bizzare icmp request timeout

Good Day

I have the following setup

All equipment 433AH

CoreSwitch - 1(ether1 + radio WDS) -wireless link- 2(WDS radio + ether1) - switch - 3(ether1 + radio WDS) - wireless - 4(radio WDS + ether1) - switch (client)

I have tried to ping from 2 to 3 or vice versa … most of the time i am getting <1ms because both is connected to a dlink 8 port switch using lan cable but every 20 icmp or so, there will be 4 or 6 request time out.

I have replaced the switch and lan cable but problem persist.

Anyone can help please?

sample result

192.168.11.4 64 byte ping: ttl=64 time<1 ms
192.168.11.4 64 byte ping: ttl=64 time<1 ms
192.168.11.4 64 byte ping: ttl=64 time<1 ms
192.168.11.4 ping timeout
192.168.11.4 ping timeout
192.168.11.4 ping timeout
192.168.11.4 ping timeout
192.168.11.4 64 byte ping: ttl=64 time<1 ms
192.168.11.4 ping timeout
192.168.11.4 64 byte ping: ttl=64 time<1 ms
192.168.11.4 64 byte ping: ttl=64 time<1 ms


Anyone has any idea? I have increased the queue size to 100 on ether1 but it does not make any difference.