I performed a molecular dynamics simulation for a peptide-receptor complex using the Amber03 forcefield and TIP3 water model. The simulation is convergent in 50ns. Based on the Amber tutorial, using this force field is fine for protein simulation, but based on some references, ff03 is not a standard Amber force field. It's known to provide only an intermediate level of agreement with the experimental data. What are more simulations with the ff03 force field tends to generate ensembles that are distinctly different from other Amber force fields, based on Amber99. [Martín-García, Fernando, et al. "Comparing molecular dynamics force fields in the essential subspace." PLoS One 10.3 (2015).]
I would appreciate it if you would give me your experiences, also some references which used this force field for the all-atom simulation of proteins. Am I need to repeat my simulations?