS. At this point the output value of the lag filter, or old value, will be 25%. Because after a short period of time after the throttle was brought to 25%, the filter output will have caught up to the input value.

Once the throttle is mashed the immediate delta TPS value will be 75% (100 - 25). This value of 75% will be used to define the row in the AE-PW table. As most AE-PW tables don't go that high, the highest row in the table is used.

Also, and this is an important concept, the real delta that the ECM sees and uses will probably be lower. The reason is that the ECM is reading the TPS sensor and lag filtering it every 12.5 milli-seconds, or 80 times a second.

So as we are mashing the go-pedal the ECM is filtering the TPS. This has the input of the filter going up in value as the throttle moves from 25% to 100%. And the output of the filter doing the same.

Even as such, the initial delta TPS values will be high. Then as the throttle is held at 100%, the output of the lag filter is catching up to it. With that the delta TPS value (current - lagged) is getting smaller in value.

As the delta TPS reduces, the AE-PW is tapering off. And once it reaches 0 the AE stops. The action of AE is calculated within the ECM 80 times a second, or every 12.5 milli-seconds. Basically as such:

Read the TPS sensor (new or input to filter).
Lookup the Filter Coefficient from the table vs coolant temperature.
Filter the TPS using the current TPS % and the old value (from the last time it was filtered).
Lookup the AE-PW from the PW table.
Do the Coolant and RPM lookups and adjust AE PW
Program the AE PW into the ECM hardware

Then 12.5 milli-seconds later, do it all over again.

Changing the filter value affects the AE in 2 ways: duration and magnitude.

A smaller filter coefficient will have the output of the filter slower to catch up to the input of the filter. Hence increases the duration of the AE.

With the slower filter (smaller coefficient) the delta will also be greater. This provides for a larger volume of AE. Note: this really depends upon how the AE-PW tables are set up. Usually they are incrementing as the delta increases. But do not have to be.

I tend to use the filter table to control the duration of the AE. And the PW tables to control the volume of AE. However, it is important to understand that the filter values also affect the volume of AE.

Various stock GM ECMs have different AE logic. The '7747 has both a TPS & MAP filter table. Where the '8746 only has a MAP filter table.

The other variable in the filter tables is the engine coolant temperature. This defines which filter value to use. Simply put, a colder engine generally requires more AE for a longer duration. So typically smaller filter coefficient are found at the colder table entries.

I'll see if I can get some graphs of this in action.

RBob.

{edited for spelling and grammer}

RBob

iTrader: (1)

Attached picture is a graph of 2 starts. The bottom axis is hours:minutes:seconds.

The dTPS & dMAP are the delta values for each. Not the AE PW, but the delta used to lookup the AE PW.

In the first takeoff the throttle is just mashed from fully closed (nearly fully closed as setting my foot on the go-pedal opened it a hair).

In the second takeoff the engine is brought up against the convertor, a little. Then the throttle is quickly opened.

From the graphs the interaction between the TPS & MAP are to the delta TPS & MAP values can be seen. Even the duration of the MAP AE is higher then the TPS AE.

Please note that this is on a port injected vehicle. A TBI vehicle may be a little different in that the delta AE values will be greater.

RBob.

devilfish

Good info. Thanks

vernw

iTrader: (2)

No, that's EXCELLENT info!!!

escort_ford84

What are the addresses for the filter values? I dont see it in the anlu hac.

RBob

iTrader: (1)

The '8746 (anlu) doesn't have tables of filter coefficients for AE. The delta TPS is based on the reading 12.5 msecs prior to the current TPS reading. For MAP, there are 2 set filter values. And a temperature threshold to choose between them:

RBob.

escort_ford84

So, would raising the value of the MAP AE filter delay or add some time to AE?

escort_ford84

and i suppose the values are in non-inverted a/d counts

Colts45s

so on an 8746, can the coefficient change in dmap adjust AE PW volume or is it mostly duration do to the lack of a dtps filter? And if it is limited to mostly duration caould you explain a little more into the effect of the PW table changes to volume? To my understanding changeing the PW tables would effect both duration and volume the same way you described the AE tables am I correct?