I just updated to the newest build to give my two cents, I used airradar in mac that gives % instead of dbm.

All I see is that if i go 20dbm i get 65% 25dbm i get 74% and 27dbm (max) between 79 and 80%. (all in 2.4g)

Something to note is that if i go 17 dbm on 5ghz i get 52% and if i choose 30dbm (max) i get 60%, i dont understand why even with the higher dbm on 5ghz i get lower signal vs when compared with 2.4ghz.

It is clear however than changing the power has some increased coverage, but that 5ghz its still the weakest link.

(Last edited by abolition on 17 Apr 2011, 13:29)

You shouldn't be seeing differences between builds because there's been no changes to wireless drivers or anything else that should affect performance. You should probably look at your testing methods.

Well, please slap me in the face, because neither Windows' own signal strength bar thingy, nor the web GUI, nor inSSIDer report any change in signal strength. In fact, I've just set txpower to 1 (both via iwconfig and the GUI), and according to the GUI the signal "strength" is 65%...

Could someone please post his settings here, so i can copy them? I'm guessing that the problem lies there.

EDIT: Just managed to change the dBm on the 5GHz band,and that does seem to be working. However, I still can't get the 2.4GHz band to work with me.

(Last edited by Quipeace on 17 Apr 2011, 16:04)

Thanks, now it seems ok

From my testing something around 24-25dbm for the 5ghz band is the most stable.  Using InSSIDer I spent about 30min adjusting the settings and watching the graph.  Each dbm step increases signal, but when I set it past 26-27dbm, the average increases, but I start seeing larger ripples more like a sine wave.  These ripples can be seen in file transfers as the transfer rate fluctuates with some regularity.  At 24dbm, the fluctuations are significantly smaller.  I don't quite understand the science behind antenna/radio interactions, but it seems like while the radio can push that power level, the 5ghz antennas seem to distort or be otherwise unable to handle the higher power levels.

At 24dbm on 5ghz I'm currently averaging around -60dbm signal at the laptop.  Using 15dbm on 2.4ghz I'm also averaging about -60dbm at the laptop.  There is no competition in my area on the 5ghz band, but there's a handful of neighboring 2.4 networks.  I can push the 2.4g signal to 27 and end up with about -40dbm at the laptop, but all the neighboring networks are below -80dbm signal to me, so I figured I'd turn it down to where I end up with about a +20bm gradient and also reduce my network footprint some as I get full coverage @ 15dbm.

How are you changing the power levels?  Web Interface or SSH Command Line?

In my experience the Web interface is much more tedious to change the radio power levels, its less responsive and not as easy to tweak the settings and also quite often displays config options that are not valid/possible.  Its great for a lot of other things, but I'd recommend using the command line instead.

SSH into router
vi /etc/config/wireless
press I
scroll cursor to txpower entry
change as desired (within spec) making sure the number remains in single quotes '
press ESC
if you are happy with the changes you made:  type :x and then press enter   (include the : )
if you want to cancel without saving the changes:  type :quit! and then press enter  (include the : )

when back at the command prompt type
wifi
this will reset the wireless radios and use the new settings.  It will also show you error messages if you are on a wired system (sometimes it works on wireless, but since your restarting the radio, usually you have to reconnect the SSH session and you miss any error messages)

to find out what power levels it should let you set type
iw phy1 info |grep MHz
(case sensitive)  phy0 = 2.4ghz phy1 = 5ghz

You might have to change your channel to achieve a higher dbm setting.

config 'wifi-device' 'radio0'
        option 'type' 'mac80211'
        option 'macaddr' '**********'
        option 'hwmode' '11ng'
        list 'ht_capab' 'SHORT-GI-40'
        list 'ht_capab' 'TX-STBC'
        list 'ht_capab' 'RX-STBC1'
        list 'ht_capab' 'DSSS_CCK-40'
        option 'country' 'US'
        option 'htmode' 'HT40-'
        option 'channel' '7'
        option 'distance' '35'
        option 'noscan' '1'
        option 'disabled' '0'
        option 'txpower' '15'

config 'wifi-iface'
        option 'device' 'radio0'
        option 'network' 'lan'
        option 'mode' 'ap'
        option 'ssid' '**********'
        option 'encryption' 'psk2'
        option 'key' '***********'
        option 'hidden' '1'
        option 'macfilter' 'allow'
        option 'maclist' '*****************'

