According to the FAQ, the power displays 255 by default, even though it's capped by the driver "unless you enable override".

Now, I've installed the wl package. When I enter wl txpwr1 it responds by telling me that it's transmitting at 251mW, which as the FAQ states is incorrect. What I'm most confused about is what is it capped to? nvram get pa0maxpwr on my unit lists 0x4e, or 78 in decimal. Does that mean my wrt is transmitting at 78mW? If I change the power to 78 using wl txpwr1 -m 78 will I still be transmitting at 78mW, or will I be transmitting at some frational value of 78 (ie, does 255 corespond directly to 255mW, or does it follow something like ActualTXpwr=SetTXpwr*78/255?).

One of the reasons I ask is because all other firmwares I've tried have listed 28mW as the default value, but it looks to me that OpenWRT is using 78mW as the default value. I'd really like to know what this is actually transmitting at.

iwconfig should show your current xmit power. RC4 default is 19dbm which is ~79mw.

- DL

The 255 means "try to set the driver to 255mW" .. the driver rejects this and sets the power at the maximum, pa0maxpwr.

It's also worth pointing out (again, from the FAQ) that increasing the transmit power does not automatically improve your range, the AP may transmit further but the client will not send any stronger signal back to the AP. A much better solution is to change the antennas on the ap (and/or the client) to directional antennas which will give you a gain, amplifying both the signal transmitted and the signal received.

Thanks. Other firmwares seem to default to 28mW (~14dBm?). Anyone happen to know what the Linksys default firmware is? Just wondering why OpenWrt defaults to the driver maximum limit when the other firmwares (Sveasoft, DD-WRT, etc) don't seem to. Is OpenWrt emulating the default firmware, or are the other firmware's emulating the default firmware?

OpenWrt makes no decisions as to what the power is set to.  The values setting the transmit power are in NVRAM before you put OpenWrt on and the binary-only non-OpenWrt-originated driver reads these values and acts on them. 

The only difference is that OpenWrt does not hide it from you like the other firmwares try to do (resulting in your confusion).

I can't pull up current linksys specs pdf but iirc they have variously spec'd it at 15dBm or 18dBm depending on version. I'll check the setting on the next (and last, before I switch to ASUS) of the wrt54's I've got in stock.

And note that (as already pointed out) 19dBm is by no means the limit of Openwrt. But also as already noted, IMO there's little point in increasing for most applications vs using antenna gain.

- DL

(Last edited by dl on 11 Feb 2006, 10:16)

No need to check if it's a hassle, I was really just curious with that point. So now a final quesion.

Let's say I don't want to transmit such high power (because I have a high gain antenna and don't want too much brain damage) or I want to limit the ability of my neighbors to even try and connect, avoid conflicts with phones, etc. I don't see a setting in the nvram for setting the output power other than pa0maxpwr. Is this what I should use, or is there a config file elsewhere? Or should I just write an init script that issues the "wl txpwr1 ..." command?

I use a startup script, but for convenience set my own nvram variables:

wl txpwr $(nvram get wl0_txpwr)
wl antdiv $(nvram get wl0_antdiv)
wl txant $(nvram get wl0_txant)

- DL

So are you all saying the 250mw claims being made by Sveasoft are bogus?

Can I override the iwconfig driver to force more than 84mw of power out of the radio?  If so, how?

Thanks

Read the thread again, they are saying that largely speaking it's a pointless exercise.

the later linksys wrt54g models, like v3.1, use SiGe's Power Amplifier 2528L; it's the little 16 pin IC next to the BCM2050 radio under the sheet metal cover.
Application note: http://www.sige.com/support/files/appli … on_0.1.pdf

Spec of the PA is 24 dBm in 802.11b and 21 dBm 802.11g.  Hence some more than 19 dBm should be possible, but perhaps only with hardware (PA bias) changes. Would the supply voltage of the third amplifier stage be 5 V then an extra 2 dB are in the spec for g mode. Beyond that the Error Vector Magnitude deteriorates (phase errors which degrade g mode performance). Can confirm experimentally that unmodified any setting beyond appr. 75 mW may show in iwconfig or wl txpwr1 but in actual fact does not happen.
Note that these power values are momentary power; as the link needs to be receiving as well and is not duplex the long-term average power is considerably less.

