AW: [802.3ae_Serial] Issue with 1500 Power
Hi,
My question is now about the additional 3 dB you mention. Are the
additional 3 dB you are talking off the :
Vertical eye closure penalty ** (max) 3.0 dB
as stated in the receiver table? If this is true then maybe some text would
be helpful because this means the worst case sensitivity you require for
interworking is 3dB higher ( Which are the missing 3 dB).
You state a stressed sensitivity (in OMA) of -11.8 in the table I find -11.4
dBm.
However also in this case I have problems to see how an eye penalty can be
used for attenuation budget. For my interpretation this does mean that in
case of a signal with 3 dB eye penalty the full attenuation cannot be
bridged. This means that either it has to be made sure that this penalty is
not there (but why specifying it ) and should be removed or it has to be
made clear under which conditions the 13 dB attenuation can be reached as
this makes this optional. This is at least not really easy to understand.
So if I now follow this and add this penalty to the sensitivity we would
require a receiver that supports a OMA sensitivity of -14.4 (or following
your figures -14.8 dBmOMA) This translates into average power sensitivity
(for infinite ER) of about - 18 dBm + ER penalty. Is this consideration
correct in your eyes now?
Regards Juergen
> ----------
> Von: Peter Öhlén[SMTP:Peter.Ohlen@xxxxxxxxxxxxx]
> Gesendet: Mittwoch, 27. Juni 2001 13:27
> An: Rahn, Juergen (Juergen); _Serial PMD Ad Hoc Reflector (E-mail)
> Betreff: RE: [802.3ae_Serial] Issue with 1500 Power
>
> Here is a short power budget the way I see it. For simplicity, let the
> ISI & dispersion penalty be 0. All powers in OMA [dB].
>
> The transmitter output power is -1.39 dBm. At the receiver you have
> -14.39 dBm. I think the confusing part is the stressed sensitivity.
>
> The stressed sensitivity is measured with an eye having 3 dB eye
> closure. In addition to that, there will be a small additional vertical
> eye closure due to sinusoidal jitter, which I don't account for here.
> With the stressed eye the sensitivity in OMA (measured between the
> nominal 1 and 0 levels) is -11.8. If we instead of the stressed eye have
> an eye without the 3 dB eye closure, the sensitivity would be
> -11.8-3=-14.8 dBm. This would give you roughly 0.4 dB margin.
>
> To make this more accurate, you would have to take the sinusoidal jitter
> in the stressed eye into account (--> more margin), as well as some
> penalties that could occur in the real link (e.g. interferometric noise
> and other known and unknown penalties --> less margin).
>
> Hope this helps,
>
> /Peter
>
>
> > -----Original Message-----
> > From: Rahn, Juergen (Juergen) [mailto:krahn@xxxxxxxxxx]
> > Sent: den 27 juni 2001 12:30
> > To: stds-802-3-hssg-serialpmd@xxxxxxxx
> > Subject: AW: [802.3ae_Serial] Issue with 1500 Power
> >
> >
> >
> > Hi,
> > as discussed in the PMD call I understand there is a mismatch
> > in the values
> > defined for the 1550 case. The minimum transmitter OMA is -1.39dBm +
> > penalties. Link attenuation is set to 13 dB , so this would give a
> > sensitivity requirement of -14,39 dBm OMA ,. however the stressed
> > sensitivity (In OMA) is defined to be -11.4 dBm. so there are 10 dB
> > difference between those two values. The trade off that has
> > been introduced
> > (that I do not like, but this is a different discussion) will
> > shift this
> > complete link power level. When comparing this with powers
> > and sensitivities
> > as defined in ITU the following appears: ITU defines 2 dB
> > path penalty with
> > this OMA we would end at a minimum transmitter power average
> > of about - 1
> > dBm which is in line to ITU. with 109 dB attenuation there is
> > margin in the
> > ITU numbers, with 11 dB and ITU sensitivity (Which has been
> > confirmed by
> > measurements) we add up to 0. ( -14 dBm average power
> > sensitivity and 2 dB
> > penalty gives us the G.691 application). When simply taking
> > this transmitter
> > power and 13 dB attenuation we end up with 2 dB better sensitivity
> > requirement as currently experienced by measurements (worst case EOL).
> > This may start this conversation.
> > Regards Juergen
> >
> >
> >
>