config 'wifi-device' 'radio1'
        option 'type' 'mac80211'
        option 'macaddr' '**********'
        option 'hwmode' '11na'
        list 'ht_capab' 'SHORT-GI-40'
        list 'ht_capab' 'TX-STBC'
        list 'ht_capab' 'RX-STBC1'
        list 'ht_capab' 'DSSS_CCK-40'
        option 'country' 'US'
        option 'htmode' 'HT40-'
        option 'distance' '35'
        option 'disabled' '0'
        option 'channel' '153'
        option 'txpower' '24'

config 'wifi-iface'
        option 'device' 'radio1'
        option 'network' 'lan'
        option 'mode' 'ap'
        option 'ssid' '*********'
        option 'hidden' '1'
        option 'encryption' 'psk2'
        option 'key' '*********'
        option 'macfilter' 'allow'
        option 'maclist' '***********'

2.4ghz steps 5/10/15/20/25/27dbm -- InSSIDer 2.0 -- Intel 5300

(Last edited by rjs_boy on 17 Apr 2011, 20:20)

does the option distance work? and if so how?

It appears to work.  Doesn't seem to make huge differences, but the last time I played with it, I started at 1000 and timed a 1gb file transfer from the farthest point I would be using the computer.  I then halved the value and retested.  Did that a few times to see the changes, ended up for me something around 30-40 seems to work best.  From what I can tell the value is distance in Meters from the AP, some documentation stated that I needed to double the value due to the fact  that the signal needs to make a full trip there and back.  But other documentation seemed to indicate that it was smart enough to automatically double.  The conflicting documentation is is why I started at 1000 and just tested it from there.  But about the farthest distance I go is around 30m from the router, and even set at 35 I've gone almost 50m away and still maintained a totally functional connection.

I think I gained all of 0.5mb/s - 1mb/s  transfer speed after all the fiddling.  Not huge, but its hard to resist tweaking to get the most out of a product.

quick retest using netio  (every test gives different value with netio...not fond of it, but its simple)

Distance '380'

TCP connection established.
Packet size  1k bytes:  6898.46 KByte/s Tx,  2722.17 KByte/s Rx.
Packet size  2k bytes:  7200.98 KByte/s Tx,  3305.24 KByte/s Rx.
Packet size  4k bytes:  7551.96 KByte/s Tx,  3869.44 KByte/s Rx.
Packet size  8k bytes:  7485.04 KByte/s Tx,  4351.88 KByte/s Rx.
Packet size 16k bytes:  7508.33 KByte/s Tx,  6115.91 KByte/s Rx.
Packet size 32k bytes:  7638.34 KByte/s Tx,  7609.10 KByte/s Rx.
                        7380 avg             4662 avg

Distance '140'

Packet size  1k bytes:  7457.29 KByte/s Tx,  3127.19 KByte/s Rx.
Packet size  2k bytes:  7523.64 KByte/s Tx,  3938.44 KByte/s Rx.
Packet size  4k bytes:  7755.00 KByte/s Tx,  3598.47 KByte/s Rx.
Packet size  8k bytes:  7961.33 KByte/s Tx,  6331.76 KByte/s Rx.
Packet size 16k bytes:  7736.30 KByte/s Tx,  7373.80 KByte/s Rx.
Packet size 32k bytes:  7787.96 KByte/s Tx,  7433.50 KByte/s Rx.
                        7703 avg             5300 avg

Distance '70'

Packet size  1k bytes:  7125.49 KByte/s Tx,  2898.95 KByte/s Rx.
Packet size  2k bytes:  7415.68 KByte/s Tx,  3703.28 KByte/s Rx.
Packet size  4k bytes:  7700.08 KByte/s Tx,  3145.28 KByte/s Rx.
Packet size  8k bytes:  7856.05 KByte/s Tx,  5890.49 KByte/s Rx.
Packet size 16k bytes:  7974.77 KByte/s Tx,  6214.25 KByte/s Rx.
Packet size 32k bytes:  7875.81 KByte/s Tx,  8052.05 KByte/s Rx.
                        7657 avg             4983 avg

Distance '35'

Packet size  1k bytes:  6614.77 KByte/s Tx,  3940.53 KByte/s Rx.
Packet size  2k bytes:  7253.10 KByte/s Tx,  3247.60 KByte/s Rx.
Packet size  4k bytes:  7822.00 KByte/s Tx,  4199.13 KByte/s Rx.
Packet size  8k bytes:  7739.40 KByte/s Tx,  5325.47 KByte/s Rx.
Packet size 16k bytes:  7596.10 KByte/s Tx,  7075.39 KByte/s Rx.
Packet size 32k bytes:  8104.07 KByte/s Tx,  8128.09 KByte/s Rx.
                        7521 avg             5319 avg

Doing the large file transfer and timing it was much more consistent for testing.

