Packet loss when using RB711UA-2HnD

RouterOS Wireless Networking
1

Hi,

I’m experiencing issues with a wireless IP Phone connected to an RB711UA-2HnD.

I’ve configured the RB with a simple configuration:

  • reset the configuration
  • remove default configuration
  • create bridge (add port ether1 and wlan1 to the bridge)
  • add IP address to bridge interface
  • enable wlan1 interface
  • configuration of wlan1:
  • ap-bridge mode
  • 2GHz-B Band
  • 20MHz and Frequency 2412
  • Security Profile : default (no encryption)

The wireless IP Phone does get connected to the RB, it does get an IP address.
But when the wireless IP phone pings its gateway (almost 50% of the packets are lost).

The odd thing is, if I connect a Linsys WRT54G it works straight away with basic configuration using 2GHz-B band and no encryption.
No packet lost.

Any idea ? I’m stuck on that one.

Thanks for any help

2

the phone supports only B mode?
Are you using both chains - do you have 2 antennas connected to the router?
Try setting the distance setting to indoors.

3

Thanks for the quick answer, much appreciated.

Actually the IP Phone supports B & G.
It works with B & G modes with the Linksys.

It doesnt work in G mode with the RB.
Following error is in the logs for this mode:
00:XX:XX:XX:XX:XX@wlan1: disconnected, received disaccoc: sending station leaving (8)
00:XX:XX:XX:XX:XX@wlan1: connected
00:XX:XX:XX:XX:XX@wlan1: disconnected, received disaccoc: sending station leaving (8)
etc…

RB is using both antennas, both are connected to the router.

Distance set to indoors doesn’t improve the loss rate.

I tried to enable both chain0 and chain1 but still lots of packets are still lost.

4

how far away from the AP that IP Phone is located. If it is located very near, then try to lower the output power.
What RouterOS version are you using?

5

AP is quite close indeed. But I tried to walk away and still lots of pack loss.
With a wifi analyzer that shows the strenght of the signal, it starts at -50dBm from close to the RB
to -80dBm while walking away still with the same bad results.

I tried to lower the Tx Power “all rates fixed” from 13dBm to 0dBm to -10dBm, same bad results.

RB was in version 5.3, it was upgraded to 5.22, and as it was showing the same bad results
it was upgrade to 6.0rc7, but still loss of packets while pinging the DNS server or gateway.

Thanks

6

Try using tx-power 10.
Make a support output file when you are sending pings to the gateway and they timeout. Then send that file to support@mikrotik.com

7

Thanks, file has been sent over.

I changed the Tx Power to 10dBm, still lots of packets lost.

8

Additional information:

After sniffing around, icmp packets are actually making their way all the way round.
from the IP Phone => to the RB => to the DNS server => to the RB

A sniff just between the RB and DNS server (actually on ether1 of the RB)
shows that icmp packets are flowing just fine.

For an unknown reason it is the last path between the RB and the IP Phone over wifi that doesn’t work well.

Sniff on the DNS server:

16:52:40.149655 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 53764, length 28
16:52:40.149817 IP dns-server > ip-phone: ICMP echo reply, id 13330, seq 53764, length 28
16:52:41.150540 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 54020, length 28
16:52:41.150692 IP dns-server > ip-phone: ICMP echo reply, id 13330, seq 54020, length 28
16:52:47.184896 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 55044, length 28
16:52:47.185059 IP dns-server > ip-phone: ICMP echo reply, id 13330, seq 55044, length 28
16:52:47.189218 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 55300, length 28
16:52:47.189358 IP dns-server > ip-phone: ICMP echo reply, id 13330, seq 55300, length 28
16:52:47.231278 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 55556, length 28
16:52:47.231334 IP dns-server > ip-phone: ICMP echo reply, id 13330, seq 55556, length 28
16:52:48.238903 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 55812, length 28
16:52:48.238963 IP dns-server > ip-phone: ICMP echo reply, id 13330, seq 55812, length 28
16:52:49.770744 IP ip-phone > dns-server: ICMP echo request, id 13330, seq 56068, length 28

Using the ping tool on the RB we have the following results:
Ping To: ip-phone
Timeout: 5000ms
Time Reply Size TTL
1949ms 50 64
1475ms 50 64
1515ms 50 64
2538ms 50 64
1574ms 50 64
1631ms 50 64
1613ms 50 64
2572ms 50 64

Again, there is no delay when the IP Phone is connected to the Linksys

9

I made some additional wireless captures using Wireshark.

  • IP Phone pinging its default gateway using a Linksys : capture-ping-gw-linksys.pcapng
  • IP Phone pinging its default gateway using a RB751G-2HnD v5.22 : capture-ping-gw-routeros.pcapng

Captures are available with attached zip file.

The odd thing is that when using the RB, multiple duplicate echo reply packets are sent back to the IP Phone.
We don’t have this while using the Linksys.

The same issue happens when the IP Phone pings its DNS Server which is beyond the RB. Packets look ok
until the reach the RB back and need to be proceeded over the wireless link to the IP Phone.
Duplicate packets happens then.

Any idea would be very much appreciated.
capture.zip (118 KB)

10

Problem solved.

Issue was coming from the UAPSD (Unscheduled Automatic Power Save Delivery) on the IP Phone.
Disabled it on both side (IP Phone and RB) and now everything flies…