Alfa comparison test: AWUS036 H vs NH

[VulpiArgenti]

I recently bought the newer NH version of the AWUS036, seduced by 2000w transmit power plus N-speed data transfer. I've made a pseudo-scientific comparison; same 5 dBi antenna, HD install BT5R1 32 then BT5R2 64, controlling as many variables as possible.

Internet

Number of APs seen in wicd: about the same

Speed connects to router (WPA2): H much quicker

Ping router:

H     100 packets transmitted, 99 received, 1% packet loss, time 99162ms
       rtt min/avg/max/mdev = 1.240/4.726/177.438/19.295 ms

NH   100 packets transmitted, 100 received, 0% packet loss, time 99232ms
       rtt min/avg/max/mdev = 1.359/677.013/2927.134/612.316 ms, pipe 3

Download speed test:
H 286 KB/s
NH 577


Monitor mode/airodump

Airodump scan: H sees 30% more APs

Reception PWR: H much better, by an average of 10 dBm
(Note that the value for PWR is reported by the driver, therefore may not be comparable).

RXQ on fixed channel/AP: very variable. H reports itself as more sensitive (eg 24 vs 12) , but both seem to collect a similar number of packets.

Injection: both excellent.

Transmit power: Hacking the regdb/CRDA allows:

wlan1 (H)      IEEE 802.11bg  ESSID:off/any  
                   Mode:Managed  Access Point: Not-Associated   Tx-Power=30 dBm   
wlan3 (NH)    IEEE 802.11bgn  ESSID:off/any  
                   Mode:Managed  Access Point: Not-Associated   Tx-Power=35 dBm !!

I tried to test this by DOSing my AP from a distance (PWR -85). NH reported fewer ACKs (eg H [58|62 ACKs] vs NH [18|28 ACKs]), but this is probably a function of its lower sensitivity. There seemed to be no difference in the effectiveness of the deauth.
I suspect there is no advantage in increasing transmit power at the expense of reception sensitivity.


Soft AP

Setup: no problems on either, airbase and dhcpd work.

Soft AP transfer speed (for the victim):

Driver	Max kbps	Ave KB/s	Victim browsing experience
H	350	45	slightly better than dialup
NH	70	3	almost unusable
(iwlagn	700	64	slow broadband)

Other

Macchanger: no problems with either.
5GHz band: not seen by either (need AWUS051NH for this)
Virtualbox: H stable, NH flakey


Conclusion

If I could have only one, it would definitely be the AWUS036H.



P.S. If anyone has any tests they want me to run I'll do my best to oblige.
P.P.S. Anyone want to buy my NH?

[zimmaro]

hi,
VulpiArgenti thanks for the excellent/interesting analysis of the "comparison". (if I may say so) I had already had some doubts during this similar-thread:
http://www.backtrack-linux.org/forum...hlight=txpower
good job!thanks-again
+1 36H

[Scamentology]

35 dbm is higher than 2 watts. This may explain why your card is malfunctioning (you will most likely destroy your card running it that high without modding it). No higher than 32dbms would be my recommendation and 33 dbms is just under 2 watts.

http://www.moonblinkwifi.com/dbm_to_watt_conversion.cfm

Try your test again (if you ran the test at 35dbm - you didn't specify the power used in the test) and if you haven't melted it yet you will see an improvement.

Thanks for sharing your findings.

[VulpiArgenti]

Hi Scamentology

you didn't specify the power used in the test

I followed your tutorial, and mostly use 32dBm. Surprisingly, I haven't seen any difference between 30-35 (benefit or harm). This is the reported power; who knows what it is actually putting out?

Why do you say my card is malfunctioning? It seems to work well, just not as sensitive as the H.

(I am well aware of regulatory issues; this is in a lab setting)

[Osirium]

VulpiArgenti ...agree almost in full with what did you say.

However the best performance that I ever had it with wireless cards was based on this one:

NextG USB-Yagi Plug & Play 11N Long Range WiFi antenna 2200mW..this is the absolute thing that should be part of a Wifi Warrior kit

[VulpiArgenti]

Intriguing - I'll look into it. Thanks.

[Scamentology]

Hi Scamentology

Why do you say my card is malfunctioning? It seems to work well, just not as sensitive as the H.

NH
txpower = 30
ping
8 packets transmitted, 8 received, 0% packet loss, time 7000ms
rtt min/avg/max/mdev = 0.469/0.874/2.660/0.690 ms

Speedtest.net
3 meg connection
2.60 Down
.73 up

_____________
txpower = 32
8 packets transmitted, 8 received, 0% packet loss, time 6997ms
rtt min/avg/max/mdev = 0.493/0.522/0.558/0.037 ms

Speedtest.net
3 meg connection
2.83 Down
.73 up

Both tests were done about 6 feet from the router.

You can see how ( by comparing numbers ) I would think its malfunctioning.

(I am well aware of regulatory issues; this is in a lab setting)

I stated no objections to this. In fact I encourage this and appreciate the effort.

The fact that we get such different outcomes with this card means this card is NOT stable enough to recommend, but the one I got works great.

VulpiArgenti
NextG USB-Yagi Plug & Play 11N Long Range WiFi antenna 2200mW..this is the absolute thing that should be part of a Wifi Warrior kit

I will be purchasing one of these tonight, Will let you know next week. Thanks for the heads up.

[grikster]

Great report, looking forward for your next analysis.

Thanks

[bitbomb]

Download speed test:
H 286 KB/s
NH 577

Assuming this is 577 KB/s. I guess download speed is the only area that the NH offers any advantage

[VulpiArgenti]

Thanks for posting your tests Scamentology.

My main disappointment with the NH was 1) sensitivity, and 2) poor performance as a soft AP. It would be interesting to hear anyone's results on these.

[Off-topic: looking forward to seeing a post on that wicked NextG]