OpenWrt Forum Archive

Topic: TP-Link tl-wr741nd using wrong antenna?

The content of this topic has been archived on 13 Apr 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

Can somebody verify this on another tl-wr741nd?

I just got a tl-wr741nd and flashed backfire 10.03. Everything works fine but the external antenna is not being used. I'm using mac80211/ath9k

This router has only one external antenna but internally it has two, one of them as a PCB trace antenna. I have tried putting a reflector on the external antenna, using a 14db panel antenna and I see no change in reception (always the same amount of aps when scanning, same power, etc). I also expected it to be much better with the stock antenna.

So, I've come to the conclusion it's NOT using the external antenna but the internal pcb one. Maybe diversity is disabled but the antenna is not set correctly or something. txantenna, rxantenna and diversity seem to be all ignored in /etc/config/wireless.

You can see the antenna connections on this pic from the wiki:  http://wiki.openwrt.org/_media/toh/tp-l … l-pins.jpg

i opened a ticket too: https://dev.openwrt.org/ticket/7576

some info about the ar9285 wireless nic from Luis Rodriguez (from atheros):

> AR5416 and AR9280 are 2 stream.
>
> AR9285 is 1 stream.
>
> All AR9280s are 2 stream and have a 2x2 RX/TX chain configuration.
>
> The AR5416 has an 3x2 RX/TX chain configuration but it must be noted
> this then means one stream can use two chains with selection diversity
> -- sort of like antenna diversity but for one chain.
>
> AR9285 devices are 1-stream and use antenna diversity with some
> enhancements
, I think one of the enhancements is LNA, which I believe
> means we get to dynamically configure the power amplifier depending on
> the noise on the driver and I think this helps with noisy
> environments.
>
> AR9170 is 2 stream 2x2 chain config.
>
> AR9271 (USB) is 1 stream with antenna diversity support (and LNA IIRC).
>
>  Luis
>

more:

802.11n does allow for 1 stream devices with a 1x1 TX/RX chain setup
but the hardware *can* have two antennas like your typical legacy
device with diversity enabled. This is the case for AR9285, for
example.

Luis

(Last edited by naplam on 7 Jul 2010, 13:47)

update: as a station i still have the problem but i think as AP maybe it's working ok. I made some tests and it looks like turning the directional antenna does change the power received in other stations. It remains to be seen if this is just the tx antenna being properly selected... maybe the rxantenna is still not working properly.

naplam wrote:

update: as a station i still have the problem but i think as AP maybe it's working ok. I made some tests and it looks like turning the directional antenna does change the power received in other stations. It remains to be seen if this is just the tx antenna being properly selected... maybe the rxantenna is still not working properly.

I made another test today by checking the power of an associated sta as seen by the router in ap mode, it seems it's not a matter of ap or sta, it's a matter of reception or transmission: When receiving, i get the same power no matter what i do with the ap, thus it looks like it's transmitting on the external antenna but receiving on the tiny internal one only.  When transmitting (router as sta), i can see power vary by moving the directional antenna, wrpping the stock antenna in tinfoil etc., but when i scan as sta i get very few aps presumably due to the wrong antenna being chosen to receive.

Hi Guys,

I have been fiddling with the TL-WR741ND  and noticed this problem too.  I decided to do some testing by disabling the internal pcb trace antenna (by removing a 0 ohm link) and soldered on another rpsma socket.
This way I could attach two antenna's , one to the original rpsma socket and another to the newly added socket.

I then attached an antenna to the new socket , did a scan for nearby AP's.  All the Ap's I know to be around this area were detected.  Apon removing the antenna, i then rescanned. This time it found far fewer AP's.

The oposite was then tried with an antenna attached ONLY to the original socket. A laptop running netstumbler was used to monitor the output from the AP, the antenna then removed resulting in a sharp drop in the netsumbler signal plots.

I did both of these tests several times. It seems to confirm that the internal trace antenna is used for RX and the antenna attached externally is for TX.

Sorry for the long winded waffle smile I am not the best at explanations.


Hoping this info is of use,

Brad

Hello again,

I forgot to mention that the firmware used was TP-link's and NOT Openwrt (I have yet to be able to get the wireless working at all when openwrt is installed, sorry thats probably best in another thread smile )

This seems to indicate that the antenna problem is not with Openwrt code but maybe underlying source code from the manufacturer of the Wifi IC? 

... Brad

Brad, I think this must have been fixed in OpenWrt already, nbd gave me some patches months ago and we tested it thoroughly, so the patches must have been committed already

naplam wrote:

Brad, I think this must have been fixed in OpenWrt already, nbd gave me some patches months ago and we tested it thoroughly, so the patches must have been committed already

naplam, did your successfully configure rxantenna?
Can your provide appropriated wireless config?

smile

Hi All,
Did anybody get antenna diversity working properly on wr741ng, version 1.8?
I tried with the latest trunk (kernel kernel - 2.6.37.4-1), but still getting the same rx signal, with or without antenna.
Here is what I have in /etc/wireless/config

config wifi-device  radio0
        option type     mac80211
        option channel  11
        option macaddr  00:23:cd:1d:11:f4
        option hwmode   11ng
        option htmode   HT20
        list ht_capab   SHORT-GI-40
        list ht_capab   TX-STBC
        list ht_capab   RX-STBC1
        list ht_capab   DSSS_CCK-40
        # REMOVE THIS LINE TO ENABLE WIFI:


Is there any way to apply diversity patch (http://lkml.org/lkml/2010/8/6/358) without compiling - i.e. using firmaware generator?
Is there a firmware with this patch included?
Thanks.

ATTITUDE ADJUSTMENT (bleeding edge, r26290) uses Linux Kernel 2.6.37. Yeah, its bleeding edge and not stable, but well, it does work quite nicely.

And no, as long as there are not binary patches, you cannot use the firmware generator to accomplish this. You won't need to compile the toolchain, but crosscompile the source after applieying patch with a precompiled toolchain (aka SDK). Or try bleeding edge ;-)

Thanks for the response Orca.
Please excuse my ignorance, but with 'ATTITUDE ADJUSTMENT (bleeding edge, r26290)' are you referring to the latest trunk for wr741nd -
http://downloads.openwrt.org/snapshots/ … pgrade.bin?

I was using that one, but still getting no difference in RX signal, with or without antenna attached to the router.

Where did I see this problem before?
Not directly related but smell like a general problem.

Is there anyway to see if antenna diversity patch (http://lkml.org/lkml/2010/8/6/358) is included in the firmware I've installed?
If the patch is included in the installed firmware, should I have 'option diversity' in /etc/wireless/config?
Thanks.

BradA wrote:

I have been fiddling with the TL-WR741ND  and noticed this problem too.  I decided to do some testing by disabling the internal pcb trace antenna (by removing a 0 ohm link) and soldered on another rpsma socket.
This way I could attach two antenna's , one to the original rpsma socket and another to the newly added socket.
Brad

Can anybody mark on the image (http://wiki.openwrt.org/_media/toh/tp-l … l-pins.jpg) where to add external antenna instead of pcb trace antenna?
Thanks.

The discussion might have continued from here.