Hi,
I have just started tuning the neural network tuning system on the 2020 silverado 6.2. I have a supercharger with water to air intercooling fitted to otherwise stock internals.
Issues so far:
The 2020 silverado has a LIN BUS enabled MAF sensor. Barometric pressure, Intake air temp and probably a few other things are sent via LIN Bus to the ECM, the old "swap the intake temp sensor to the LT4 style MAP+Temp sensor" doesn't work. I have had some success modifying the manifold air temp model to bias more towards coolant but obviously depending on the level of heat soak in the intercooling system this doesn't read actual intake temp which confuses the model a bit. Anyone found a better solution for this? Is it possible to wire in the factory sensors to the E90 to enable SCIAP and temperature?
Having said that the stock cal holds together really well. I did have to re-tune the MAF curve at the top end (using stock airbox and MAF) as we are near the end of the range of the standard sensor and obviously the spark and fueling to get knock and EGT's under control but that was straightforward.
As far as VE tuning goes I have stock parameters which are "Volumetric Efficiency Aifrlow" and "Mass Airflow (Sensor)" I can just compare these two, because I know the MAF is good (comparing steady state LTFT+STFT and Widebands) so if I now tune the VE airflow to match the MAF airflow I should be right? Is this the method others are using? I don't see that I need to fault out the MAF etc and tune open loop as the older Gen4 system if I have these variables.
Also the neural network is pretty tricky. I just want to target the high load mid rpm VE range which is overpredicting, most of the VE airflow looks good, but it re-calculates the whole curve coefficients so I am always changing the whole map. Each time I spin the online calculator the numbers change a bit even with the same inputs so it's tricky to know if I am fixing my problem or causing more problems in other areas. The load axis is also just marked "Load" with 0-1. I changed this to 1.5 because I am roughly putting in 50% more air, but does anyone know what parameter is acutally used to calculate "load" is it cylinder air? Likewise with intake cam position, it appears to be referenced to something other than "degrees of advance."
It would be great if HP tuners had some method of predicting airflow from these tables or some sort of compare function.
Anyway if anyone else is having success using the neural network tuner on GM engines I would love to hear your thoughts.