Fuel Charge Cooling Coeff vs Charge Temp Approx Table

[DerekAE86]

Am I right to assume these two settings will interact with each other?

So if your Charge Temp Approximation Table is way out, changing the Fuel Charge Cooling Coeff will give you poor results, and vice versa?

What's the best method for tuning these? The helpfile says a typical value for petrol is a Fuel Charge Cooling Coeff of "10".
Should I just leave it at 10 and attempt to get the Charge Temp Approximation Table correct first?

Currently changing the Fuel Charge Cooling Coeff doesn't appear to have any effect for me unless I set it to 0, which turns it off?

Tuning for a steady Lambda of 1.000 with a Target of 1.000 currently drops me down to 0.850 when I target 0.900, but if I set it to 0 it will achieve the 0.900 target.

I'm confident all my Injector settings are correct. So that just makes me think my Charge Temp Approximation Table is way out and I need it to be influenced by ECT more than IAT in this region.

But wouldn't changing the Fuel Charge Cooling Coeff after getting the Charge Temp Approximation Table roughed in throw all out again? Is changing the Fuel Charge Cooling Coeff from the default of "10" when using pump U98 even necessary?

[Vaughan]

18 hours ago, DerekAE86 said:

Am I right to assume these two settings will interact with each other?

Yes Fuel Charge cooling is added on top of charge temp approximation to estimate the temperature of the charge air after the fuel has been added based on the target AFR.

The fuel charge cooling coefficient shouldn't have any effect at a target Lambda of 1 but does for richer or leaner targets.

In terms of how to tune it that is probably best answered by @Adamw

[Adamw]

Charge temp approximation is best tuned by varying the air temperature while holding coolant temp relatively stable.  On the dyno this happens quite naturally if you have decent fans etc.  The thermostat will keep the coolant temp relatively constant and you can vary the air temp by switching the fan off during a pull or ducting hot air into the intake or blocking intercooler airflow.  

Charge cooling can be more tricky as incorrect deadtimes can also have a big influence on the "lambda step" test.  I usually hold at medium load at about 2500 so im working at PW's that are outside of the non-linear region and then with the lambda target at 1.00 adjust a big patch in the fuel table so it is running at 1.00.  then step target to 0.95 and 1.05 and note down what the measured settles on for each target.  Tweak deadtime and charge cooling until I get measured lambda to follow target reasonably well in both directions.  After every adjustment you need to go back to a target of 1.00 and adjust the fuel table so it is running close to 1.00 again before doing the step tests again.  Incorrect deadtime tends to give you a different amount of overshoot or undershoot depending if you are going richer or leaner, whereas incorrect charge cooling tends to give you the same amount of overshoot or undershoot in either direction.    

 

[DerekAE86]

7 minutes ago, Adamw said:

Incorrect deadtime tends to give you a different amount of overshoot or undershoot depending if you are going richer or leaner, whereas incorrect charge cooling tends to give you the same amount of overshoot or undershoot in either direction.    

That's a nice tip. Do you find deadtimes change from what the injector manufacturer quotes, or is this more when you don't know the exact deadtimes?

 

[Adamw]

Deadtime has a large influence from the injector driver used to test the injector, so yes, the generic values the injector supplier quotes are rarely correct.  

[DerekAE86]

So it turns out this overshooting of the Lambda Target was actually a bad O2 sensor.

It was exaggerating the reported difference from stoich. Found it yesterday at the dyno when the car's O2 sensor showing a much large swing compared to the dyno sensor in the exhaust pipe.