Re: [8023-10GEPON] [POWER BUDGET] resolving differences over PIN vs. APD
All,
With the group's forbearance, I'd like to explain a confusion I had at the
last meeting, explain the source of it, and then come to my direct issue,
and then a pointed observation.
In the last meeting, I mistakenly had the idea that the 'APD' group wanted
to use APD only for class B++, and PINs for the other classes. I was wrong!
They indeed want to use APD's for all three classes.
I wondered to myself, why am I being so stupid? So I looked at the Takizawa
presentation again, and I notice that (page 9), the APD for class px10 and
px20 have a -26 dBm sensitivity, while for B++ it is -28 dBm. I think it
was the shift of numbers that made my brain make the wrong turn.
So, my question: why is the sensitivity of the B++ APD pegged at -28 dBm,
while for the lower classes it is -26 dBm? No rationale was given for this.
Previous Email correspondence asked about the 2dB reduction that the PIN
people put into their sensitivity (-16 dBm, rather than the usual -18 dBm),
and how come it wasn't taken for the APDs...
My pointed observation: I think that the sensitivity for B++ does not
include the extra 2dB 'hit' so that the transmitter stays under the magic +3
dBm number (which is the 'edge of possibility' for an unamplified
transmitter.) In other words, the APD team derated the PX10 and PX20
detectors by 2dB to reflect the 'low cost' factor, but they just couldn't do
it for the B++ budget.
Bottom line: I think there still is some equalizing we need to do on these
budgets. If "team PIN" hits itself with a 2dB penalty for 'low cost
optics', then "team APD" should also be hit with a similar number (for all
classes). To do otherwise is not a fair comparison.
This is probably the source of some of the differences on cost estimation.
Some folks seem to think that PINs will always be 2x cheaper and APDs -
maybe this is true if APDs continue to be the premium grade that the current
B++ proposal implicitly requires. Other people claim that APDs will come
down in cost premium - maybe this is also true, if the APD sensitivity is
derated by 2dB off of its 'nominal P2P' value.
Please discuss.
Regards,
Frank E.
-----Original Message-----
From: Lingle, Jr, Robert (Robert) [mailto:rlingle@OFSOPTICS.COM]
Sent: Thursday, June 14, 2007 12:13 PM
To: STDS-802-3-10GEPON@listserv.ieee.org
Subject: [8023-10GEPON] [POWER BUDGET] resolving differences over PIN vs.
APD
All,
As chair noted in his email of 6/12, please recall the discussion we had on
Monday evening in Geneva about pros and cons of PIN vs. APD in ONU.
Attached are the straw polls and points of argument.
I have identified three items in red to focus on initially.
<<APD vs PIN pro-n-con.pdf>>
1. An objection was raised to having an amplified EML at OLT due to
non-linear interactions between D/S and analog video overlay. David Piehler
supplied a response that will be posted separately to start a thread.
In two cases we had task force members stating directly contradictory
opinions:
2. one stated that PIN ONU leads to lowest fully subscribed cost, while
another stated that PIN ONU leads to highest fully subscribed cost.
3. one stated that high power EMLs (+2 to +3 dBm minimum output power) for
use with APD at ONU are commercially available today, while another stated
that such high power EMLs are not available today.
I would like to ask for volunteers who hold opinions on point #2 and point
#3 to volunteer to defend these those positions on the Reflector, with
back-up information.
Warmest Regards,
Robert
Robert Lingle, Jr.
Fiber Design and Transmission Simulation
OFS Corporate R&D, Atlanta
404-886-3581 (cell)
770-798-5015 (office)