(Last edited by rjs_boy on 17 Apr 2011, 20:43)

First of all, thank you for the clear graphs and explanation, I was used to setting the txpower via iwconfig, but ofcourse, that isn't permanent, so I used the GUI for that .

Anyways, I tried using your method of vi-ing and resetting, but I just cannot seem to get it to work, I also partially copied your settings for radio0. Here's a printscreen of inSSIDer, can you spot when I changed from 15dBm to 27dBm? I'm not getting any signal on my smarphone upstairs either. Again, I used your settings and config-method.

15dBm/27dBm/15dBm/17dBm  -  Intel 5100

Could there be something wrong with the antenna?

(Last edited by Quipeace on 17 Apr 2011, 21:55)

Try a different client? Also, what are you expecting to see? I'm no radio expert, but from reading about transmit power, apparently you can get worse signal by increasing it. Other factors in your environment can affect things as well. I'm just not sure you can execute a command and see a big jump in your graph there. Also with just one client you don't have enough data to draw a conclusion.

I like the GUI for many things, but setting the TXpower...I'm just never sure if the setting actually took place or not.  So I learned how to edit the config file and have done it that way ever since.  Regardless, there should have been some noticeable difference between 15 and 27dbm.

I decided to fiddle with the iwconfig method of changing the txpower setting, and man, I wish I had known about that sooner than now.

editing the config file resets the radios, using iwconfig doesn't even drop the connections (unless set too low...oops).

One thing I noticed with InSSIDer is that its very poor at tracking other networks than the one the computer is currently connected to.  However if you disconnect from all wireless networks, it paints a much better picture of what is happening, OR if you only modify the one you are connected to.

I spent some time playing around with both the 2.4ghz and the 5ghz radios and iwconfig.  Again, I am able to notice differences ranging from 5-30dbm (27 on 2.4g).

iwconfig wlan0 txpower 15  -  2.4g to 15dbm
iwconfig wlan0 txpower 27  -  2.4g to 27dbm

iwconfig wlan1 txpower 15 - 5ghz to 15dbm
iwconfig wlan1 txpower 30 - 5ghz to 30dbm

changing each network while connected to it and running InSSIDer to monitor the signal.  I can regularly tell the difference between every 2dbm step, single dbm settings were not as clear.  However it only worked if I was connected to the 5ghz network while changing the 5ghz settings; same for the 2.4g network.  If I was monitoring the 2.4g radio while connected to the 5ghz network, it was much less clear what effect any of the changes were having to the 2.4g network.

There definitely should be more of a difference between 15 and 27dbm.
I doubt its an antenna problem, simply because you are getting a low to mid 50dbm signal.  If anything I'd suggest getting farther away from the router and retesting the difference between 15 and 27.
Do you have a V1 or V2 router, (mine is the V1 but the only way I know that is because I had it prior to them releasing a V2).  Shouldn't matter for the 2.4g radios, only the 5g antennas were changed in the V2.
Can you post your wireless config for the 2.4ghz radio?

First of all, my configuration:

config 'wifi-device' 'radio0'
        option 'type' 'mac80211'
        option 'macaddr' 'xxxxxxxxxxxxxx'
        option 'hwmode' '11ng'
        list 'ht_capab' 'SHORT-GI-40'
        list 'ht_capab' 'TX-STBC'
        list 'ht_capab' 'RX-STBC1'
        list 'ht_capab' 'DSSS_CCK-40'
        option 'country' 'US'
        option 'htmode' 'HT40-'
        option 'channel' '7'
        option 'distance' '35'
        option 'noscan' '1'
        option 'disabled' '0'
        option 'txpower' '17'

config 'wifi-iface'
        option 'device' 'radio0'
        option 'network' 'lan'
        option 'mode' 'ap'
        option 'encryption' 'psk2'
        option 'wps_pushbutton' '1'
        option 'ssid' 'xxxxxxxxxxxxx'

The 5GHz network seems to respond to the increase in power, I can clearly see changes. However, the 2.4GHz channel is still giving problems: (RED LINE)

I started at 27dBm, then dropped to 2dBm, then to 17dBm, and back to 27dBm. The signal downstairs is fine, but I'm still not getting anything on the smartphone.

