Lew - your first paragraph in the response implies that we want the same
SNR in the test signal as in an actual signal in the application. That's a
bit different than wanting to produce the same power penalty that we expect in
application, which is what I think we are trying to do. The differences are
related to the signal dependent (vs. independent) noise levels and optimum (vs.
centered) thresholding you refer to in your submission.

The simple formula accounts for extinction ratio by its definition of
Q = 1/2 of pk-pk signal / rms noise, or OMA/(2x rms noise). It
requires the assumption that the noise is independent of average power. I
believe your curves will approach the simple equation if you let ER go even
lower.

----- Original Message -----

**Sent:** Wednesday, October 27, 2004 10:08
PM

**Subject:** Re: [10GMMF] TP3 calculation of
OSNR for compliance test

What
we are calculating is the S/N of a signal with relative noise, i.e. different
amounts on each level, with a PP of 0.9dB. The TP3 compliance signal
will then simulate this relative noise with equal noise on each level yielding
the same S/N.

Also, doesn't the simple formula still need to take into account the
extinction ratio?

In
any case, you are probably right that the approximation of equal power is
fine.

Lew

-----Original
Message-----

**From:** Tom Lindsay
[mailto:tlindsay@IEEE.ORG]

**Sent:** Wednesday, October 27, 2004 3:22
PM

**To:** STDS-802-3-10GMMF@listserv.ieee.org

**Subject:** Re:
[10GMMF] TP3 calculation of OSNR for compliance
test

Lew -

Your development assumes that the rms value due to RIN is
proportional to the optical signal power. In the case of TP3 testing, we
plan on adding a white noise generator before the E/O, such that its rms
noise is ~ the same at both levels.

In this case, the math can get much simpler and should revert to the
simple textbook formula PP(dB)=-5*log[1-(Q/Qn)^2], and Qn
=Q/sqrt(1-10^(PP(dB)/-5)).

For 0.9 dB, Qn=12.1. (Compare to your equation (8)).

Comments?

Tom

----- Original Message -----

**Sent:** Tuesday, October 26, 2004 8:45
AM

**Subject:** [10GMMF] TP3 calculation of
OSNR for compliance test

Sorry for the late e-mail, my first attempt got bounced as
too large a file size.

Please find attached a calculation I have
done for the optical S/N we need to have on the TP3 compliance
signal. It also has some calculations of the resulting EDC S/N for
different receive powers and receiver front end
sensitivities.

Lew

-----Original Message-----

From:
Michael Lawton [mailto:mike_lawton@AGILENT.COM]

Sent: Tuesday, October
26, 2004 2:22 AM

To: STDS-802-3-10GMMF@listserv.ieee.org

Subject:
[10GMMF] TP3 meeting Agenda + dial in Oct 26

Dear
TP3'ers,

Here is the agenda for today's meeting:-

Tuesday
October 26th at 9am SJ, 5pm UK, 6pm Germany

Dial in (650) 599-0374,
Meeting ID:
136169

1. List
Attendees

2. Review meeting
notes from last
week

3. Agenda
additions/changes?

4. Static
Channel
Methodology

- Noise loading - review calculations from
Lew

- Channel types and exact
characteristics

o review comments from Tom reference rise and fall
time

5.
Jitter

- Petre kindly agreed to evaluate jitter testing requirements and the need
for a high
frequency

sinusoidal
interferer

6.
Link
Budget

- I think we have agreed that the testing is designed to emulate the
budget

- Do we need to do more work on the
budget?

- Is our thinking on metrics changing? (PIE-L, PIE-D, fixed length
equalisers)

- 300m question

7. OMA Measurement
methodology

- Piers is exploring a histogram vs a slow pattern approach.

I'm
not expecting we work through all the items ... however I want to keep
them on the agenda so we understand our work list and where it makes sense
we can shift around the order.

Talk to you all later.

Best
Regards

MIke