Control Channel on 2.4GHz

[john9527]

Did you restart the acsd daemon? I'm not sure if it gets restarted with the wireless services.

I didn't, but from the pid it looks like it restarted.

I separately killed and restarted the acsd daemon, and the nvram setting still held on my fork.

 

[sfx2000]

That setting controls Broadcom's own ACSD daemon.

Outside of auto - I'd agree...

ACSD is a mess these days...

 

[sfx2000]

Channel selection that we did for auto channel selection on my platform about 6 years ago...

As you can see, there's a fair amount of weighting on the scans - 1/6/11 get more attention that others... and the others, it was a distribution that we obtained from drive testing...

Within the data set, we did look at OBSS (overlapping WLAN networks) - this was really focused on high quality WiFi for VOIP solutions back in 2011 or so...

channellist=IEEE80211b:1:3,6:3,11:3,2,7,3,8,4,9,5,10

 

[eastavin]

If everyone stuck to using 1 /6 /11 the wifi in your neighbourhood would work best as the wifi protocol is designed to work with other signals on the same channel. However it sees signals on other channels that overlap it only partly as noise which degrades everything.

I have often used my Wi-Fi analyzer on my smartphone to identify neighbours that are on overlapping channels and then politely inform them and ask them if they would consider another setting. So far all have considered my input, some have moved and some have not. I think a leaflet/handout stuck in doors may be another way to bring attention to the fact there are only 3 clear channels on 2.4 Ghz.

If this becomes tiresome consider a move to the 5 Ghz band. The advantage there is the range of all the devices will be much shorter due the higher frequency and there are more channels. Together this means less annoyance to deal with. Plus there are low power channels (2 digit) and higher power channels (3 digit). This helps you choose whether you like to be unreachable outside your home or satisfy other needs with a little more power.

 

[adampk17]

I'll admit I've learned something about the 2.4 GHz channels in this thread. Thanks OP for the topic.

One question. Earlier someone mention that using 20 MHz is desirable, or perhaps more considerate, over 40 MHz. Do you sacrifice something in that choice like potential speed?

 

[dlandiss]

(5 GHz) Plus there are low power channels (2 digit) and higher power channels (3 digit).

Can you confirm that? It always looked to me like I was getting the same signal strength at either end of the 5G band.

 

[dlandiss]

One question. Earlier someone mention that using 20 MHz is desirable, or perhaps more considerate, over 40 MHz. Do you sacrifice something in that choice like potential speed?

If the 2.4G band is sparsely populated you can gain speed with 40MHz channels. You would cause (and accept) more adjacent-channel interference so I'm not sure there is any practical gain, and it hurts neighbors (and yourself) if they have radios on those adjacent channels. See attached image.

 

[eastavin]

Can you confirm that? It always looked to me like I was getting the same signal strength at either end of the 5G band.

There is no really simple "confirmation" here. I don't think you said what product you have or what application it is for or where you live.

Rules and laws vary by country, application (ex indoor, outdoor, point to multipoint, point to point, frequency band and stuff like is the antenna integral or external etc... Each country has its own regulations regarding use of unlicensed spectrum and generally they specify what the "effective isotropically radiated power" or the "maximum spectral power density" maybe from the transmitter/antenna assembly for specific segments of spectrum. Manufacturers need to adhere to the rules on frequencies and maximum EIRP for products they ship and the application it is intended for and need to show a return on investment so features and channels included vary. So the "confirmation" depends on the age of the product you have (rules in effect when built), what application it was designed for and where you live.

I looked through several FCC and Industry Canada web pages and nothing really simple popped out at me as they are all a jumble of above issues. However a consultant has prepared a simple chart you can see here that gets to your question. http://www.semfionetworks.com/uploads/2/9/8/3/29831147/5ghz-details-canada-update-may-2015.png The FCC updated its rules to something very similar in 2014.

If your product is in USA or Canada and includes channels 36-40 you will notice they are lower power. If your product does not have these channels it may all appear the same to you provided you have no channels above 144. I have an assortment of old wifi products that lack the high and low channels depending on when I bought them and how cheap I was feeling

Then just to muddy the waters a little more... remember that absorption of radio signals of a given power generally increases as frequency increases (wavelength decreases) and that differing construction materials have different absorption characteristics. (ex going through free air is better than drywall which is better than going through red brick etc.)

