How many decibels does this router radio REALLY output?

[harry newton]

How many decibels does the router/radio below really output?
('cuz 26 + 18 is too high)

In the before-and-after krack-update screenshot below, I'm confused because
the radio says transmit power is 26dBm into a built-in 18dBi dish antenna,
so the EIRP of the stock unit set for the USA would be too high at
26+18=44dBm in the USA.

And yet it's all set to the defaults (I didn't change the transmit or
antenna value since it's the stock unit that comes with the antenna).

BEFORE:
<http://wetakepic.com/images/2017/10/29/01_PB400_firmware_update_krack.jpg>

AFTER:
<http://wetakepic.com/images/2017/10/29/14_PB400_firmware_update_krack.jpg>

Of course, the AirOS firmware could secretly throttle the EIRP to 30dBM (or
whatever the USA legal limit is); but why would AirOS report bogus number
then?

What are those decibel numbers in AirOS actually *telling* me?

[Jeff Liebermann]

On Sun, 29 Oct 2017 16:42:22 +0000 (UTC), harry newton
<ha...@is.invalid> wrote:

>How many decibels does the router/radio below really output?
>('cuz 26 + 18 is too high)
>
>In the before-and-after krack-update screenshot below, I'm confused because
>the radio says transmit power is 26dBm into a built-in 18dBi dish antenna,
>so the EIRP of the stock unit set for the USA would be too high at
>26+18=44dBm in the USA.

Are you doing point to multipoint or point to point?

<https://www.air802.com/fcc-rules-and-regulations.html>

5GHz is different quite different. Might be helpful if you describe
what you're using for hardware, what band you're using, and your
wireless network topology.

>And yet it's all set to the defaults (I didn't change the transmit or
>antenna value since it's the stock unit that comes with the antenna).
>
>BEFORE:
><http://wetakepic.com/images/2017/10/29/01_PB400_firmware_update_krack.jpg>
>
>AFTER:
><http://wetakepic.com/images/2017/10/29/14_PB400_firmware_update_krack.jpg>
>
>Of course, the AirOS firmware could secretly throttle the EIRP to 30dBM (or
>whatever the USA legal limit is); but why would AirOS report bogus number
>then?
>
>What are those decibel numbers in AirOS actually *telling* me?

One of the dangers of updating firmware is that the saved or stored
values might not work with the new firmware if they've shifted memory
location between versions. This is almost guaranteed for major
revision changes in firmware. Less so for minor changes. When in
doubt, print your your settings, reset the router, and manually setup
the numbers. Krack firemare tweeks seem to be a rush job and fall in
the category of "customer tested software".

You mention that "it's all set to the defaults" which is rather
ambiguous. Did you reset the router after installing the firmware
update?

Busy patching the roof today before the sky falls on me tonite.
--
Jeff Liebermann je...@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

[harry newton]

He who is Jeff Liebermann said on Sun, 29 Oct 2017 12:41:28 -0700:

> Are you doing point to multipoint or point to point?
> <https://www.air802.com/fcc-rules-and-regulations.html>

Thanks Jeff.

Given the 2.4 GHz antenna dish is 18 dBi, the chart says the max is
a. Point to multipoint
(18dBm or 63mW transmit) into (18dBi antenna) = (36 decibels or 4 watts)

b. Point to point
(26dBm or 400mW transmit) into (18dBi antenna) = (44 decibels or 25 watts)

The radio is set up as an access point, exactly how you suggested Rod Speed
set up his antenna outside his house to feed his neighbor Wi-Fi where his
neighbor is only a few hundred feet away line of sight.

So I'm overdriving the radio by accident, I think.
I thought they throttled themselves so that they'd never exceed the limit.

> 5GHz is different quite different. Might be helpful if you describe
> what you're using for hardware, what band you're using, and your
> wireless network topology.

I have plenty of 5GHz Ubiquiti equipment (Rocket M5, for example, with a
30dBM dish), but this particular radio is one of my smaller ones, which is
the Ubiquiti PowerBeam M2 400, which is only 26 decibels of transmit into
an 18dBi antenna.

But even so, it has no problem getting to the legal limit it seems, where
it must be defaulting to the legal limit for point to point even though
it's set up as an access point hanging off a wired router acting as a
repeater.

