| Thread Links | Date Links | ||||
|---|---|---|---|---|---|
| Thread Prev | Thread Next | Thread Index | Date Prev | Date Next | Date Index |
|
Re-send - I thought I sent this to the
reflector, but it seems not. Haim, The optical gain that you get from FEC can
be well estimated from looking at the BER vs optical power curves and taking
the difference between the power required for 1E-12 (i.e. the no FEC case) and
the power required for ~ 1E-4 (the BER at the input of RS (255, 239) FEC for 1E-12 at the output. While I agree the optical improvement will
be greater in the case of an APD (because the BER vs power curve is not so
steep) typically the slope of an APD curve is not half that of a PIN and
therefore the optical improvement will not be twice as much. I also take issue with the value of 1.5 dB
improvement for a PIN. The theoretical improvement for RS (255,
239) FEC (net coding gain) is 2.8 dB. Of course this
assumes that the only penalty for the 7% increase in bit rate is additional
noise caused by a wider electrical bandwidth. If the receiver used for
the measurement has too little bandwidth even for the lower rate, then you can
see a large sensitivity penalty for 7% increase in rate. Alternatively,
if a large dispersion penalty is included in the measurement setup you can see
a big increase in this for the increased rate. If either of these cases is not true, then
the BER vs power curve for typical PIN based receivers is, if anything,
somewhat less steep than the theoretical curve leading to a larger improvement
for the with FEC case than theory would suggest. As a consequence of all of this, I am
surprised by the claim that RS (255, 239) FEC
gives only 1.5 dB improvement for a PIN receiver if this is in the context of a
continuous downstream signal. Burst mode in the upstream has, however,
been said to be potentially different because of some OLT receivers showing
very abrupt BER vs power curves. Nortel Networks UK Limited, External +44 1279 402540 ESN 742
2540 Fax +44 1279 402543 From:
Haim Ben-Amram [mailto:Haim_Ben-Amram@PMC-SIERRA.COM] Pete, Regarding your first
question. RS FEC can provide you
3-4dB optical gain, but depends on the receiver type. In case of APD, which is
limited by Shot Noise the coding
gain is quite similar to the electrical SNR. While in PIN, which is limited by
the Thermal Noise the coding gain
is around half of
the electrical gain (APD gain). Haim. From:
Robert, Thanks for the notes from the Ad-Hoc call. I have comments on three
sections: “It was stated that for a 7% overhead
code, such as RS, a 1.5 dB coding gain would be appropriate for PIN case, but a
3 dB gain for APD case. It was not known how to evaluate the pre-amplified PIN
case. This seems to be a major area requiring clarification.“ Was this for RS (255,
239) as defined in G.709? Upstream or downstream? If it
is for (255, 239) downstream I find this surprising
are there measurements to support this? Secondly, “Assuming a value of 13 dBm for minimum
Psat and subtracting 3 dB to avoid patterning and 2 dB for min/max range would
yield a minimum SOA output power of 8 dBm as a working assumption.
” I don’t think that there is any need to subtract
anything for the min/max range. If +13 dBm
is the guaranteed minimum Psat, then subtracting 3 dB to avoid distortion would give
+10 dBm minimum. Surely a
min/max range of 2 dB means that the spec for output power would be +10 dBm min and +12
dBm max? Thirdly, “It was not known how to evaluate the
pre-amplified PIN case. This seems to be a major area requiring clarification.” There is some information on the slope of the BER vs power curve for pre-amplified receivers in
Supplement 39 to the G-Series Recommendations clause 9.5.2
which is relevant to the FEC gain that might be
expected. This can be found at: http://www.itu.int/rec/T-REC-G.Sup39/en Regards, Nortel Networks UK Limited, External +44 1279 402540 ESN 742 2540 Fax +44 1279 402543 _____________________________________________ Please
find attached the notes from today's call. Please
review it carefully and discuss issues on the reflector. If you wish to
volunteer to take action items, please let me know. Otherwise I will raise
issues on the reflector and begin asking people to take action items :-) Robert -----Original
Message----- From:
Sent:
Tuesday, February 06, 2007 5:27 PM To:
STDS-802-3-10GEPON@listserv.ieee.org Subject:
Re: [8023-10GEPON] [POWER_BUDGET] ad hoc group 2nd meeting Please
find attached the file for the call today 2/6 at 7PM EST and 2/7 at 9AM. -----Original
Message----- From:
Sent:
Friday, February 02, 2007 4:37 PM To:
STDS-802-3-10GEPON@listserv.ieee.org Subject:
Re: [8023-10GEPON] [POWER_BUDGET] ad hoc group 2nd meeting Please
find attached the notes for the first Power Budget Ad Hoc meeting. Please
direct any corrections to me. Robert -----Original
Message----- From:
Sent:
Thursday, February 01, 2007 3:09 PM To:
STDS-802-3-10GEPON@LISTSERV.IEEE.ORG Subject:
[8023-10GEPON] [POWER_BUDGET] ad hoc group 2nd meeting All, I
will send the notes from this week's meeting tomorrow 2/2. As
discussed in the previous call, the next call(s) will occur on Tuesday Feb 6 at
7PM EST and Wednesday Feb 7 at 9AM EST. Notes
from this week's call as well as agenda for next week will be sent Friday. Toll-free
in US 866-263-8899 Outside
US 816-249-6061 Conf.
Code 7707985015 Robert Robert
Lingle, Jr. Fiber
Design and Transmission Simulation OFS Corporate R&D, 404-886-3581
(cell) 770-798-5015
(office) |