Page 1 of 1

AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 2:12 pm
by czolo
World record? :D
I hope you like it

MT Team:
Can you make better power control for AR93xx cards? Now it works only in all-rates-fixed mode, and it's a problem for link stability
Are you planning to support QCA9880 or other 802.11ac chipsets from Qualcomm?

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 4:08 pm
by server8
Which kind of HW?

Giuseppe

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 4:10 pm
by czolo
Which kind of HW?

Giuseppe
Check specification at the bootom of this site:
http://www.cyberbajt.pl/cms.php?get=produkt,4334

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 4:58 pm
by uldis
for those chipset cards only all-rates-fixed is working.
On what distance you get such results?

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 5:17 pm
by honzam
Why using MiniPCI express? Is there any harware limitation with MiniPCI?

http://forum.mikrotik.com/viewtopic.php?f=3&t=57864

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 6:11 pm
by czolo
for those chipset cards only all-rates-fixed is working.
Is this software or hardware limitation?
On what distance you get such results?
This is laboratory test. In outdoor test results are 10-20% worse.

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 6:12 pm
by czolo
Why using MiniPCI express? Is there any harware limitation with MiniPCI?

http://forum.mikrotik.com/viewtopic.php?f=3&t=57864
Have you found any 3x3 card for MiniPCI? AR9380/AR9390 works only with MiniPCI-Ex

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 8:12 pm
by honzam

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 8:29 pm
by czolo

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 9:24 pm
by honzam
All the those cards have only 2 chains, so MCS-15 only and 300Mbps.
Look you good? All these cards are 3x3 mimo

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 9:54 pm
by czolo
All the those cards have only 2 chains, so MCS-15 only and 300Mbps.
Look you good? All these cards are 3x3 mimo
They aren't full 3x3 MIMO. They are 2T3R, AR9160 is 2-chain chipset.

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 11:03 pm
by InoX
All the those cards have only 2 chains, so MCS-15 only and 300Mbps.
Look you good? All these cards are 3x3 mimo
Data Rates: Legacy 802.11a/b/g (1-54Mbps), 802.11n (up to 300Mbps)

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 11:30 pm
by 0ldman
2T3R cards, can they actually receive @ 450Mbps?

They actually TX/RX over all three chains but only two streams, correct?

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 11:38 pm
by honzam
OK, my mistake :) Does anyone 3x3 sector antenas deployed in real production?

Re: AR9380 - laboratory tests

Posted: Mon Mar 19, 2012 11:50 pm
by czolo
OK, my mistake :) Does anyone 3x3 sector antenas deployed in real production?
RF-elements shown 3x3 sector at MUM. Gigaeter is a panel antenna, not a sector.

Re: AR9380 - laboratory tests

Posted: Tue Mar 20, 2012 10:37 pm
by czolo
for those chipset cards only all-rates-fixed is working.
On what distance you get such results?
What with my question about new chipsets support?

Re: AR9380 - laboratory tests

Posted: Wed Apr 04, 2012 10:06 am
by demopro
outdoor test
distance 35km
antennas on both ends rocket dish 5g34 (V/H polar.) + 28dbi panel antenna in slant polar.
card - http://www.compex.com.sg/fullDescription.aspx?pID=132 with ar9390 chipset
environment full of noise
tested with nv2, nstreme , 20MHz channel , 40MHz channel, fixed data rates, default data rates, all tx-power mode

best result 90Mbps half duplex
any help? uldis, normis?

Re: AR9380 - laboratory tests

Posted: Wed Apr 04, 2012 7:38 pm
by czolo
outdoor test
distance 35km
antennas on both ends rocket dish 5g34 (V/H polar.) + 28dbi panel antenna in slant polar.
card - http://www.compex.com.sg/fullDescription.aspx?pID=132 with ar9390 chipset
environment full of noise
tested with nv2, nstreme , 20MHz channel , 40MHz channel, fixed data rates, default data rates, all tx-power mode

best result 90Mbps half duplex
any help? uldis, normis?
Can you show us:
/int wir pr adv

Important question: how do you connect the pigtails to the cards? Please describe the order in which have you connected pigtails?
I see from screens, that you have problems with signal level. Signals should me similiar.

Re: AR9380 - laboratory tests

Posted: Thu Apr 05, 2012 12:33 am
by demopro
wireless card have basic setup, most is default and experimented with every combinaton of frequency, nv2 , nstreme , data-rates, tx-power.
ch0 is vertical polarization of RocketDish 34dbi
ch1 is horisontal polarization of Rocket Dish 34dBi
ch2 is +45° slant polarization of panel antenna 28dBi

reason for the asymmetric signal is frequency 5000 MHz , for freq. 5500 signal levels rx/tx are almost identical
in one situation i had 135Mbps but not stable

Re: AR9380 - laboratory tests

Posted: Thu Apr 05, 2012 8:18 am
by czolo
wireless card have basic setup, most is default and experimented with every combinaton of frequency, nv2 , nstreme , data-rates, tx-power.
ch0 is vertical polarization of RocketDish 34dbi
ch1 is horisontal polarization of Rocket Dish 34dBi
ch2 is +45° slant polarization of panel antenna 28dBi

reason for the asymmetric signal is frequency 5000 MHz , for freq. 5500 signal levels rx/tx are almost identical
in one situation i had 135Mbps but not stable
Connections are correct. If in 5500 freq signals are identical, so I think you had to install slant antennas in good positions.
What you may try:
- set data rates only for MCS16-20
- for testing use simple 802.11 and check BER on the link (you have poor CCQ)
- best result I have with nstreme with dynamic size, but you can try nv2 too
- try to find some channel near 5500Mhz - look at CCQ & Signal strength parameter when you decide which freq is better

Maybe I could I look at your link directly (via net of course). My priv conntact: czolo[at]o2[no_spam].pl