Straightforward out-of-the-box solution for extending WiFi range

Discussion in 'Wireless Internet' started by Danny D'Amico, Dec 13, 2013.

  1. Hi Miso,
    I agree.

    Older datasheets have different information than newer ones.
    Here's a datasheet on the UniFi Access Point Long Range (UAP-LR):
    http://dl.ubnt.com/datasheets/unifi/UniFi-AP-DS.pdf

    It shows the EIRP is 27dBm + 3dBi = 30dBm (i.e., 1 Watt).
    Thank you for that suggestion. I suspect there are three options:
    1. Turn off the WiFi router SSID
    2. Keep the SSID, but put the wired access point far away
    3. Keep the SSID, but change it to a different SSID

    I think the third method might not work though, as it would be on
    a different network - but I'm not really sure since I've never used
    one of these $100 1 Watt access points before.
     
    Danny D'Amico, Dec 14, 2013
    #21
    1. Advertisements

  2. This is interesting, in that it's the solution that most
    of us can do for free (since there are a lot of a spare
    routers lying around).

    What I like about it, is that we re-use a spare router
    as a repeater.

    What I don't like is that the spare router is *still*
    anemic (they're all about 15 to 18 dBm into something
    like a 3 to 5 dBi antenna).

    But, since people around her have a spare router, it
    just might work fine!

    I'm not sure why DDRT is needed, since any router should
    be configurable into a repeater, right?
     
    Danny D'Amico, Dec 14, 2013
    #22
    1. Advertisements

  3. That makes sense.
    So, that would be 24dBm EIRP for that Buffalo (i.e., 1/4 Watt) versus something
    like 18 to 20 dBm EIRP for most Netgear routers (i.e., 1/10 Watt), so the
    Buffalo certainly has a place in this discussion - especially since it's an
    easy solution (which is part of the equation) for a grandmother to handle.

    By way of correction, I just found out that the EIRP of the UniFi access points
    I googled for (which I've never used) is only about 30 dBm (not 36 dBm) but even
    so, that's a full Watt.

    However, I think I have to give up on the access points as to complex.
    So, the Buffalo router has the advantage of simplicity.

    The advantage of the Buffalo is simplicity and 4 times the normal power.
    The advantage of the UniFy access point is sheer power (3x the Buffalo);
    but that comes at the cost of complexity. Sigh.

    What would be nice is a simple router that outputs 1 Watt (the legal limit).
     
    Danny D'Amico, Dec 14, 2013
    #23
  4. Hi Miso,
    I understand what you're saying, as it's almost impossible to get the spec
    for the antenna gain of a typical home broadband router without going to
    the actual FCC report on the specific device.

    So, pretty much, I assume *all* home broadband routers must be spec'ing out
    the EIRP, if they spec anything. So, at 24dBm (1/4 Watt) for the Buffalo,
    I would agree with you that this is the EIRP (and not the transmit power).

    The nice thing is that the Buffalo is (presumably) simple to set up, and,
    the EIRP is probably more than double the power of a "typical" home broadband
    router, which, I assume is around 20dBm (1/10 Watt) EIRP.

    However, just to state what I've found out, according to *this* datasheet,
    the Ubiquiti UniFy Access Point datasheet does specify the antenna gain and
    transmit power separately, so, I *assume* my EIRP calculations are correct:
    http://dl.ubnt.com/datasheets/unifi/UniFi-AP-DS.pdf

    In *that* datasheet, they say the antenna gain is 3dBi for the UAP-LR, and
    the transmit power is up to 27 dBm. So, I get 30 dBm effective (1 Watt).

    Of course, that huge power increase comes with an increase in complexity.
    Sigh... if only they made a 1 Watt home broadband router!
     
    Danny D'Amico, Dec 14, 2013
    #24
  5. This has always confused me.

    In a way, it makes sense, in that you can throw a football 50
    yards to your kid, but the kid can't throw it back.

    However, all the client has to to most of the time is RECEIVE
    the thrown football. So, it only needs receiver sensitivity.

    It doesn't need throwing power, right?

    It only needs the throwing power on the uplink, right?

    So, aren't most connections downlinks?
     
    Danny D'Amico, Dec 14, 2013
    #25
  6. Danny D'Amico

    Char Jackson Guest

    Using your analogy, the _requests_ are footballs thrown by the client, but
    the simple fact is that _everything_ needs to be acknowledged, (TCP anyway)
    so two-way comms are an absolute must.

    Greatly simplified.
     
