The standard BPDU timer is 2 secs - ie one BPDU every two secs. any switch that has been up for any length of time should have a high count on BPDU TX.

It does sound like you may have some issue with spanning tree.

How is your network built? Do you manually prune VLANs? use the switch trunk command to only allow VLANs where they need to go.

With 50 switches you also need to make sure the actual topology is sensible - the core switches do have the priority configured such that they are the root and backup root, don't they? rootguard can help with rogue root bridges.

What about portfast? make sure that is configured on all user ports. There are two main effects with that - the obvious one is the immediate trasition to forwarding, the other is that it suppresses the messages sent to the root bridge when a port goes up or down, which will reduce the workload on the root (and many other switches).