Windows 2008 NLB Cluster only reachable if I ping the cluster name before attempting to access.

I have 3 Windows 2008 Standard X64 servers setup in a NLB cluster with 2 physical Nics in each.
Nic 1 on each server:
Server1 Public IP-10.20.0.26/DG-10.20.0.1
Server2 Public IP-10.20.0.27/DG-10.20.0.1
Server3 Public IP-10.20.0.28/DG-10.20.0.1

Nic 2 on each server:
NLB NIC IP-192.168.75.5/DG-192.168.75.1

DNS entry has been created for NLB name with IP 192.168.75.5

Cluster Operation Mode: Unicast.

All servers are connected to a Cisco Catalyst 2950c switch.

Here is the issue:
Each server is reachable by server host name or host IP address with any app.  I'm trying Remote Desktop as these servers will be in a Remote Desktop Services farm and running a critical application.  I can't reach the servers with the NLB name or by IP with Remote Desktop (or any other app) until I ping the NLB name or IP.  Once I ping NLB name or IP I can make connections to the cluster with any app for about 30 seconds.  After 30 seconds I can no longer make any new connections using the NLB name or IP.  Once connected I don't get kicked off though.