[Mark McLeod]

With direct connection to a router the subnet mask really shouldn't hamper you.
Generally, the router only uses the subnet mask as a filter when a device doesn't appear in its own routing table and it has to decide what other routers are appropriate to send the packet to. Using network bridges (not APs) on the robots will cause the single router to populate its' routing table with all the cRIO addresses.

With the local listed devices in its routing table a router really acts as a simple switch. It doesn't have to be a router until it encounters a packet address that doesn't appear as a local device, then that packet must be routed it to other routers/APs to try to take care of. The routers tell each other what kind of local network address spaces they are handling and the local netmask defines whether packets of that type should be forwarded or not.
There are variations between manufacturers about how and when multiple routers/APs share their routing tables that can affect when the subnet mask is called into play. In a home-style network, like what we setup, the cutoff for when the netmask gets used usually occurs at the interface to an Internet service provider or a high school network router. Actually, regardless of the subnet mask, by convention the 10. private network doesn't get routed across major networks at all. So the school network is safe from our private robot traffic.

I run with both 255.0.0.0 and 255.255.255.0 all the time in training (but I admit my selection of home routers is limited to 4 or 5), so there's something odd going on.
What subnet variations does your new router offer?

Are you rapidly changing IP addresses on your devices while you run through variations?
That can play havoc with the routing table, as it takes time to decay and eliminate past addresses that a device used to be set to. Packets won't get routed to the new address until the router forgets about the old address. That aging out of past addresses differs by router manufacturer too.