surprising AC68U range improvement with Vaesel's CFE and nvram edits

[RussellInCincinnati]

With RMerlin's observations in mind, that fiddling with transmit power is not likely to do anything great...and having no intention to boost routers above FCC regulatory limits of 36dBm EIRP in my US region (see parts of this thread below)...from the initial guidance of "Engineer", have further followed "Vaesel" suggestions for editing a 1.0.2.0 (US) boot loader CFE for AC68U routers (you only need to look at the notes after the heading "Edit CFE" in his post):
http://forums.smallnetbuilder.com/showthread.php?t=19262
), for maxing transmitted WiFi signal milliwatts from 100 to 780.

By the way Vaesel describes steps of changing firmware post-CFE change, that are at least not essential in the case of router already running Merlin ...49. That is to say, after you install the new CFE and issue a "reboot" command, your Merlin ...49 firmware is still good. All you have to do with the Merlin menus is go in and do a "restore factory default" action. And you're back in ...49 with a new potential for higher transmit power, that you will soon finalize with some nvram set commands later.

A compliment to that post describing a specific CFE (bootloader) mod, is this "hggomes" post suggested by "Engineer" on the general procedure for burning CFE 1.0.2.0 US (or any other CFE) into RT-AC68U or TMobile TMAC1900 routers, using the "mtd-write v2" Linux utility:
http://slickdeals.net/forums/showpost.php?p=70991738&postcount=2576

After copying out, editing and burning the changed CFE back into a router that was already running the latest Dec 2014 ...49 firmware, had to reset the thing back to factory defaults through the Merlin menus. Then you must reconfigure all your custom settings, but particularly finish up re-establishing any wireless settings.

Now there are a few Xshell4/SSH or telnet login and then "nvram set" commands AFTER using the Merlin menus...to finish setting up the higher-powered wireless connections. That is to say the several nvram set commands you need to do to polish off this power boost, apparently get overwritten, each time you change wireless settings through the Merlin menus. So you do the "nvram set" commands last.

Not forgetting that you have to enable SSH or telnet to be able to reach the Linux prompt in the router, to do the nvram set commands or indeed any CFE retrieval or reflashing.

Am reporting that the increase in range and link stability, and reduction in ping times, to distant iPad, tiny USB WiFi adapters and cheap Alcatel smart phones is immediately and pleasantly noticeable. And works great with Merlin ...49 firmware.

In my busy areas have confined 2.4ghz router settings to 20mhz bandwidth only, so the speeds are not mind-blowing. But being able to walk out of my large brick building where the router is several walls away from the outdoors, and go out to a parked car on the street and still have a 2.4ghz connection, is a little bit of an eye-opener.

Am also surprised because am sure RMerlin knows what he is talking about, and he has lucidly described the probable uselessness of increasing the transmit power of just one side of a WiFi "conversation". After all every WiFi transaction involves both a transmit AND a receive.

Considering also that you may be hurting something by sending more power through tiny bits, am offering this non-statistically-significant experimental result more as a fun topic for weekend fiddling. Rather than a broad recommendation that we can all get closer to Heaven this way. Especially since it is real easy to destroy a router by loading it up with a bootloader you have edited by hand, that has a single wrong character anywhere in the Ethernet MAC addresses you copy into your new CFE.

Anyway thanks much, RMerlin, Engineer, hggomes and Vaesel.

Not sure exactly what nvram show output to look for, to verify that the transmit power is higher. But testing certainly shows the difference anyway.

 

[TonyH]

Hi,
Did you happen to check the temperatures after TX power is pushed to said 780mW?
Or just feel router case and still warm to touch? Almost increase of 10 fold of TX power
ought to make a difference and at what cost? Premature failure of radio chips?

 

[Engineer]

A compliment to that post describing a specific CFE (bootloader) mod, is this "hggomes" post on the general procedure for burning CFE 1.0.2.0 US (or any other CFE) into RT-AC68U or TMobile TMAC1900 routers, using the "mtd-write v2" Linux utility:
http://slickdeals.net/forums/showpost.php?p=70991738&postcount=2576

Hey, I get a little credit for that post too!


Thanks for posting. I'm now tempted to play with the CFE and settings from the posted link!

 

[RussellInCincinnati]

radio changed from 48 to 51-52 degrees Celsius, not worrisome

Hi,
Did you happen to check the temperatures after TX power is pushed to said 780mW?
Or just feel router case and still warm to touch? Almost increase of 10 fold of TX power
ought to make a difference and at what cost? Premature failure of radio chips?

Well since statements about the future can't be proven to be true or untrue, and we have no idea of what the long term service life will be of Asus routers/boards/power supplies in any case, and we don't know what environments or loads will surround or stress any particular router, doubtful that we can armchair speculate the relative/possible reduction in service life caused by running another 0.7+ watts (say 3 BTUs of heat) through a router that consumes about 10-15 watts of power.