A possibility that I cannot completely rule out is that the system is actually doing the full 21 dBm but is reporting less. After all the values reported from the software are anything but calibrated measurements. Unfortunately I do not have the equipment to measure power reliably at 2.4 GHz.

(Last edited by doddel on 10 Mar 2006, 23:34)

Increasing power is not a pointless exercise in one very good example.

With WDS increasing transmit power on all the routers used allows a much greater distance between routers without buying expensive antennas.   

In the past I used (I hate to admit) Sveasoft and WDS to build a wireless network for a small residential neighborhood.  Each router's clients only needed the range of the house, but needed to be able to relay to the next router in the WDS network in some cases a couple blocks.  Boosting transmit power made this possible with only the 7dbi antennae from Linksys. 

I am in development for building a more extensive WDS network for a business park.  Again increase transmit power will allow the links to get from building to building without expensive high gain or external antennae.  So long as the individual clients are in range of one router they will work fine. 

The only downside to this is that clients always seen high signal strength but cannot connect if their transmission is out of range of the router.  You must educate the users on this anomaly.

> Increasing power is not a pointless exercise in one very good example.

more important is "radiated power", than output power. If you get more output power from WiFi radio, mean that you produce more heat and noise, nothing more. If you look with spectrum analyzer, then above 80 mW, signal from almost every WiFi transmitter goes wider (producing additional noise for sidechannels), but actual power reading remains almost same. And even, if you somehow can rise power from 32mW (15dBm) to 250mW (24dBm), then radiated power difference with same antenna will be only 9dB. Just use better antenna and get extra 9dB and maybe more, with less noise and heat. Up to 100mW output power is usable, when you are in the wild, already using maximal possible antenna gain, and need just a little more signal strength for a very long link. 250mW is pointless and is usable only for fooling people.

(Last edited by netmaster on 12 Mar 2006, 16:49)

My point was, the vast majority of people who ask about this are n00bs.

Do they apply high-gain antenna first? No. Do they optimize other aspects of their operation first? No. They go straight for increasing transmit power with the rubber ducky units so they can fool themselves that they are getting better reception on their laptop with it's 30 mW card and stubby antenna. Because they see more bars in the signal meter on the client laptop, they are convinced that they have actually done something COOL.

If you are doing point-to-point with high-gain antenna, you have a valid reason for it. But I see no reason for making it easy for people, given this option most will just increase the local interference problems by a factor of 9 for no useful purpose.

Excellent point.    As to how to make the distribution both novice and expert friendly, I don't have any good ideas.  Personally I'd prefer the distribution focus mainly on seasoned linux users, and let distros like DD-WRT build to simple GUIs for the novices. 

The biggest problem I had when switching to OpenWRT was my long distance links that I replaced Sveasoft with.  I had done all my research on antenna and had a solid working installation, except when Sveasoft would crash or fail to boot.  Because of lack of documentation and avoidence of the subject on these forums, it took me a while.

Needless to say, I figured it out and never looked back.

You'll be interested in dctrl if you are doing long-distance.  That originated here, from nbd,  and lets you adjust ACK timing. I think DD-WRT has it now also.

hello. i have the following problem with this. i'm using high gain antenna (directional) on my WL500gP. i have just upgraded from whiterussian who let me change the tx power. i need this setting for my wireless connection to even work. on my wr 0.9 with maxed out txpower i had good stable connection. now... nothing! the connection is between my router and one mac computer which are about 300 m away with two solid concrete walls with a lot of metal in them between.. i really need this to happen. thanks.

edit: yes i have tested the wireless connection closer to the router, it works. that to clear up any confusion about me not turning wifi on first.

(Last edited by dlbogdan on 21 Mar 2008, 23:18)

nevermind... i reverted to whiterussian.... much much better