AP goes "stand-by" after a period of SM inactivity

There is an AP in our cluster which serves only one SM. After a few hours of a client inactivity the Internet became inaccessible for the client till pinging the AP. Internet is accessible after ping. They do it every morning.
There is nothing significant in the AP or the SM logs.
Is it a bad AP or some kind of wrong configuration?

Check the “Bridge Timeout” configuration on both units. Set them to the max, 1440, and see if the problem goes away.

Increasing of “Bridge Timeout” didn’t help. But the problem disappeared when I disabled the RF Public Network Interface of the SM. The RF interface parameters were the same as the NAT parameters.

I’d traced the network connection when the problem had occurred again. This is the picture of the situation:
1. The client have internet access. I can’t ping the SM.
2. The client lost internet access. He can’t ping anything behind the AP. The SM and the AP are on session. I can ping the SM.
3. The client tries to ping an outside host. I see his ingoing and outgoing traffic both at the AP and the SM. But the client has no response.
4. The client pings the AP and internet became accessible. I can’ ping the SM.

Can somebody explain that?