    Char Jackson, Dec 14, 2013
    #26
  7. Danny D'Amico

    Char Jackson Guest

    Any wireless router can be used as an access point, and of course any access
    point can be used as an access point. So to me, it seems like a better
    solution to put a normal-powered AP (whether that AP starts out as a WiFi
    router or an AP) in the area where it's needed, versus trying to find
    something that's high-powered and possibly causes interference for 2 city
    blocks in every direction. Exaggerating to make a point.

    Regarding SSIDs and channels, if you go with multiple AP's, I would use the
    same SSID on all of them but each would be on a different channel. Your
    channel choices (for the 2.4GHz band) are pretty much limited to 1, 6, and
    11. Experiment with those to see where you get the best results.

    By the way, using multiple AP's does NOT mean the secondary AP's are on a
    different network. If you end up that way, you're misconfigured.
     
    Char Jackson, Dec 14, 2013
    #27
  8. I see now. Better than before.

    The request is the football thrown 150 feet by the father radio to the kid
    client.

    The kid client acknowledges receipt, but, only weakly, such that some of
    the acknowledgements never make it back to the father.

    So, the father throws the same football again, and again, because only
    some of the acknowledgements are getting back to him from the kid client.
     
    Danny D'Amico, Dec 14, 2013
    #28
  9. Danny D'Amico

    miso Guest

    I believe the TX power is for the transmitter plus antenna, especially
    since the antennas are integral. Hence the link I gave you where the
    poster thought the same thing.

    The N450 has removable antennas, but it is MIMO, so you really wouldn't
    do that. They are longish around around 5dBi.
     
    miso, Dec 14, 2013
    #29
  10. Danny D'Amico

    miso Guest

    Some of the chipsets spec the lowest useable signal for reception. So this dongle is good to -95dBm at the 11mbps setting. Thus if dad had
    a really really long arm, you could make your argument. But in real
    life, it is better to have a symmetric communications system.
     
    miso, Dec 14, 2013
    #30
  11. Danny D'Amico

    miso Guest

    Look at the receive spec of this dual bander. It craps out at -80dBm at
    the 11mbps spec. Like I said, these dual band devices tend to be worse
    than the single band devices.
     
    miso, Dec 14, 2013
    #31
  12. Wow. You have a point. -80dBm is way too insensitive.
    I generally don't read the sensitivity specs, simply because I *assumed*
    they're usually pretty good (in the -97dBm to -99dBm range).

    -80dBm is pretty bad.

    I've found, only from my experience, that I need about 12 to 15dBm of
    "headroom" between the signal and the noise in order to reliably connect.

    So, with a sensitivity of -80dBm, that would mean I would only connect
    when the signal is at about -65dBm to about -68dBm.

    And, as this picture shows, even my current 2.4 GHz rooftop radio, which has
    a good clean WiFi signal from only a couple of miles away, is only getting
    a signal of about -55dBm (which is a great signal strength):
    http://farm4.staticflickr.com/3790/11375313475_125699bec3_o.jpg

    My noise alone is -88dBm, as can be seen from that screenshot, so, no signal
    worse than about -72dBm to about -75dBm would even be connectable.
     
    Danny D'Amico, Dec 15, 2013
    #32
  13. Oh oh ... that's bad news.

    That would make the true (claimed) transmit power about 5dB below the
    24dBm, so that makes it about 19dBm transmit power.

    Given that my assumption is that most home broadband wifi routers have a
    transmit power of about 15dBm to about 18dBm being put into a 3dBi to
    5dBi antenna, that would mean that the Buffalo isn't all that more
    powerful than the rest of the pack.

    Of course, if it's as much as 3dB more powerful, it's still twice
    the power; but, it could be that it's only about 1dB more powerful,
    which isn't all that much.

    Did I do the math correctly to form that tentative conclusion?
     
    Danny D'Amico, Dec 15, 2013
    #33
  14. Danny D'Amico

    miso Guest

    Those receive specs are not the noise floor. In theory, you should be
    able to maintain 11mbps reception at that level. But probably that is
    measured with a direct connection and a step attenuator. So I get your
    point that you might want to be 10 to 15 dB above the minimum detectable
    signal to account for crud in the aether, multipath, etc.

    Those Tube-u usb dongles kick some serious ass for not much money. For
    yucks I ran the tube-u with a 16dBi gain panel antenna from a hill in
    Berkeley pointed at San Francisco. I never saw so many WAPs on a kismet
    screen. A few were Axis cameras with no cypto, but I have no idea if
    they could be viewed.

