Crust index in Erdas

[felixio]

hey people, i have the trouble with CI index in Erdas, when i am useing the image from landsat 5 tm with atmospheric correction it gives me the whole white picture in the results,

the formula CI=1-((red-blue)/(red+blue)) also i tryed to change R band with NIR, but......does not work at all

when i tryed to insert in the model the bands seperately the reslusts are the same, just white pic,

can anyone help with this type of issue?