sheldon
|
Hi,
If you have access to ADE, look at the adc_8bit model, DNL is implemented in the model. One cavaet, you may need to add re-normailization so the output range with DNL is correct. INL is easier, it just requires implementing a non-linear transfer function from input to output. For example, V( mid) = V( in) * gain + HD2 * V( in)^2 + HD3 * V( in)^3 ... where V(mid) is the actual voltage that is used for the conversion. Depending on whether or not your ADC model is differential out, you can drop the HD2 term.
Sheldon
|