There is no performance impact comparing to external redundancy diskgroup
Not true . We used NORMAL redundancy in one of our critical DB's 5TB Data Diskgroup. The performance was horrible due to excessive I/O from all those extra disks. We had to move everything to a diskgroup with external redundancy.
More disk don't cause more I/O, but increase I/O bandwidth. If you experience I/O starvation using ASM redundancy, it could be that it may not be configured properly, for instance, having different sizes of failure groups which result in unnecessary data rebalancing, or cross-linked storage arrays causing unnecessary disk controller traffic.