I'm using your build for a week now, and it's working like a charm, thanks.

I searched the forum and read the wireless docs, and it seems there is no support for a wireless mac filter.
Does anyone found a solution for this?

There is support now I believe, the wiki is outdated. Try it

Sigh, new problem: miniupnpd seems to restart every 30 minutes, anybody else seeing this?

root@OpenWrt /root# logread | grep upnp | grep bye
Apr 18 06:37:31 OpenWrt daemon.notice miniupnpd[4773]: received signal 15, good-bye
Apr 18 07:07:32 OpenWrt daemon.notice miniupnpd[4996]: received signal 15, good-bye
Apr 18 07:37:33 OpenWrt daemon.notice miniupnpd[5219]: received signal 15, good-bye
Apr 18 08:07:34 OpenWrt daemon.notice miniupnpd[5442]: received signal 15, good-bye
Apr 18 08:37:35 OpenWrt daemon.notice miniupnpd[5665]: received signal 15, good-bye
Apr 18 09:07:36 OpenWrt daemon.notice miniupnpd[5888]: received signal 15, good-bye

It's receiving a SIGTERM, which is a normal kill signal. It's not crashed, it's just terminated. Can't for the life of me figure out what the hell is causing this. I attached some logging to the init script and hotplug scripts to see if they were the cause, but no go:

Apr 18 09:07:36 OpenWrt user.notice miniupnp: init script called
Apr 18 09:07:36 OpenWrt user.notice miniupnp: init script called
Apr 18 09:07:36 OpenWrt daemon.notice miniupnpd[5888]: received signal 15, good-bye
Apr 18 09:07:36 OpenWrt user.notice miniupnpd: fw hotplug script called, args: 
Apr 18 09:07:36 OpenWrt user.notice miniupnp: init script called
Apr 18 09:07:36 OpenWrt user.notice miniupnpd: removing firewall rules for eth1 from zone wan
Apr 18 09:07:36 OpenWrt user.notice miniupnpd: adding firewall rules for eth1 to zone wan
Apr 18 09:07:37 OpenWrt user.info syslog: SNet version started
Apr 18 09:07:37 OpenWrt daemon.notice miniupnpd[6111]: HTTP listening on port 5000
Apr 18 09:07:37 OpenWrt daemon.notice miniupnpd[6111]: Listening for NAT-PMP traffic on port 5351

You can see the init script being called, then the firewall hotplug script for miniupnpd being called. I'm thinking maybe uci state has something to do with it, but I don't understand the 30 minute interval. My eth1 dhcp lease time is 60 minutes, so why is the firewall hotplug script called?

The udhcpc client tries to refresh the lease after half of the time expired.

Did you maybe remove the default "accept dhcp" rule and now run into https://dev.openwrt.org/ticket/4108 ?

I am not experiencing this issue on r26698.

Ok, I identified a syntax error in my miniupnpd iface hotplug handler that caused the restarting of miniupnpd on any interface changes. I've also implemented checks in the miniupnpd/minidlna init script to see if minissdpd is running before starting.

@jow

I thought of that too, but the rule is there. Maybe something similar is happening though, I'll investigate some more.

EDIT: Looks like udhcpc triggers a hotplug ACTION=UPDATE for eth1, and my syntax error in the hotplug handler caused it to restart miniupnpd. So my bad

(Last edited by arokh on 18 Apr 2011, 15:58)

@arokh Will this http://github.com/albertz/shairport work on your build? Maybe you can add this in build?

Yes, it will work It won't fit on ROM, but I'll provide packages on my repo as soon as someone packages everything up properly. See https://forum.openwrt.org/viewtopic.php?id=29491

I don't have a USB soundcard but I do have an Airport Express, so I'm not going to do the work myself

Tried it before, and it gave me an error when I restarted wifi, so I figured it was like the wiki said, not working yet.

Today I tried it again, same error, I put it in the "config 'wifi-device'" section....wrong!
So I put it in the "config 'wifi-iface'" section, and it's working now.

Thanks arokh

Edit: I modified "macfilter/maclist" in the "Wireless configuration" section  in the Wiki,

(Last edited by joszz on 24 Apr 2011, 10:33)

Awesome!

AFAIK you need a jack to use airtunes (or hdmi like on apple tv) how will this work on our router?

With an usb audio card

I already have an apple tv, not really useful for me XD apple tv and router are separated by 30 cm lol