Trying to set up Opiates testing by GC-MS. Trouble obtaining acceptable calibration line (range 200 - 16000 ng/mL). Using linear regression w/equal weighting.  Any suggestions?  If I try quadratic regression, what type of weighting should I use?  Thank yo

Linear regression with equal weighting does not produce an acceptable calibration line.  Trying to obtain a calibration range of 200 ng/mL - 16000 ng/mL.  Would quadratic regression be better?  If so, what type of weighting should I use?   Thank you in advance.