[MLB-WIRELESS] Free Space Optics Info

[Todd Hubers]

Thanks Nick, that's a fantastic real-world description, which can help me
and others better understand FSO.

I am wondering whether FSO TX and RX beams can be made much fatter -
equivalent to having a satellite dish for RF systems (ie. a concave
mirror). I have read that optics are a significant cost component of FSO.
By dispersing the beam more, spiders, heavy rain and movement at least
should be reduced factors. Also an increased overall beam power should be
possible without compromising and eye safety is no compromised, possibly
helping to overcome fog and possibly condensation without header: the same
intensity (per mm of diameter) of the beam can be maintained, but with a
bigger beam a high power light will transmit overall. Fog: I think fog is a
considerable challenge in temperate climates such as Geelong/Melbourne. But
it's no like we're researchers or manufacturers of FSO links, so we're not
about to test such potential solutions.

1. I was wondering why short links had problems - seems like there are
minimum power outputs of TX. Perhaps a physical shade (like sunglasses)
could help for short links?
2. How does an FSO link compare in cost to an RF one?

For super short distance 10Gb, it's probably more economical to just haul
> the fibre.

Not for multi-user. See ournet.alivate.com.au - I'm thinking of a mesh
network with no long backhaul. Links are both last-mile and backhaul.

Todd

On Mon, Mar 30, 2015 at 2:51 PM, Nicholas Hobbs <Nickhobbs at optusnet.com.au>
wrote:

> I run a number ( 6 pairs) of FSO links to connect a number of our sites
> together. (5 at 1Gbit and 1 @ 100Mbit), some of which have given continuous
> service for 6+ years. Shortest link is approx 60m, longest is about 600m
>
> They are generally very reliable, however;
>
> In practice most people who sell and have to support them will not
> recommend them over 1km unless they are very high end multi beam self
> aligning units. Most units will have a documented db loss over distance and
> weather conditions. I switch to the 70/80Ghz microwave for runs over 800m
> and under 3km. 18 and 38Ghz for longer hauls.
>
> They require very stable mounts at both ends, think 75 or 100mm diameter
> pole, under 500mm long onto a very solid wall/floor mount. Both
> transmitting optic beams needs to line up with a receiver only 25-40mm wide
> at the other end.
>
> Whilst they may advertise 'no interference', fog, heavy rain, condensation
> on the lenses if they have no heaters and spiders (seriously) all present
> challenges.
> I have a pair of links that are 200m runs and get taken out by the fog in
> winter and they are located in inner Melbourne about 30-40metres in the air.
> I have one link with intermittent issues due to building
> movement/expansion/contraction. 40mm is a very small target when it's up to
> 1km away and doesn't give much margin for movement.
> Super short distances can be problematic as the units can potentially
> overload the receiving sensor.
> East-West links can be problematic due to sun glare in the
> morning/evening. As the sun tracks North/south in the skyline through the
> seasons, it may only be an occasional problem, and really tricky to
> troubleshoot unless you're there when it happens! (Had this happen too. My
> one East/Westish link drops at around 4.30pm for 5mins for 2-3 days in May
> when everything lines up perfectly!
>
> So in short they're really good sub 1km as long as you know the conditions
> they are in and the units can't move an inch (literally).
>
> For super short distance 10Gb, it's probably more economical to just haul
> the fibre.
>
> My 2c
>
> Nick
>
> On 30 Mar 2015, at 2:04 pm, Todd Hubers <todd.hubers at gmail.com> wrote:
>
> Well that is even more limited than what I recalled from our Geelong meet
> discussion.
>
> If that's the case, an FSO 1 gbps link over 2.5km is quite good then. I
> have a feeling the upper limit is very affected by fog (and other weather
> factors), and that a backup link is needed if full speed needs to be
> maintained the occasional time the fog is too thick. This puts it close to
> par with RF IMO (subject to price comparison). If used in a pure mesh
> network there may be one RF route and another FSO route, rather than having
> the redundancy on the link, it would be more economical to have redundant
> routes (given enough nodes).
>
> Cheap short p2p (in the order of ~100m [and up to 1km] distance) LED FSO
> links would possibly beat RF links. I'm considering setting one up,
> pointing across the road. Such links are keyed very simply suffering
> little, if any, MODEM delay. Much research has been (apparently 100s of
> Millions of USD) spent on developing long distance and fast FSO with no
> success. Lightpointe is one such effort, but bundles RF redundancy (=
> fail). It would be great if FSO research could focus on the 50-100m range
> to reach ~10Gbps speeds cheaply - maybe a good candidate for crowd funding
> one a successful prototype is completed.
>
> Would be a nice hack day - building an LED system that could reach at
> least 100Mbps.
>
> Todd
>
> On Mon, Mar 30, 2015 at 12:34 PM, <mw at freenet.net.au> wrote:
>
>> 24GHz Airfiber @ full speed is good for about 1km if you stay within the
>> ridiculously low ACMA regulations ;)
>>
>>
>>
>> There is a 5GHz model, though, that claims similar data rates which can
>> push a bit further, but you need a lot of free spectrum ;)
>>
>> Cheers!
>>
>>
>>
>> Mike.
>>
>  _______________________________________________
> Melbwireless mailing list
> Melbwireless at lists.wireless.org.au
> https://lists.wireless.org.au/mailman/listinfo/melbwireless
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.wireless.org.au/pipermail/melbwireless/attachments/20150330/0eb2c81e/attachment-0001.html>