Comment by ac29

7 years ago

I work with 900MHz radios and infrastructure, here's my thoughts:

> The frequencies are what I'm confused by as well. To have a half mile range at relatively low powers it would need to be a low frequency band.

> I'm guessing they haven't bought licenses from the FCC, so my bet is that they're using the 900 MHz ISM band.

Seems likely, and probably even works in controlled tests over relatively short distances. 900MHz meshing radios certainly exist, like this one (no affiliation with them, haven't used the product): https://www.rajant.com/products/breadcrumb-wireless-nodes/lx...

>Even with a high 30 dB signal to noise ratio, the Shannon channel capacity in a 26 MHz band is 50 Mbps.

Impossible in urban areas without highly directional antennas (which aren't small). The effective noise floor in most even semi-developed areas I've worked in is ~-100 to -80dBm - it can be a struggle even with expensive radios and fixed infrastructure. 30dB SNR just isn't going to happen over any appreciable amount of range on this sort of device (a high gain parabolic dish? sure). Using all 26MHz of the spectrum with an omni directional antenna (again, as would be required) also isn't going to happen except over relatively short distances.

900MHz is good technology, and can get robust, long distance links with the right infrastructure (I most work on 1-10mi links). The limitations on a pocket sized device with a pocket sized antenna mean this is unlikely to work @ 900MHz- especially in their fantasized scenarios of being in areas with no cell phone coverage (cell carriers have access to better bands, and far better infrastructure).

8 comments

ac29

Reply
athrowaway3z  7 years ago

> cell carriers have access to better bands, and far better infrastructure

What would be the best band for this kind of infrastructure? ( Disregarding any existing FCC licenses )

  • ansible  7 years ago

    > What would be the best band for this kind of infrastructure? ( Disregarding any existing FCC licenses )

    The frequency bands that have been licensed for cellular infrastructure.

    Its all a trade-off. Higher frequencies get you better data bandwidth, but require more power and get attenuated more by stuff (buildings, trees). Lower frequencies get you much better range for a given power level and better building penetration, but lower data bandwidth.

    Exceptions to this are the "crap" bands, like 2.4GHz and 60GHz. Microwave ovens (which are great interferers, BTW) use 2.4GHz because it is absorbed by water. Great for heating your food. No so great trying to transmit data in an outdoor environment with rain and fog.

    60GHz gets absorbed by the oxygen in the atmosphere, so that's another band that's only good for very short ranges.

    The bands the cellular carriers have now are a good trade-off between power requirements, data bandwidth, and other issues (like absorption). The ISM bands exist basically because nobody would pay to use them.

  • ac29  7 years ago

    The biggest advantage to the cellular bands is that they are exclusively licensed, so noise is a non-issue, or at least an issue that you are in control of, since you are the only one allowed to build infrastructure in a given area.

    Cell carriers have a wide variety of bands they can deploy, from frequencies above 1GHz, which doesn't propagate well through free space or obstructions, but allows for high-density, high-bandwidth use, to frequencies below 1GHz (600, 700 and 800MHz bands in the US), which propagate well in free space and reasonably well around and through obstructions (but cant be deployed with as high a density, since signals propagate well).

    That being said, the infrastructure is the biggest advantage.

    My gut says that absent FCC regulations and allocations, for wide area coverage something like 500-900 MHz is an ideal spot between propagation and required antenna size. Guess where TV stations are (historically at least)?

angel_j  7 years ago

What if you used the noise as the carrier wave for the signal? Like, if in an urban area you have a "stable" probability distribution of noise, modulate that.

  • Robin_Message  7 years ago

    You're being downvited because, whilst this is a clever sounding idea, unfortunately it doesn't work like that.

    The noise is random, so you can't modulate it because you can't predict what it will be. A probability distribution isn't enough; say you want to modulate by XORing with a random, uniformly distributed stream of zeros and ones. Knowing that distribution doesn't enable the receiver to extract the data.

    (Fun fact: GPS signals are under the noise floor, but detectable because they are encoded with a predictable pseudorandom carrier which can be found even in the noise.)