It only has to paint the barn with 2.4GHz WiFi to feed the cameras which
are about a half kilometer away, which is nothing for this type of WiFi
radio.

> You mention that "it's all set to the defaults" which is rather
> ambiguous. Did you reset the router after installing the firmware
> update?

The radio rebooted itself after the firmware update.

What I mean by defaults is the radio is set up to be an access point at a
certain IP address in bridge mode, but that's all I set. The rest is
whatever defaulted after I set it.

So I didn't touch the antenna setting.
Nor did I touch transmit power.

So it *defaulted* to the 44 Watts, which, for a home radio, is overkill
unless I'm going 10 miles but it's not even half a mile that I'm going.

So I'll dial down the transmit power.
Thanks.

NOTE: For those reading this, if you need an access point to paint an area
between hundreds of feet and a couple of miles away with 2.4GHz or 5GHz
WiFi signal, you can get Ubiquiti or Mikrotik radios for about the same
price as you paid for your last home router - but which are so much more
powerful (aka directional) than a radio with an omni, that it isn't funny.

> Busy patching the roof today before the sky falls on me tonite.

Good luck with the roof!

Our power went out again yesterday (it goes out once a month) and we didn't
even have much wind coming up from your side of the ocean.

[Jeff Liebermann]

On Sun, 29 Oct 2017 21:35:35 +0000 (UTC), harry newton
<ha...@is.invalid> wrote:

>He who is Jeff Liebermann said on Sun, 29 Oct 2017 12:41:28 -0700:
>
>> Are you doing point to multipoint or point to point?
>> <https://www.air802.com/fcc-rules-and-regulations.html>

>Given the 2.4 GHz antenna dish is 18 dBi, the chart says the max is
>a. Point to multipoint
>(18dBm or 63mW transmit) into (18dBi antenna) = (36 decibels or 4 watts)
>
>b. Point to point
>(26dBm or 400mW transmit) into (18dBi antenna) = (44 decibels or 25 watts)
>
>The radio is set up as an access point, exactly how you suggested Rod Speed
>set up his antenna outside his house to feed his neighbor Wi-Fi where his
>neighbor is only a few hundred feet away line of sight.
>
>So I'm overdriving the radio by accident, I think.
>I thought they throttled themselves so that they'd never exceed the limit.

I deduce by reading between your lines that you have an unspecified
model 2.4Ghz radio and that you're using it in point-to-multipoint
topology. Therefore, you're limited to 4 watts EIRP which translates
into +26dBm into an 18dBi dish.

>But even so, it has no problem getting to the legal limit it seems, where
>it must be defaulting to the legal limit for point to point even though
>it's set up as an access point hanging off a wired router acting as a
>repeater.

As far as I know, there's no connection between the transmit power and
the mode of operation. Since the radio has no way to know what gain
antenna is attached, it has no way to determine if the radio is used
in accordance with 15.247. Setting up your Ubiquiti for bridging,
which is what's normally used for point to point, should not have any
effect on the transmit power.

Note that if this were a cellular system, which has ATC (automagic
transmitter control), the SNR (signal to noise radio) from the
receiver at the other end of the link would be reported back to the
transmit radio and adjusted for the minimum RF level necessary to
maintain a decent BER (bit error rate) or PER (packet error rate).
However, wi-fi is not cellular, so forget about doing it the way
you've been guessing that it works.

>> You mention that "it's all set to the defaults" which is rather
>> ambiguous. Did you reset the router after installing the firmware
>> update?
>
>The radio rebooted itself after the firmware update.

Reboot is NOT the same as reset to defaults. When you reboot, all it
does is unload the current setting from working memory, and reload the
saved settings from NVRAM into working memory. When you update the
firmware, there's no guarantee the setting saved in NVRAM are going to
work correctly.

>What I mean by defaults is the radio is set up to be an access point at a
>certain IP address in bridge mode, but that's all I set. The rest is
>whatever defaulted after I set it.
>
>So I didn't touch the antenna setting.
>Nor did I touch transmit power.

Let me make this really simple. If you're radio is going goofy things
after a firmware update, punch the reset button and put everything
back to the factory defaults. Then, configure it to your favorite
working numbers and see if the problem goes away.

