OpenWrt Forum Archive

Topic: WRT54G electric consumption

The content of this topic has been archived on 1 May 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.

Hi all

I have a power supply delivering different voltages, and in facts, the WRT seems to be able to run at only 4.5V and still 1 A current intensity. I am wondering if anyone has the same phenomenon with its machine. If so, why does linksys provide a 12V/1A power unit ?

If it works fine, and I believe it does, it would be great to make solar and/or wind autonomous units.

Thanks for your answers.

There are lots of example - including some on I'm sure - about how the WRT54G (and similar boards) will take anything from 5v to 40v DC power, and work fine.

If you have a problem at any voltages other than 4.5V.....then I can't say what's causing that!

There is absolutely no reason you could not use a solar panel or wind generator to power one of these units, provided the voltage never dropped too low.


Don't have the total schematic but there is documentation of the 1501-33 power regulator that is in my wrt54G V2.
See: … AP1501.pdf
With very few external components this switch-mode buck converting power supply IC can downconvert voltages and deliver up to 3 Ampere at 3.3 V. It integrates the regulating circuitry and the power transistor
Question is whether the raw input voltage is still also used somewhere in the WRT54G; don't know and don't fancy probing all those little solder pads, risking short circuit. But if it is not, it explains why the input voltage is not critical as long as the regulator circuit gets enough to function. The datasheet of the 3.3V IC version states that minimum input voltage should be 4.75 V ! They must have chosen a 12 V raw supply just to work with lower input currents and have less ohmic losses, or perhaps just because 12 V is the most popular and cost effective.
Don't understand the constant current that you see. Would expect a constant power (consumption of the circuits plus losses), i.e. input current inversely proportional to voltage at the input. But many current meters don't see the higher harmonics.

(Last edited by doddel on 30 Jul 2005, 21:38)

I measured power consumption (including power supply losses) of Linksys WRT54G v2.0 (with ADM multiplexer) and WRT54GS v1.1 (with Broadcom multiplexer). I have to admit that I have no idea how efficient is its power supply.
WRT54G v2.0 consumes 7.8W
WRT54GS v1.1 - 6.1W

Conclusion: at least current units (G v2.2+, GS v1.1+) should consume 5W or less - without loss in power supply.

EDIT: There are very little variations regarding current load of the device - within boot the device consumes basically the same amount of electricity as during normal operation.

(Last edited by tosuja on 1 Aug 2005, 17:40)

On the wrt54g 3.1, the power goes into a 12V regulator, then into a 5V, 3.3V and 1.8?V circuts.

Ditto regarding shorting pins.  I think I did that and blew up the BCM5325 chip.  No more ethernet :-(

Does anyone know if you can run one of these from Cisco POE?

There's a good page on the WRT54G's power supply here … yPoweredAP.  It points out that although the AC1501 voltage regulator is capable of a 40V DC input, the input capacitors are only rated to 25V.  Keep in mind also, that over time capicitors will dry out and de-rate as points out - so the input capacitors may only be able to handle 18V in a few years time.

This link: also has some really good data on the power consumption figures.
And this link … 00011.html has some more good general info about the WRT54G power supply.

(Last edited by danversj on 7 Aug 2005, 01:43)

The newer units V3 and later comsume very little juice. 250ma at 12V. Around 3 watts.

Wallace78 wrote:

Here some graphs on power consumption.
It is in italian but graphs don't need translations. … amp;sid=18

I can not understand why the power consumption measured in this site is so different from what I have calculated... They must have done something wrong. It can't be as low as 2-3 Watts.
I've repeated the measurements (with the WRT54GS idle in AP mode) and I got the same results (and other measurements agree).

doddel wrote:

don't know and don't fancy probing all those little solder pads, risking short circuit.

That's what happend on my router... Two days ago, I was doing some analysis on the PWM of the voltage regulator circuit when a probe slipped causing an instand short circuit between two pins of the AC1501 sad Now I get a wrong pulse on the regulator's output, which results lower output voltage (arround 2.4V)
If I'm lucky only the voltage regulator is burned (and I'll change it with an equivalent regulator), but I'm affraid that the damage is bigger. Now I have to desolder it to check it alone.

