Mailing list for discussions and development of KORUZA, a system for wireless optical networks (in English).

List archive Help


Re: [koruza] Nice project!


Chronological Thread 
  • From: Musti < >
  • To:
  • Subject: Re: [koruza] Nice project!
  • Date: Thu, 14 Aug 2014 08:24:25 +0200

Hi Caleb,

thanks for your message.

SFP modules are not exactly ideal for FSO, but sufficient to built up to
100m links, having about 20dB link margin, which is fine even for worse
case weather condition, from the free-space optical path loss point of
view, not accounting for pointing errors and other mechanical imperfections.

I have shown it is possible to couple optical power from SFP to SFP via
free-space over 100m with only about 1-3dB loss on a clear day. Long
term operation is affected by mis-alignment due to random events as well
as due to thermal expansion of the structure. Thus KORUZA has motorized
alignment so this can be actively compensated.

TP-Link MC220L has an ethernet PHY chip, so some CRC should be done in
it, but not much more as a Layer 1 device. It uses Marvell 88E1111 chip
inside:
http://www.marvell.com/transceivers/assets/Marvell-Alaska-Ultra-88E1111-GbE.pdf
Feel free to see more details about it.

I have inquired about getting the chip and building a media converter,
however none of the companies making such things wish to deal with
anyone prior to NDA signing and a serious amount of money.

As for the failures, the most obvious one is definitely thermal
expansion, which is sufficiently slow to be auto-tracked once a good
enough algorithm is implemented.

Wind and vibration do cause some signal fluctuation, however it was not
observed thus far as very significant.

At least here in Slovenia, there was no extra loss observed due to dirty
lenses and they have are rather dirty.

As for the prototype, I am now working on finalizing all the parts and
building some more prototypes to test everything and release an alpha
version this fall, probably in a month or so.

Kind regards,
Musti



On 14.8.2014 5:34, Caleb James DeLisle wrote:
> Hello Musti and others,
>
> This is a very interesting project, thanks for taking it on.
> I'm curious about a few things (other than when will you ship a prototype
> :) )
>
> I was originally a bit wary of the idea of using an SFP module
> as a light source but you seem to have written a paper studying
> it. I'm not qualified to judge it as an EE but I am still curious
> about how you came to the feasibility conclusion and how much power
> those little guys have.
> I can't seem to find the paper so if you can point me in the right
> direction I'd be most grateful.
>
> Also I'm curious whether the TPLink device you selected does any
> processing of the ethernet packets. Most interestingly does it
> validate checksums and discard packets with bit errors or is it
> possible to use forward error correction higher up the stack.
>
> As always, I'm curious about the failure modes you're seeing in your
> tests, what's going wrong? Are they mostly getting knocked out of
> alignment by wind/micro-earthquakes/magic? Deposits on the lenses?
> Temporary fog/smog?
>
>
> Thanks for making such an interesting project.
>
> Caleb
>



Archive powered by MHonArc 2.6.18.

Top of page