>So it *defaulted* to the 44 Watts, which, for a home radio, is overkill
>unless I'm going 10 miles but it's not even half a mile that I'm going.

Everything you do seems to be overkill.

>So I'll dial down the transmit power.

Dial, as in a knob and potentiometer? How quaint.

>> Busy patching the roof today before the sky falls on me tonite.
>
>Good luck with the roof!

I think I have all the potential leaks plugged. Of course, I've been
saying the same thing for last 15 years or so.

[harry newton]

He who is Jeff Liebermann said on Mon, 30 Oct 2017 19:28:04 -0700:

> I deduce by reading between your lines that you have an unspecified
> model 2.4Ghz radio and that you're using it in point-to-multipoint
> topology. Therefore, you're limited to 4 watts EIRP which translates
> into +26dBm into an 18dBi dish.

Yes. I was setting up an access point.
Your information helps because I want to stay legal.

For example, since a neighbor 15 miles away (by road) lost her Internet
yesterday in the Santa Cruz Mountains, I temporarily set her up using this
spare Rocket M2 I had lying around outside gathering leaves & rust:
<http://wetakepic.com/images/2017/11/01/wifi3.jpg>

It won't hold up against a windstorm as all I did was hand bolt it to the
top of a wooden stepladder just outside her house, on the ground, near
enough for the only (albeit short) Ethernet cord she had to reach the POE
injector just inside the house.
<http://wetakepic.com/images/2017/11/01/wifi4.jpg

Luckily, as you know, the weather is calm these days out here so it will
hold up until her ISP gets around to fixing the blown access point she was
connecting to. Meanwhile, we're getting pretty good (-60dB) signal pointing
to an access point that is over 3 miles (over 5 kilometers) away with this
temporary rig.
<http://wetakepic.com/images/2017/11/01/wifi_2.jpg

> As far as I know, there's no connection between the transmit power and
> the mode of operation. Since the radio has no way to know what gain
> antenna is attached, it has no way to determine if the radio is used
> in accordance with 15.247. Setting up your Ubiquiti for bridging,
> which is what's normally used for point to point, should not have any
> effect on the transmit power.

This is good information becuase, for some reason (the forums?) I was under
teh impression tha tthe radios were "hard coded" to not exceed the legal
EIRP under any circumstances.

Yet, as can be seen in the screenshot below, this Ubiquiti Rocket M2 I
temporarily set up for my neighbor miles away, is transmitting at 24 dBm
into a dish that is 24 dBi, for a whopping calculation EIRP of 48 dBm.
<http://wetakepic.com/images/2017/11/01/wifi.jpg

> Note that if this were a cellular system, which has ATC (automagic
> transmitter control), the SNR (signal to noise radio) from the
> receiver at the other end of the link would be reported back to the
> transmit radio and adjusted for the minimum RF level necessary to
> maintain a decent BER (bit error rate) or PER (packet error rate).
> However, wi-fi is not cellular, so forget about doing it the way
> you've been guessing that it works.

In this case of the temporary setup, the measured noise is about -84 dBm,
with a signal of -60 dBm, so we have a headroom of about 24dBm, which is
way more than the 10 dBm to 15 dBm I find out, empirically, that seems to
be the minimum required for a stable connection.

> Reboot is NOT the same as reset to defaults. When you reboot, all it
> does is unload the current setting from working memory, and reload the
> saved settings from NVRAM into working memory. When you update the
> firmware, there's no guarantee the setting saved in NVRAM are going to
> work correctly.

I agree. In fact, I sometimes get *better* performance when I update the
firmware, and sometimes *worse* (especially when they change how the radio
reacts to denial-of-service attacks which consume the CPU bandwidth (it
seems).

I had to talk to my ISP to get him to block certain IP addresses on his
side, instead of me having to do so one by one on my side.

But I agree, there are three things that are different:
a. Merely rebooting the radio (aka power cycling)
b. Resetting the radio to factory default (aka factory settings)
c. Updating the radio firmware (aka the operating system)

> Let me make this really simple. If you're radio is going goofy things
> after a firmware update, punch the reset button and put everything
> back to the factory defaults. Then, configure it to your favorite
> working numbers and see if the problem goes away.

