I want to fit the EC50, and its standard error, not the logEC50.

In fitting dose-response curves, it is usually best to fit the logEC50. Why? But if you want to fit the EC50 of a dose-response curve, you can.

Enter data so X is concentration (or dose), and not the logarithm of concentration.

If you use Prism 7

The equation you want is built in. Choose one of the dose-response folders of equations. Note that some are for dose (or concentration) vs. response and others are for log(dose or concentration) vs. response. Choose an equation where X is expected to be dose or concentration, and Prism will fit the EC50 or IC50 directly, with standard errors. It will also report the confidence interval and Prism lets you choose asymmetrical confidence intervals which accurately portray the uncertainty. 

If you use Prism 6 or earlier

You'll need to enter the relevant equation as a user-defined equation. For stimulatory drugs, the equation you want is:

These rules usually work for computing initial values:

• Bottom=1*Ymax

• Top=1*Ymin
• EC50=1*(Value of X at Ymid)
• HillSlope=1*sgn(YatXmax-YatXmin)

That last rule just means: +1 if the curve goes uphill, -1 if it goes down.