    Anyway, today dual band is not so good. I won't pronounce it crap
    because the 5.GHz band certainly has less players on it, so it could
    potentially be better. But the gear certainly doesn't have the specs of
    the 2.4GHz products. The bad part is all the clients these days tend to
    be dual band, so they are probably compromised a bit.
     
    miso, Dec 15, 2013
    #34
  15. It always confused me what the difference is. I never really understood
    these things as the only things I know are those encountered in daily
    use of the radio as my WiFi signal.
    http://farm8.staticflickr.com/7419/11378789605_1226611246_o.jpg

    I guess the "noise floor" is *my* noise (i.e., the environmental noise)
    of about -87dBm (which, I'm told, is quite high).

    So, even though the receive sensitivity of the Rocket M2 that I'm using
    is around -94dBm to -98dBm, I'm not sure how much that matters, in
    practice anyway, given my noise is far above that level already.
     
    Danny D'Amico, Dec 15, 2013
    #35
  16. I have a Bullet M2, which is hooked up to a different antenna about
    a mile or two away - for my second Internet feed - which worked just
    fine screwed directly into a planar antenna (low line losses).

    I just tried logging into it, but I forgot the password so I'll have
    to climb up there and reset it in order to see what noise levels it
    sees. Darn. I like the newer POEs that have the reset button in the
    POE, instead of having to use the reset button in the radio on the roof.

    I'm in the Santa Cruz mountains, so our line of sight is a good 20
    miles in most directions. We can, theoretically, pick up antennas
    from all over the place.

    My dream is to see if I can, technically, connect to a Starbucks
    20 miles away in San Jose. I'm not sure how that works, but, I see
    that it's all about the antenna.

    Luckily, the RocketM2 comes with a nice RocketDish antenna that is
    high gain; because I'm realizing it's more about the antenna than
    about the transmitter or receiver.
     
    Danny D'Amico, Dec 15, 2013
    #36
  17. In theory, the dual band should be the best of all worlds.
    It should have less noise.
    And, it should have double the decibels right out of the box.
     
    Danny D'Amico, Dec 15, 2013
    #37
  18. On Fri, 13 Dec 2013 01:48:52 +0000, Danny D'Amico wrote:

    UPDATE:

    After many hours of trying to get the settings just right, just now
    I was able to tremendously extend the WiFi range of my laptop, as a test,
    simply by connecting a Ubiquiti NanoBridge M2 feedhorn (sans dish antenna)
    to the Ethernet port.

    Here is my signal strength at the feedhorn, as seen through the laptop:
    http://farm3.staticflickr.com/2869/11399128184_ebab5e1de1_o.png

    Notice the noise is a tiny at -99dBm while the signal strength is huge
    at -44dBm (with a SNR of -44 - -99 = 55, if I did the math right).

    This gets me 130Mbps between my Linux laptop & the home broadband router.

    Here are the network settings that were necessary to make this work:
    http://farm3.staticflickr.com/2854/11399127954_02139418fd_o.png

    And, here are the access-point specific wireless settings to make it
    connect to my home broadband router's SSID:
    http://farm4.staticflickr.com/3733/11399100445_af45bef4c0_o.png

    With the dish antenna, that Nanobridge M2 has a gain of 41dB (i.e.,
    23dBm transmit power + 18dBi antenna gain), which is far too powerful.

    Since that calculates (if I did the math right?) to over 12 Watts, I
    had to lower the gain by removing the dish ... which dropped the gain
    down to 23dBm + 3dBi, or 26dB (which is a 0.4 Watts).

    Even that was far too powerful for use in my house, so I dropped the
    transmit power of the feedhorn radio down to 6dBm, so with the 3dBi
    feedhorn-only gain, the screenshots above are at 6+3=9dB (0.008W) EIRP.

    Even with the gain reduced as low as I could make it, I still got
    a connection strength of -44dBm and a connect speed of 130Mbps, so,
    it's at least a proof of concept that this is one way to extend the
    WiFi range of your laptop.

    My goal will be to try to connect to my home broadband router from a
    mile or two down the road... so that's what I'll try next.

    PS: Jeff Liebermann should be proud of me!

    Here's the howto I wrote up ... (it can also be used at coffee shops!)
    BEGIN: How to use a Nanobridge M2 as your laptop wireless NIC!

