[time-nuts] Using lasers for data transmission

Michael Wouters michaeljwouters at gmail.com
Sun May 1 08:36:37 EDT 2016


Dear Ilia

On Sun, May 1, 2016 at 6:40 PM, Ilia Platone <info at iliaplatone.com> wrote:
> The problem would be modulating a 10GBASE-T signals into a single laser
> beam, and demodulating it using (I think) an APD.
>

The White Rabbit cards use SFP (small form-factor pluggable) lasers
that plug into the card. These incorporate both the laser(s) (there
can be an uplink and downlink laser) and APD. Modulation in these
devices is simply by varying the current. This puts a frequency chirp
on the laser which exacerbates dispersive effects. You can do better
with an electro-optic device like a Mach-Zehnder intensity modulator -
but this isn't necessary.

> except the one depending on light travel, that shouldn't be a problem if
> using White Rabbit, there could be some problem with the modulating and
> transmitter/receiver delay response times.

I remember reading that the SFPs are a source of jitter but
nonetheless, sub-ns timing is achievable.

> I mean that lasers offer 100ps rise time, and the APDs I found offer 5ps
> rise time, these must be multiplied by all the wires that GigE needs, which
> should be 4 pairs if I remember correctly, and some are bi-directional, plus
> the modulation process.
>
> Ilia.
>

Cheers
Michael


More information about the time-nuts mailing list