Questions on Stressed RX Test E-O and E-O-E Linearity

1
Questions on Stressed RX Test E-O and E-O-E
Linearity

• Lew Aronson Finisar
• Abhijit Shanbhag - Scintera
• August 24, 2004

2
Question 1 on TP3 Linearity

• With the ISI generator before E-to-O, there will
  be intramodal dispersion which will lead to
  cross-products and non-linear ISI after the PIN
  (due to square-law, similar as with chromatic
  dispersion), which the EDC may not be intended
  for.
• Not an issue for TP3 test since there is a very
  short fiber between source and DUT (I.e. no
  significant optical intramodal dispersion in TP3
  test)
• Believe references exist that E-O-E process
  turns out to be linear. Cross products average
  out.

3
Question 2 on TP3 Linearity

• With No Fiber, I(t) is the data-modulated output
  of the ISI generator and h(t) is the laser filter
  function (assumed to be perfectly linear), the
  o/p would be sqrt(I(t))h(t)). The PD output is
  then (sqrt(I(t))h(t))2. The EDC is designed
  to compensate for channel of the form
  I(t)(sqr(p(t))), where p(t) is the cumulative
  (optical) channel impulse response.
• Potential Issue with Low Bandwidth or Non-Linear
  Source
• Can be Controlled by Signal Characterization Test
  Portion of Stressed RX Test Definition.
• Test Signal for TP3 Stressed RX Test Should be
  Specified and Measured Optically into a Linear
  Reference Receiver (e.g. with isolated one
  pattern).
• Mismatch between electrical ISI generator and
  received ISI can be detected and corrected by
  either by using better E-O source or compensation
  of electrical ISI in order to achieve the
  required ISI.