After 3 days lost the connection again.
Supout file was sent to the support.
Ticket#2011080966000281
Dzieva.
turn down your power till you get to about -60 or so. I think you will find the radios are more stable when they are not driving max output.
I’ll try
Thanks,
Dzieva
I decreases 5db in tx power and continued the same problem.
anyway thanks
Dzieva
if you are not able to resolve lowering the power, you have to switch to a single chain.
Using single chain with very high signal (around -5x) is the only way to workaround disconnection at the moment using 802.11N in nstreme and nv2.
regards
Ros
Hi,
is qour tx ccq higher then rx ccq? → then
try to shield your antenna against neighbours, or separae it al least 1m from each other.
try to change channel and watch CCQs at the neighbour AP´s.
if you ar using a SXT try SXT RF shield form DCpower - it removes surrounded RF noise from your SXT by 10-20 db and reception of signal become very clear.
Impressed with shielding waveguide around the SXT however i would reduce the size further if the interference is co-location source(s) like other AP’s or Ptp on a mast this to have reduced windloading in exposed windy locations, also cannot tell from picture if aluminium is flexible in the center if yes on a windy day this could cause signal variations but the concept is good maybe other users could offer more practical advice on this.
This would be a wireless card problem?
Regards,
Dzieva
No a MT firmware problem.
regards
Ros
I don’t think NV2 disconnects is a simple as just MT firmware issue it maybe more complicated than that and perhaps a combination of several factors with 802.11 and nstreame there is no disconnects and perhaps it’s just more tolerant to what I assume is interference , with NV2 there is more performance advantages to be gained in using this protocol if it was a simple firmware upgrade don’t you think MT would have addressed this by now and avoided the bad press on this issue, however they tell users they cannot reproduce the disconnection issues in the test lab and ask for effected users to send back debug files to support,
I cannot understand why MT don’t send out a engineer with test equipment to several effected users sites and professionally analyse on site why NV2 is effected so much, from these tests MT would be in a better position to address this issue rather than the present solution,
I remember wireless parameters which we can set for a year before MT realized that they do not get through to the wireless card. I think it was retry in 3.x. This wouldn’t happen if they had an engineer with a spectrum analyzer. This not solved problems look like they
still not hired one.
Impressed with shielding waveguide around the SXT however i would reduce the size further if the interference is co-location source(s) like other AP’s or Ptp on a mast this to have reduced windloading in exposed windy locations, also cannot tell from picture if aluminium is flexible in the center if yes on a windy day this could cause signal variations but the concept is good maybe other users could offer more practical advice on this.
the shield is pure aluminum with V2A screws and it ist tested under windload.
You can turn the backplate 90° to have the SXT verticaly polarized on chain 0.
we are also shield Jirous Gentlebox 320 dual and got the RF reception of 5 clients from 13Mbit/20%ccq wireless rate to 39-58Mits /60-80% in a 20mhz channel on single chain clients.
only 1 client is dual 7 km the other are singlechain 8-12km away.
contact me directly if you want further information
sonny
Hi Rpingar,
Thanks for this, switch to a single chain solved the problem.
No more disconnections.
Thanks very much
Dzieva
are you so kind to give a karma?
regards
Ros
Why not make the shield round like a tube?
Could you cut down the beam width with the right shielding?
Unless you design the rf shield in a horn technique you could end up with multipath signals being picked up during RX, careful design and knowledge of wavelengths is essential or you could solve one problem but generate another but i do agree with the shape could be circular but prehaps exponential design but not tube.