Will further note my and my clients' experience of usually getting tired of/disinterested in computer gadgets long before they generally "wear out". All the regrettable mistakes have seen in computer purchases over the years, weren't a matter of buying things at list price or pushing equipment too hard, it was always a problem of buying stuff that wasn't compatible/hardly ever got used.

Did note that my one AC68U originally showed 2.4 and 5ghz radio temps of 48 and 49 degrees Celsius under light usage. After boosting the output power the radios are now at about 51 and 52 degrees C.

 

[john9527]

I generally stay away from these power discussions and don't want to get into another discussion about regulatory limits (there is another thread dedicated to that)...

but just want to point out that the 'Vaesel' changes lead to an EIRP that is significantly over these limits in every locale when you factor in the antenna gain.

While I fully support tweaks that unlock country specific channel and power restrictions that are probably done to simplify manufacturing, distribution and support, I have a bit of a problem with changes that clearly violate intended regulatory limits (and could lead to further regulations for the manufacturers).

Just my personal opinion...

 

[RussellInCincinnati]

Am going to study this issue before considering changing 5ghz transmit power away from the Merlin defaults (i.e. to 780) in any but a temporary, experimental exercise on a single router.

My general understanding up until now is that there is some kind of 1000 milliwatt regulation on the 5 gigahertz channel in the United States. Will investigate how antenna gain goes into the regulatory calculations.

Also should have made it clear that despite my mentioning above of 2.4 gigahertz effective range changes with a cheap cellphone, in my test environment the only channels IN USE connecting to the test router are 5 gigahertz client adapters. Neither are there any other users of the 5 gigahertz channels that are detectable around the test-boosted router.

Thus after your reminder John9527 have turned off the 2.4ghz radio on the boosted router.

There is no burning ambition on my part to make the power experiments permanent. Will confess to some burning curiosity. And have already greatly benefited from the exercise in getting technically comfortable with editing nvram settings, and installing replacement bootloaders in Tmobile/Asus AC1900 routers. That is going to come in handy for many other reasons besides fiddling with transmit power.

So will not shed any tears if I end up turning these mods off. Especially since my current way of solving WiFi range problems to distant clients is using intermediate 30-35 dollar EdiMax WISP repeaters. Can't imagine any clients where range is truly important AND they can't afford such cheap repeaters, which also help isolate intranet traffic from the internet gateway.

 

[RussellInCincinnati]

OK some rough calculations. The Asus antennas on the AC68u are probably less than 5dbi gain, noting this on an Asus forum:
Here's the reply from Asus tech support.>
>Sir,

The antennas are 3.5 dbi each. This does not multiply when combined with other antennas. They are just 3.5 dbi each.

Kind regards,

Randall C.
ASUS Technical Support

from forum http://vip.asus.com/forum/view.aspx?id=20121207140134463

Can also see from
http://www.wisp-router.com/page.php?11

"Several of the FCC part 15 rules govern the transmit power permited in the ISM bands. Here is a summary of those rules:
Maximum transmitter output power, fed into the antenna, is 30 dBm (1 watt).
Maximum Effective Isotropic Radiated Power (EIRP) is 36 dBm (4 watt).
You can obtain the EIRP by simply adding the transmit output power, in dBm, to the antenna gain in dBi."

...that if you add 29dBm signal from an 800 milliwatt transmitter to a 4dBi gain from stock Asus antennas you get something like 33dBm EIRP. Which appears to be less than the US FCC limits. This is NOT my final thoughts on the issue, am doing more research on the matter.

 

[john9527]

This is why I generally stay out of these discussions

Here's a link to the last discussion I had on this same topic (and the various documents with different interpretations...you can pick and choose as you see fit).

http://forums.smallnetbuilder.com/showpost.php?p=143759&postcount=24

Have fun all...

 

[RussellInCincinnati]

Pondering the effect of having THREE 780 milliwatt radios on, with a 3-stream router like the Asus rt-ac68u...

There is an absolute max power allowed of 4 watts, 36dBm, considering antenna gain, from the sum total of a device's radios. Let's say the total pre-antenna power from three 780 milliwatt radios is 2.5 watts, or 34dBm. Let's also say that, according to http://www.cpcstech.com/dbm-to-watt-conversion-information.htm
that "If the power at the transmitter is lowered by 1dB, the antenna gain can be increased by an additional 3dB"

So if we have 34dBm pre-antenna total power from a router, and we theoretically had 6dBi gain antennas in use, that would end up right at the 36dBm US regulatory limit. My understanding is that the stock Asus antennas are 6dBi gain or less, based on the "Randall C" message quoted above in this thread from Asus tech support.

Again not my final thoughts on the matter, but having three 780 milliwatt radios in operation on a unit with modest-gain antennas might be OK.

Also note the existence of products like the Amped Wireless SB1000
http://www.ampedwireless.com/products/sb1000.html
that advertise boosting power to 1000 milliwatts. That Amped Wireless advertisement carefully (if in tiny print) notes that the boost has only been certified OK when used with a single certain Amped router, the SR150. However the existence of the advertisement gives me an overall clue that the idea of a 1000 milliwatt radio is not wildly against regulations now in the US.