Thoughts on Starry's new EHF wireless internet tech and propagation issues

[TVHeadedRobots]

I just read about Starry's new wireless internet tech (http://www.engadget.com/2016/01/27/aereos-founder-is-now-taking-on-isps-with-wireless-internet/) that supposedly uses the EHF band to transmit gigbit/s speeds over distances of up to a km. I'm perplexed as to how they might be handling the technical hurdles involved with non-line-of-sight comms in the EHF band. As I understand it there are many possible issues with propagation using the EHF band in the presence of obstructions; reflection, diffraction, etc. I also was under the impression that there is significant atmospheric absorption at these frequencies. Is transmission of a signal even possible over km distances in an application that is not line-of-sight? They are suggesting that you can stick a small antenna just outside your window on your house. Hoping to start a conversation with people far more versed in RF than myself and this was the best place I could think of to do so.

Thanks.

[HAL-42b]

I'm not very acceptive of brand name internet.

If this happens it must happen on the end user side not on the ISP side.

Otherwise you will not only pass ISP's traffic on your property at no charge but you will be also paying their power bills. No way!

On the consumer side the microwave links work well point-to-point. Ubiquity and other manufacturers have products on the market doing that already. The problem is the routing in what is essentially a mesh network. If the routing problem is solved there won't be a need for ISPs any more.

[TVHeadedRobots]

HAL-42b,

Maybe I missed something but I don't think that they were suggesting that the end user would be a relay. You would only be a consumer of the service. My question is how are they overcoming the technical obstacles of transmitting a, say 60GHz signal with 1 Gb of data riding on it every second with out line-of-sight. I'll look at the Ubiquity stuff to see if I can answer my own question with their docs. I actually didn't realize that they made stuff that operated in those frequencies.

Thanks.

[HAL-42b]

@ 60GHz it has to be strictly line of sight. They probably have a motorized mount inside that plastic can.

[tggzzz]

My question is how are they overcoming the technical obstacles of transmitting a, say 60GHz signal with 1 Gb of data riding on it every second with out line-of-sight. I actually didn't realize that they made stuff that operated in those frequencies.

I first saw a 60GHz link in 1996. It wasn't phased array, but had a 20dB antenna that was ~4cm long. It also had a thin piece of plastic over the antenna that acted as an anti-reflection coating in the same way that there in an anti-reflection coating on your camera's lens.  And that clearly indicates you aren't in Kansas anymore:)

A limiting problem back then was simple thermodynamics: the power transistors are necessarily small, and the power output is limited by the ability to remove the heat from a small area. There will have been some improvements since then, especially w.r.t. having many transistors in a phased array transmitter, but I'd still look carefully at that topic.

As for line of sight only. 60GHz signals have multipath as you would expect, but narrow-angle beams make multipath less predictable and stable. Cue phased array transmitters, but I'd like to see how they determine where to point the beam this millisecond.

I note that the consumer's antenna has to be mounted in a window. This is realistic given how 60GHz is not good at penetrating many materials - including foliage. 60GHz is, of course, in the "oxygen hole", so range will be limited anyway.

If there is a link budget published anywhere, I'd be curious to read it.

[TVHeadedRobots]

tggzz,

    Thanks for the insight. Anti-reflective plastic coating on the antenna, not in Kansas indeed. I did a bit more digging and found an article (http://techcrunch.com/2016/01/27/starry-internet-is-aereo-founder-chet-kanojias-latest-salvo-in-the-war-on-isps/#.vxfixtd:t3V2) that mentions the frequency is 30GHz (I was guessing when I mentioned 60GHz.) It also mentions that they are using an active phased-array to attempt to make this work and that the service is only aimed at densely populated urban areas. That seems to make a bit more sense than the first article I read, which didn't make mention of this caveat.

    If I find any documentation on the tech (link budget or not) I'll post it here.

Thanks.

[tggzzz]

tggzz,

    Thanks for the insight. Anti-reflective plastic coating on the antenna, not in Kansas indeed. I did a bit more digging and found an article (http://techcrunch.com/2016/01/27/starry-internet-is-aereo-founder-chet-kanojias-latest-salvo-in-the-war-on-isps/#.vxfixtd:t3V2) that mentions the frequency is 30GHz (I was guessing when I mentioned 60GHz.) It also mentions that they are using an active phased-array to attempt to make this work and that the service is only aimed at densely populated urban areas. That seems to make a bit more sense than the first article I read, which didn't make mention of this caveat.

    If I find any documentation on the tech (link budget or not) I'll post it here.

Thanks.

28GHz (IIRC) is old tech for last mile connectivity. Very definitely susceptible to foliage attenuation; path loss studies should be easily available.

[rfeecs]

There was a lot of hype a while back for LMDS, which was around 30GHz:
https://en.wikipedia.org/wiki/Local_Multipoint_Distribution_Service

It never caught on at the consumer level.  I think people underestimated the cost of doing things at 30GHz.

There is similar hype for various "5G" wireless network proposals that use millimeter waves, beam forming, etc.

[TVHeadedRobots]

Found this on 28GHz networks and outdoor propagation. Not exactly apples to apples, but interesting. http://faculty.poly.edu/~tsr/Publications/ICC_2013_28.pdf

[HAL-42b]

That was interesting. They do mention using steerable antennas, not phased arrays. This would be much cheaper than phased arrays.