    0. I reset the Nanobridge M2 radio to default settings as per this video:


    I connected the POE to the Nanobridge M2.
    I reset the Nanobridge M2 back to factory defaults by holding the reset button down for 10sec (until all LEDs flashed)

    1. I set the Nanobridge M2 to be the Linux laptop wireless NIC as per this video:


    2. I turned off the wireless NIC inside the laptop with the hardware switch.
    Note: I could just as well have run this command on Ubuntu 13.10:
    $ sudo ifconfig wlan0 down

    3. I set the IP address of the laptop to be on the 192.168.1.XX subnet.
    $ sudo ifconfig eth0 192.168.1.99
    $ ifconfig (make sure eth0 is 192.168.1.something & that wlan0 is not up)

    4. I physically connected the Nanobridge M2 to the eth0 port of the laptop.

    5. I pinged the Nanobridge M2
    $ ping 192.168.1.20
    PING 192.168.1.20 (192.168.1.20) 56(84) bytes of data.
    64 bytes from 192.168.1.20: icmp_seq=1 ttl=64 time=0.572 ms
    64 bytes from 192.168.1.20: icmp_seq=2 ttl=64 time=0.460 ms
    64 bytes from 192.168.1.20: icmp_seq=3 ttl=64 time=0.286 ms
    etc. (control C to escape)

    6. I logged into the Nanobridge M2
    $ netscape 192.168.1.20 (ubnt, ubnt)

    7. I set the "Network" tab as follows:
    AirOS:Network
    Router (default is Bridge)
    WLAN Network Settings->DHCP (default is DHCP)
    LAN Network Settings->IP Address->192.168.10.20 (default is 192.168.1.1)
    [x]Enable NAT
    [x]Enable DHCP Server
    Range Start=192.168.10.100
    Range End =192.168.10.200
    Change->Apply

    8. I rebooted the Ubuntu PC (with the wlan0 card still turned off)

    9. I set eth0 to be on the same (new) subnet as the Nanobridge M2:
    $ sudo ifconfig eth0 192.168.10.101

    10. I pinged the radio:
    $ ping 192.168.10.1
    PING 192.168.10.1 (192.168.10.1) 56(84) bytes of data.
    64 bytes from 192.168.10.1: icmp_seq=1 ttl=64 time=1.15 ms
    64 bytes from 192.168.10.1: icmp_seq=2 ttl=64 time=0.255 ms
    64 bytes from 192.168.10.1: icmp_seq=3 ttl=64 time=0.310 ms
    etc. (control + C to escape)

    $ ping 192.168.10.20
    PING 192.168.10.20 (192.168.10.20) 56(84) bytes of data.
    64 bytes from 192.168.10.20: icmp_seq=1 ttl=64 time=1.71 ms
    64 bytes from 192.168.10.20: icmp_seq=2 ttl=64 time=0.308 ms
    64 bytes from 192.168.10.20: icmp_seq=3 ttl=64 time=0.242 ms
    etc. (control + C to escape)

    11. I logged into the Nanobridge M2:
    $ netscape http://192.168.10.20 (ubnt, ubnt)

    12. I set up the "Wireless" tab to connect to the home broadband router SSID:
    AirOS:Wireless
    SSID->Select (I sorted the signals by signal strength & encryption)
    I selected my WPA2-PSK encrypted network SSID.
    I scrolled to the bottom & hit select.
    Change->Apply

    Note: I also had to set the DNS server by turning off DNS proxy
    Primary DNS server = 8.8.8.8
    Secondary DNS server = 4.4.4.2

    Voila!

    Once I set up DNS (which wasn't described in the video), I was able to
    connect to the Internet, and, in fact, am using this connection to type
    this up to help myself (in the future) and others.

    END OF: How to use a Nanobridge M2 as your laptop wireless NIC!
     
    Danny D'Amico, Dec 16, 2013
    #38
  19. Of course, that 130Mbps is only between the laptop and the router.

    The actual Internet speeds from my rooftop antenna to the Internet,
    about 20 miles away from a real wire, is almost ten times slower than
    those speeds, at about 10ms ping, 19 Mbps down, & 18 Mbps up:
    http://farm4.staticflickr.com/3743/11399640063_f0187217bd_o.png

    One mistake I made that I just corrected is that I hadn't enabled the
    NTP time server on the Nanobridge M2, which I hadn't used for months,
    so the time & date were off in the prior screenshots.

    Here's the latest, with the power dialed down as low as I can possibly
    make it, and even then, I get -39dBm signal strength from my laptop
    to my home broadband router, across floors and walls in the house:
    http://farm4.staticflickr.com/3746/11399676383_5ca5354629_o.png

    I can't wait until daylight to see how far I can go to connect my
    laptop to my home broadband router.

    ADVANTAGE: This method should extend the WiFi range of the laptop up
    to a few miles (depends on the antenna & access point though).

    DISADVANTAGE: 120V power is necessary to run the POE, so, I can't
    do it on foot; but I can test it a mile or two away from the house
    in my car with an inverter.
     
    Danny D'Amico, Dec 16, 2013
    #39
  20. Danny D'Amico

    Tony Hwang Guest

    Hi,
    Do you work fpor Ubiquiti? Nothing new you are doing there.
     
    Tony Hwang, Dec 16, 2013
    #40
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.