torque model tuning s197

[Pappacosworth]

I'm tuning Kenne Bell charged 3V engine, and I'm trying to correct torque tables. I have MVPI1 Pro, and I have real torque signal from Dynomite chassis dyno in input 4, PLX map/boost in input 4. Am I correct that I should put data in Torque model - engine torque? In datalog threre is quite large IPC torque error values.
I use driver demand table from same year GT500.
Same challenges there are, it has Bosch modified 1000cc injectors that I can't find correct parameters. Maf is PMAS HPX-E, and there is plenty of reserve. Voltage went only to 3.2V
Boost is 125 KPa max, engine is MM built with locked custom cams. Hellcat fuel pump, I got that working ok.
Attached tune, and log. Log is not the latest and doesn't include real torque data

1.hpt 5r.hpl

[RobCat030]

This is how I do it. Just curve fitting and extrapolation. Personally I prefer to keep most of the axis resolution down below 1 and only dedicate two or three rows to boosted load, but that's a preference.


1 torque table.hpt

[Pappacosworth]

I don't know if it is related, but eq commanded has lean spikes near full power.
dyno1.png
12.hpl

[Thatwhite5.0]

I don't know if it is related, but eq commanded has lean spikes near full power.
dyno1.png
12.hpl

It's a related but it's not the cause. You'll also have throttle oscillation and spikes in airload and MAF lbs. I'd add the "torque" channels to your list, you'll get a better idea. Your torque/inverse table must calculate with each other. They have an inverse relationship. The inverse table should mirror the torque table and doing so will calculate together. The 3vs aren't as sensitive as the coyotes, they're forgiving and you just need to be ballpark.

torque-error.GIF
You can see why your torque errors are causing huge jumps in airload from .3 load to 1.0+. 2250 rpm to 3000 rpm can not and will not calculate correct with a 130 ft lb increase. You need to smooth that out and calculate the inverse. The table has to be some what realistic to the amount of airflow, air load, torque the cars actually producing.

[B E N]

Plot ETC throttle angle on RPM and Air load. Anything over a few degrees of error needs adjustment. Work it until you get all the errors dialed out. You will also need to adjust MAF failure load to get the AFR to all fall in line.

IPC seems to be driven by more than just torque model, raise the limit and focus away from it... it will fall in line as the rest of the tune comes together.

[Pappacosworth]

I got it much better, I smoothed torque tables. Throttle stays open, commanded and wideband eq are good. 630hp from crank at 125 KPa.

[Thatwhite5.0]

Throttle Angle error has a lot to do with Driver Demand. Some disagree, I say you can fix torque errors to a certain point with driver demand because I have. You can also fix torque errors editing MBT spark. Those are for minor torque errors. The original torque and inverse tables would never fall in line. They weren't calculated
IPC includes torque, DD and throttle body models, probably more than just those. You can adjust either one using angle error. Adjusting one too much will skew the model cause issues like OP.

IPC error on a 3v is different. They're easier to get running right, though, the error will never be at 0ft lbs. Torque errors on a 3v will be both positive and negative through out your logs. Some errors will cause drivability issues and those are what needs fixed. You can feel and hear the torque errors switch from positive to negative.

[B E N]

I agree about driver demand. It is fairly easy to histogram as well.