It is quite simple. You need to check what is the voltage for 0% on the TPS check analog input, then what is for 100% and interpolate linear (as you have simple linear sensor).
Then you can move the throttle and check what is the value of TPS check error.
It should be about 0.1-0.2V. Look what is the biggest error, add a little (eg. 0.15V) and it should work perfect.
To activate the error-checking function between the main sensor and the check sensor, it is necessary to assign an analog input for the check sensor in the TPS/PPS configuration. Next, a 2D check tolerance table should be defined. In this table, for a given position of the main sensor (in %), the expected voltage from the check sensor is specified. If this voltage differs from the actual voltage by more than the value specified in the error tolerance paramete, the ECU will detect an error and deactivate the throttle power supply.
I think the name of the table is confusing better would be to call it Check voltage vs throttle position. is more self explanatory. tolerance suggest that should be small value there …