I agree. It's even *easier* than that because I save the configuration
files, so, in the end, all I have to do (for Ubiquiti equipment) is press
the reset button on the POE injector for about 15 seconds (until all the
lights flash on the radio), and voila! Factory settings.

Then I change my PC laptop to the 192.168.1.xx subnet, taking care not to
use 192.168.1.0, 192.168.1.1, and 192.168.1.255, or 192.168.1.20.

Then I log into the radio at http://192.168.1.20 with a username and
password of "ubnt" "ubnt", and load the previously saved configuration
file.

I love that the Ubiquiti equipment defaults to station mode, which means
you can plug it into any computer Ethernet port where it becomes your Wi-Fi
connection capable of connecting your laptop to an access point *miles*
away (depending on the access point of course) where you see a desktop
connected to the access point over 3 miles (over 5 kilometers) away.

> I think I have all the potential leaks plugged. Of course, I've been
> saying the same thing for last 15 years or so.

Let's hope this winter is wet, but I could do with a lot less rain than we
had *last* winter, where our power went out every two weeks or so, for
months!

You're a little closer to civilization than I am, but we're both in the
same weather patterns.

Thanks for your helpful advice!

[Jeff Liebermann]

On Wed, 1 Nov 2017 13:59:19 +0000 (UTC), harry newton
<ha...@is.invalid> wrote:

>He who is Jeff Liebermann said on Mon, 30 Oct 2017 19:28:04 -0700:

>> As far as I know, there's no connection between the transmit power and
>> the mode of operation. Since the radio has no way to know what gain
>> antenna is attached, it has no way to determine if the radio is used
>> in accordance with 15.247. Setting up your Ubiquiti for bridging,
>> which is what's normally used for point to point, should not have any
>> effect on the transmit power.

>This is good information becuase, for some reason (the forums?) I was under
>teh impression tha tthe radios were "hard coded" to not exceed the legal
>EIRP under any circumstances.

The maximum transmit power and operating channels of the radio are
controlled by the country code setting. Anything sold in the USA is
suppose to be "locked" to US power limits:
<https://www.cisco.com/c/en/us/td/docs/wireless/wcs/3-2/configuration/guide/wcscfg32/wcscod.pdf>

>Yet, as can be seen in the screenshot below, this Ubiquiti Rocket M2 I
>temporarily set up for my neighbor miles away, is transmitting at 24 dBm
>into a dish that is 24 dBi, for a whopping calculation EIRP of 48 dBm.
> <http://wetakepic.com/images/2017/11/01/wifi.jpg

If it's point to point, that's legal.
For fixed Point-to-Point it's 1 watt (+30 dBm) minus 1 dB for each 3
dB of antenna gain greater than 6 dBi.
FCC Part 15.247(b) and (c), and Part 15.407(a).

>In this case of the temporary setup, the measured noise is about -84 dBm,
>with a signal of -60 dBm, so we have a headroom of about 24dBm, which is
>way more than the 10 dBm to 15 dBm I find out, empirically, that seems to
>be the minimum required for a stable connection.

The fade margin, headroom, SOM (system operating margin), Eb/No
(energy per bit / spectral noise density), SNR (signal to noise
ratio), or whatever you wish to call it, varies with the data rate.
The faster you go, the more fade margin you need for a constant BER
(bit error rate) or PER (packet error rate). It also varies with
modulation method.
<http://www.eletrica.ufpr.br/evelio/TE111/Eb_N0.pdf>
From my archives:
Minimum signal to noise ratio (Eb/No)
for various speeds at a BER = 10E-5 using
802.11b and 802.11g
Speed SNR(dB)
11 6.99
5.5 5.98
2 1.59
1 -2.92
54 24.6
48 24.1
36 18.8
24 17.0
18 10.8
12 9.0
9 7.8
6 6.0

>> Let me make this really simple. If you're radio is going goofy things
>> after a firmware update, punch the reset button and put everything
>> back to the factory defaults. Then, configure it to your favorite
>> working numbers and see if the problem goes away.
>
>I agree. It's even *easier* than that because I save the configuration
>files,

Nope. I don't think you understand the problem. The structure of the
configuration files might change between firmware versions. Often,
it's something subtle, like the length of the field length in an
array. If you restore the previously saved configuration file, you're
putting everything back the way it was before it became screwed up.
The idea is to use the new and improved structure, from the upgraded
firmware, not whatever is lurking in your saved config file. To play
it safe, after a few surprises, I only restore config files to the a
device with the same firmware that was used to create it.