Edward

 

[dlandiss]

I don't think you said what product you have or what application it is for or where you live.

The products, as might be expected on this forum, are both ASUS routers: RT-N66U and RT-AC68U. Both are running John's Fork of Merlin's firmware, and both are limited to 200mW on both 2.4G and 5G.

However a consultant has prepared a simple chart you can see here that gets to your question. http://www.semfionetworks.com/uploads/2/9/8/3/29831147/5ghz-details-canada-update-may-2015.png The FCC updated its rules to something very similar in 2014.

If ASUS, Merlin, or John limit the power to 200mW or less--and set 80mW as the default--that chart does not affect the results for most users in this forum.

If your product is in USA or Canada and includes channels 36-40 you will notice they are lower power. If your product does not have these channels it may all appear the same to you provided you have no channels above 144.

Both the RT-N66 and RT-AC68 have the higher and lower sets of 5G channels, as does every recent ASUS router I have seen.

Then just to muddy the waters a little more... remember that absorption of radio signals of a given power generally increases as frequency increases (wavelength decreases) and that differing construction materials have different absorption characteristics. (ex going through free air is better than drywall which is better than going through red brick etc.)

True, but I would not expect the absorption difference between 5200MHz and 5700MHz to be noticeable. In fact, this study surprisingly found only 2 or 3dB difference between 2.4G and 5G, even for wet cinder blocks.

As everyone says, "YMMV," but the original statement that "there are low power channels (2 digit) and higher power channels (3 digit)" seems to be contradicted by the fixed maximums permitted by the firmwares most often discussed in this forum--and probably by ASUS itself.

 

[eastavin]

True, but I would not expect the absorption difference between 5200MHz and 5700MHz to be noticeable. In fact, this study surprisingly found only 2 or 3dB difference between 2.4G and 5G, even for wet cinder blocks.

As everyone says, "YMMV," but the original statement that "there are low power channels (2 digit) and higher power channels (3 digit)" seems to be contradicted by the fixed maximums permitted by the firmwares most often discussed in this forum--and probably by ASUS itself.

My comments were general and not restricted to only Asus product. I am a newcomer to the Asus router family. So thank you for pointing out the the 66/68 have always had all the channels. In my collection of 10 year old+ routers from other manufacturers that was not always the case.

Yes that is a good study of differing materials. "only 3dB" is a 50% drop in power. Fact of radio propagation is the shorter the wavelength the more loss and the shorter the usable distance. Here is a note that may be helpful. http://www.linksys.com/us/support-article?articleNum=134478

I will leave it Asus RF engineers with the appropriate RF test gear to determine if and how their radio assemblies meet FCC rules on EIRP. Just because we can set a power output on a transmitter in firmware as users and fail to consider antenna gain and directivity does not make it legal. We may end up exceeding the EIRP. That is why I generally advise home users not to mess with power output or alternate antennas unless their goal was to reduce the EIRP. https://www.digikey.com/en/articles...standing-antenna-specifications-and-operation

 

[dlandiss]

Yes that is a good study of differing materials. "only 3dB" is a 50% drop in power. Fact of radio propagation is the shorter the wavelength the more loss and the shorter the usable distance. Here is a note that may be helpful. http://www.linksys.com/us/support-article?articleNum=134478
.
.
.
https://www.digikey.com/en/articles...standing-antenna-specifications-and-operation

Quite right, -3dB is half the power which would result in about a 30% decrease in range. That could easily be recovered with better antennas, which have the added advantage of improving reception as well as transmission.

Thank you for the two links. The Linksys article was pretty basic. The DigiKey article was a good reminder that anything put near an antenna affects it behavior. I have to keep reminding our installer to not put a router behind a heating duct.

There is one factual error in the Digi-Key tutorial: it would be pretty tough for an antenna to be resonant "at whole number ... fractions of the frequency of interest." Full-wave and half-wave are obvious, and by stretching the point one might include quarter-wave with a ground plane but that's really a half-wave antenna. But I can't see resonance at 1/3 wave, or 1/5, 1/6, etc. Shorter antennas like that have to be forced into resonance by adding reactance somewhere in the antenna but gain suffers.

