Improvements to Interferometric Noise (IN) methodology
Hi all,
Peter Ohlen pinted out that the posted version of the excell file was
somehow corrupted. I am trying again, now just with that file. My original
message that did not go through is also forwarded. The other file is
already posted on the reflector and is OK.
Peter
Petar Pepeljugoski
IBM Research
P.O.Box 218
Yorktown Heights, NY 10598
e-mail: petarp@us.ibm.com
phone: (914)-945-3761
fax: (914)-945-4134
---------------------- Forwarded by Petar Pepeljugoski/Watson/IBM on
01/23/2001 01:01 PM ---------------------------
(Embedded Petar Pepeljugoski
image moved 01/16/2001 11:48 AM
to file: (Embedded image moved to file: pic19379.pcx)
pic28808.pcx)
To: stds-802-3-hssg-serialpmd@ieee.org
cc:
From: Petar Pepeljugoski/Watson/IBM@IBMUS
Subject: Improvements to Interferometric Noise (IN) methodology
It seems that the first message did not go through. I am trying again.
Peter
Petar Pepeljugoski
IBM Research
P.O.Box 218
Yorktown Heights, NY 10598
e-mail: petarp@us.ibm.com
phone: (914)-945-3761
fax: (914)-945-4134
---------------------- Forwarded by Petar Pepeljugoski/Watson/IBM on
01/16/2001 11:46 AM ---------------------------
Petar Pepeljugoski
01/16/2001 12:10 AM
To: stds-802-3-hssg-serialpmd@ieee.org
cc:
From: Petar Pepeljugoski/Watson/IBM@IBM Research
Subject: Improvements to Interferometric Noise (IN) methodology
Dear all,
To help our discussion regarding the interferometric noise (IN), for
tomorrow's teleconference I prepared a small document which contains some
useful formulas for the analysis of interferometric noise. The treatment of
the IN is probabilistic and is the one which is universally accepted in the
literature dealing with it (please see references). I believe that the
formulas can be easily used to simulate the effects of IN and help in
getting more accurate results. I will have soon some simulation results
that I will share with you.
Furthermore, independent of the probabilistic approach, I would like to
suggest several improvements in the way the penalty is calculated in the
approach that Krister used.
1. The interferometric noise occurs between the transmitter and the
receiver, i.e the signal does not enter the receiver and is not affected by
receiver's bandwidth. Therefore, in calculating the signal with
interference at point TP3 we should assume infinite bandwidth. The receiver
bandwidth can be applied to the received signal that experienced
interference.
2. If we assume that the signal at TP2 has overshoot, and therefore need to
modify the worst case interfering signal, then we also need to reduce the
ISI penalty at TP3, since the overshoot has faster rising edge, and will
produce opposite effect on the ISI in the fiber.
3. If the signal does not have an overshoot, then the worst case
interfering signal is the signal with maximum amplitude - therefore the ISI
considered is the one that the model will predict with infinite receiver
bandwidth.
The point that I am trying to make is that we can't set all the parameters
at their worst case values. Some of them are mutually exclusive and affect
the penalties in opposite ways.
I also replaced the ISI penalty in Krister's spreadsheet, to the value that
one gets when the bandwidth is set to infinite, in this case it is 1.06 dB.
The chart is in the attached spreasheet.
I hope that you will find my agruments plausible.
(See attached file: interferometric noise3a.xls)
Regards,
Peter
Petar Pepeljugoski
IBM Research
P.O.Box 218
Yorktown Heights, NY 10598
e-mail: petarp@us.ibm.com
phone: (914)-945-3761
fax: (914)-945-4134
pic28808.pcx
pic19379.pcx
interferometric noise3a.xls