Hi,

This is my scenario :

AP with RouterOS 5.2
RB333 fw 2.18 + R52Hn + Antenna AirMAX AM5G19
Protocol: Nv2
TDMA: 2
Cell Radius: 10
Client Tx/Rx -57/-59

Client with RouterOS 5.2
RB411 fw 2.29 + R52n + Antenna Dual Pol 20dbi

This client is disconnect several times with : lost connection, medium-access timeout (client side) - disconnected, not responding (ap side)
I tryed also these settings:
Hw. Retries : 15 (Same problem)
Periodic Calibration (disable or enable) (same problem)
RouterOS 4.17 Client/Ap Side (same problem)
Disconnection is most frequently when the client use skype.

Are you using an extension channel? I have a similar issue with a pair of SXT's I am trying to resolve, and for me the issue seems to be tied to both chains in TX/RX and an extension channel enabled. If I disable extension or disable TX or RX on one of the chains, the stability clears up.

Hi,

No, extension channel was disable in all scenarios. I installed other 8 CPEs, I fixed the problem with :
Downgrading to 4.17
Using old wireless driver
Periodic Calibration: Enable
Hw Retries: 15
Adaptive Noise Immunity: None
Extension Channel: Disable

Thank to Mikrotik support, practically useless as the new wireless driver (nv2)...

You aren't the only one having issues with this. I opened a ticket on it Monday and haven't received a response to the ticket yet. Normis did reply to my thread and say they can't reproduce it and that V5.3 may fix it. Here is that thread.

http://forum.mikrotik.com/viewtopic.php ... 74#p264974

Chadd

Wow chadd

Switching to 5.3, timeout problem persist, new bugs.... wow wow
We are moving one of our tower towards UBNT.

Wow chadd

Switching to 5.3, timeout problem persist, new bugs.... wow wow
We are moving one of our tower towards UBNT.

I am about to do the same. I keep hanging in there with the hope that a great fix is around the corner but it doesn't seem like it is in the cards.

I have been fighting MT wireless disconnects be it "extensive Data lost", "no beacons received", to now "medium access timeout", "control frame timeout" and "too many poll timeouts" for years now and I have finally come to the realization that this isn't something MT cares to or can fix.

It would be different if this was happening on clients with poor connection strengths and CCQ but I have clients with -63/-59 and 88%/97% CCQ randomly disconnecting for time out issues. There is no rhyme or reason to it and MT can't reproduce it so where does that leave us?

Chadd

Hi,

Chadd if you have clients with -63/-59 and 88%/97% CCQ with random disconnection try to set in client side Antenna Gain to 17.
I have 3 clients (SXT 5HnD) with the same problem after moving Antenna Gain (Client side) from 10 to 17 I fixed the issue.

Are you using an extension channel? I have a similar issue with a pair of SXT's I am trying to resolve, and for me the issue seems to be tied to both chains in TX/RX and an extension channel enabled. If I disable extension or disable TX or RX on one of the chains, the stability clears up.

Extension channel adds another XXmhz channel above or below the current channel. If you're turning it on and having issues, make sure that XXmhz channel above or below your current channel is free and not used. You could be turning it on and getting interference from another link.

Extension channel adds another XXmhz channel above or below the current channel. If you're turning it on and having issues, make sure that XXmhz channel above or below your current channel is free and not used. You could be turning it on and getting interference from another link.

In my case, both channels were clear. This weekend I upgraded both radios from 5.2 to 5.4 and my timeout issues with extension channels cleared up. If you haven't tried yet, see if 5.4 works better for you.

the same problem here, 2xRB411AH with R52Hn, extension channel disabled, (20MHz only), frequency 5700, signal ~ 50-62 on both sides and both chains, ROS 5.14. Jirous EXTREME 24dB antennas, 600m ptp link.. Strange... units were connected together, however, client not didnt pass any data thru, wasnt even pingable until I went back to 802.11a mode. Both units acting like bridge. Then it went to normal, but link troughput went to 42mbps, as supposed to with "a" mode and nstreme. The problem appeared after several hours of normal link operation.. cca 15hours. Im gonna to stick with old stable protocols

Wow chadd

Switching to 5.3, timeout problem persist, new bugs.... wow wow
We are moving one of our tower towards UBNT.

I am about to do the same. I keep hanging in there with the hope that a great fix is around the corner but it doesn't seem like it is in the cards.

I have been fighting MT wireless disconnects be it "extensive Data lost", "no beacons received", to now "medium access timeout", "control frame timeout" and "too many poll timeouts" for years now and I have finally come to the realization that this isn't something MT cares to or can fix.

It would be different if this was happening on clients with poor connection strengths and CCQ but I have clients with -63/-59 and 88%/97% CCQ randomly disconnecting for time out issues. There is no rhyme or reason to it and MT can't reproduce it so where does that leave us?

Chadd

From my own experience NV2 is great but is prone to disturbance by local interference, I solved my issues with NV2 by providing better signal attenuation from each antenna, this is hardware related and MT cannot address unless they release a special NV2 antenna. Until then we will have to solve our own hardware or location (or both) NV2 issues

5.9 had major improvements in signal to noise levels under NV2.

5.14 addressed NV2 disconnects.

If you haven't tried 5.14 NV2, give it a go.

Also, if the link isn't stable with nstreme or 802.11 modes, it's not going to be stable with NV2.

I am having an issue with a microcell AP we have installed. This is the setup: RB433 using Nv2,16 dbi 90 degree sector on a tower. RB411L in "station bridge" mode that is being powered by a UBNT nanostation (PoE Passthrough). the signal from RB411L to tower is -57/-63 on channel 2412. the Nanostation is set in access point mode on channel 2462 with 5 clients hanging off it running airMAX. I am continously getting "lost connection, sychronization timeout" and "medium-access timeout" on Mikrotik to tower. all clients have public IP's from said tower. Our company is moving towards UBNT but are not ready to overhaul that current tower (hence the mikrotik because of Nv2 TDMA protocol). I feel this is a great enlightenment opportunity to find some information. Thanks to whomever may respond.