PS Thank you danversj for linking to my site smile

(Last edited by kioan on 14 Sep 2006, 04:05)

I once plugged an 120 VAC mains adapter into 220 VAC socket - RT31P2-NA gave smoke...

That's why you need to read the fine print on your mains adapter smile

I was even dumber than that!
Since the adapter was for americas (and I'm in Europe) - it did not fit into the socket and I used the special fitter to plug it in...

Hi there, I'm using a 7,5 VDC power supply but even if everything seems to work fine, I have all the lan and wan leds always on. But I repeat everything seems to work fine, I find this problem after some days without understanding what happens to the router.

Thanks a lot to everyone who contribute developing this "big thing" :-)


I experimented using a WRT54G v1.1 on a old car battery that was in very bad condition - would not work any more on a car - after having charged it. My findings : The drain power consomption was around 6 watts @ 12 V. The WRT kept on working down to 4.5 V. And it lasted 5 days...

I don't have any good condition battery to test it but I guess you can expect the WRT to stay alive for at least 2 weeks on a good 45 Ah/12V fully charged car battery.

Also, i remember while googling on that subject, that someone had a set up consisting in "deep charge" or "deep cycle" (I' can't remember the correct expression) batteries and solar panels. Maybe just forget the solar panels in Evry... very cloudy sky around here ! (I live in Morsang s/Orge).


Hi All,

We are deploying a pilot network using WRT54GLs, solar panels and deep cycle batteries. What is the power profile of the WRT54GL (though wall adapter seems to be supplying 12 V  1A), can we design our system at a lower voltage and be stable ? (like 9v 1A or less).

Any thoughts and experience would be valued.


If these numbers are accurate, maybe one could power a WRT54G(S) via USB 2.0. I believe it supplies 5V ~ 1A, so 5 watts. It would be an interesting experiment.

UPDATE: I was corrected, its 5v @ 0.5A apparently, so maybe not enough power.

(Last edited by db90h on 25 Nov 2006, 23:14)

Experience here is that as long as above the minimum voltage of appr. 7V and under the max. voltage the power supply IC in the WRT can handle (more than 20), anything goes as long as the product of external voltage and external current is at least 4-5 W for each wrt54gl.
A very reliable way of power supply is from a power supply with 13.6 V output (15V regulator IC with two diodes in series from regulator ground to real ground) that supplies one or more wrts and parallel to that a dry lead-gel battery. The battery is connected to the supply and wrts via a 10 ohm resistor that has a diode in parallel. When external power is gone the battery supplies the wrts via the diode. When external power returns but the battery would by then be empty, the resistor limits charging current thus leaving enough voltage for the wrts to start (empty batteries are effectively a short circuit). Without external power the battery only supplies the wrts because the regulator IC output circuit can only deliver but not sink current, so no diode needed there.
Very simple and hence reliable.

(Last edited by doddel on 26 Nov 2006, 16:57)

I tryed putting the stock DC power adapter and WRT54GL1.1 through an ampermeter, and i got results of 0.24A. At the beginning was 0.18A then after it jumped to 0.22A and at end i got 0.24A after it has loaded all the things, scripts etc.  I haven't tried it during a download or if the CPU is 100% used, but as far as i can see these results that i got, differ a lot from that you got. I got 11V(wire+internal resistance)*0.24A=2,64W

(Last edited by metala on 11 Dec 2006, 13:02)

Power consumption for my WRT54GS v1.1:

With WLAN on: about 250 mA
With WLAN off: about 170 mA

Power consumption drops to 170 mA right after I turn wlan off (nvram set wl0_radio=0; wifi).

EDIT: I measure from the AC/DC adaptor's wire, so that doesn't include loss caused by the adapor. I personally will use this unit in my car (with 12V system), so that doesn't matter to me :-)

(Last edited by keitsi on 11 Dec 2006, 18:04)

Hello hemshia....wit regards to ur query, the adapter if not connected to any device doesnot consume any electricity. It is as good as a disconneted circuit. Only in case of a four quadrant adapter i.e. if theres any lamp on the adapter which glows while it is on, a minimal electricity is consumed.

For any further queries visit

The discussion might have continued from here.