[harry newton]

He who is Jeff Liebermann said on Wed, 01 Nov 2017 08:45:13 -0700:

> The maximum transmit power and operating channels of the radio are
> controlled by the country code setting. Anything sold in the USA is
> suppose to be "locked" to US power limits:
> <https://www.cisco.com/c/en/us/td/docs/wireless/wcs/3-2/configuration/guide/wcscfg32/wcscod.pdf>

Thanks for that chart Jeff, as my Ubiquiti equipment, bought in the USA
whole, can only set itself to the USA, Canada, and, I think, Mexico (I'd
have to look) whereas my similar Mikrotik equipment (but bought as boards
and assembled in the USA) can be set to any one of 200 odd countries.

> If it's point to point, that's legal.

That explains the enigma, because it's certainly *saying* the EIRP is 48
decibels!
<http://wetakepic.com/images/2017/11/01/wifi.jpg

> The fade margin, headroom, SOM (system operating margin), Eb/No
> (energy per bit / spectral noise density), SNR (signal to noise
> ratio), or whatever you wish to call it, varies with the data rate.

Yes. I know. And I know you know.
I've seen all the datasheets, as have you.
For others though, here's a datasheet for just the radio part of my
Ubiquiti Rocket M2 setup (where the antenna is 24 dBi added to that).
<https://dl.ubnt.com/datasheets/rocketm/RocketM_DS.pdf>

> The faster you go, the more fade margin you need for a constant BER
> (bit error rate) or PER (packet error rate). It also varies with
> modulation method.

Yup.
As a rule of thumb, I look for 12 to 15 dBm "headroom" to solve all those
complexities. :)

> <http://www.eletrica.ufpr.br/evelio/TE111/Eb_N0.pdf>

Heh heh .... looking at the author's comments, I knew an electrical
engineer pHd named Bob Pease in the valley.

Same guy?

Tall. Thin. Sharp featured. Nerdy as hell. Don't ever get caught in a
corner with the guy discussing diffraction of electromagnetic beams!

Not like you or me. You looked like a hippy parks guy, while I look like a
pocket-protected balding computer nerd.

> From my archives:
> Minimum signal to noise ratio (Eb/No)
> for various speeds at a BER = 10E-5 using
> 802.11b and 802.11g
> Speed SNR(dB)
> 11 6.99
> 5.5 5.98
> 2 1.59
> 1 -2.92
> 54 24.6
> 48 24.1
> 36 18.8
> 24 17.0
> 18 10.8
> 12 9.0
> 9 7.8
> 6 6.0

Well, what's interesting about your numbers is that sometimes it needed a
*lot* less than my rule of thumb, and sometimes it needed a lot more.

Sigh. Such are rules of thumb.
Luckily, the setup I showed has the requisite headroom to handle the values
in your chart!

> Nope. I don't think you understand the problem. The structure of the
> configuration files might change between firmware versions.

OK. I belatedly get it. While it has never happened to me, I belatedly
understand your heart-felt advice to be wary of such changes.

Since I only use a small subset of Ubiquiti and Mikrotik stuff, and mostly
a very small subset of Ubiquiti radios, in practice, that hasn't happened
*to me*.

It has happened to me on Android, Linux, iOS, Windows, etc., but not on
AirOS or WinBox, yet.

> Often,
> it's something subtle, like the length of the field length in an
> array. If you restore the previously saved configuration file, you're
> putting everything back the way it was before it became screwed up.

See above. I agree. In principle.
Luckily, setting up a radio is easy most of the time.
a. It's either an access point,
b. Or it's a station (e.g., the barn cam)

> The idea is to use the new and improved structure, from the upgraded
> firmware, not whatever is lurking in your saved config file. To play
> it safe, after a few surprises, I only restore config files to the a
> device with the same firmware that was used to create it.

You have far too much real-world experience with such radios!
I have less experience in what can go wrong!

I appreciate the clarifying advice!
Thanks.