Fortunately that does not hinder Digi-Key's main point, which is that the size of the ground plane (or any other nearby conductor) alters the radiation pattern. We learned that back in the 1960's with CB whip antennas on the back bumper of cars: the range ahead of the car was always better than behind it.

Newcomers to the world of RF have to learn these things one at a time, often by experience because the textbooks get lost in mathematical theory and teachers only have time to cover the basics. I learned a fair amount about antennas in college, but most of it outside of class<G> in the past 58 years.

 

[eastavin]

Quite right, -3dB is half the power which would result in about a 30% decrease in range. That could easily be recovered with better antennas, which have the added advantage of improving reception as well as transmission.

There is one factual error in the Digi-Key tutorial: ... I can't see resonance at 1/3 wave, or 1/5, 1/6, etc. Shorter antennas like that have to be forced into resonance by adding reactance somewhere in the antenna but gain suffers. .

Congratulations on a long career in electronics. Kids growing up now days miss out on the simplicity of vacuum tube theory I say. There is nothing like knowing how 1 field effect transistor functions for actually understanding the stuff.

Yes better antennas can help. But as you know at some point distance wise the loss is just too great or the noise floor too high. If we calculate a simple space loss for a given signal at 2.4 GHz and around 40 feet (typical house perhaps front to back) we find that the signal loss is about 61dB. raise the frequency to 5.1 Ghz (low end of 5gig wifi) and the loss is 6.6 dB higher than at 2.4Ghz. 5.8Ghz and the loss is 7.7dB higher. That translates into a power 1/4 of the 2.4 ghz at the low end of the 5ghz band and about 1/6 at the high end of the band. Now start adding other losses like materials that you mentioned... add in a noise floor and it gets to be fun real fast. Crank it up to 100 yards and the 61dB loss is around 79dB and so on. Its a wonder that we have gotten this stuff to work at all over the years eh?

Good catch on the Digikey text. I think the editor tried to clarify the author's work and messed it up. Pretty sure they meant 1/4, 1/2, 3/4... even 1/8, 5/8 would work. not 1/3.

 

[dlandiss]

Kids growing up now days miss out on the simplicity of vacuum tube theory I say. There is nothing like knowing how 1 field effect transistor functions for actually understanding the stuff.

My electronics professors were pretty savvy guys. One thing they did that really helped us make the solid-state transition was to teach vacuum tube circuit analysis with a Voltage Source or Current Source mathematical model of the tube. So, when transistors first became available (we got a CK722!) all we had to do was swap models.

If we calculate a simple space loss for a given signal at 2.4 GHz and around 40 feet (typical house perhaps front to back) we find that the signal loss is about 61dB.

In practice I find smaller losses. If I read -30dBm on 2.4G about 10' away in free space, and then move diagonally opposite in the house about 50' away and one floor down, through the floor and 2 thick (115-year-old) walls, I can still see -65dBm or about 35dB loss. Of course, this is depending on how the Wi-Fi chip reports received signals so it may not be laboratory-accurate.

raise the frequency to 5.1 Ghz (low end of 5gig wifi) and the loss is 6.6 dB higher than at 2.4Ghz. 5.8Ghz and the loss is 7.7dB higher.

The loss-vs-frequency curves are not smooth--they have peaks and valleys. I did an informal test here and without touching the router or access point antennas, in fact without even leaving my chair, I found the RSSI to be the same on channels 48-44 as on 149-153. In fact, the 149-153 signal sometimes completely dropped out (see attached).

Pretty sure they meant 1/4, 1/2, 3/4... even 1/8, 5/8 would work. not 1/3.

One of my most favorite antenna projects was a ground plane for 144MHz. I noticed that at a certain length greater than 1/4-wave the resistive part of the impedance was 52 ohms, but some reactance was added. That was easily tuned out with a small series variable capacitor so I had a ground plane antenna that inherently matched 52-ohm coaxial cable, and had a small amount of gain as well. That gave us a reliable 144MHz link from W0QEV at Washington University in St. Louis to my home in Illinois about 25 miles away.

Aren't antennas fun? You can try different configurations with inexpensive materials, and improve transmission AND reception at the same time.