Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: [802.3_NGAUTO] Insertion loss data at temperature

Hi, I sent the channel data to George directly for further analysis. I was not able to find a curve that fits the measured data very well only using c1 and c2. I am looking forward to the outcome. Regards Thomas

2018-05-25 20:46 GMT+02:00 George Zimmerman <george@xxxxxxxxxxxxxxxxxxxx>:
All - I was looking at the data Thomas presented in rosemont ( slides 3 and 4, and notice that the high frequency crossover of the insertion loss limit and the measurement is at about 1 GHz. The issue appears to not be a flat offset but that the shape of the curve is incorrect.  From Thomas’ curves it looks like the insertion loss curve predicts about 2dB more loss than even the high temp curve at 2.5 GHz. 

I believe we need a new curve fit. (I’m volunteering to do the work) Can you provide the insertion loss data versus frequency at temperature?

If we use the existing curve, out insertion loss limit has a steeper slope vs. frequency than in should. If we just added an offset (say x dB, either 0.5 or 1) to fix it below 1GHz (almost mid-band for some of the 10g modulations), we are asking the phy to tolerate approx. 2+x dB more loss around Nyquist for most of the modulations we are considering.  Not that this is a big deal, but it will make a measurable difference in margins.

George A. Zimmerman, Ph.D.
CME Consulting, Inc.
Experts in PHYsical Layer Communications

To unsubscribe from the STDS-802-3-NGAUTO list, click the following link:

To unsubscribe from the STDS-802-3-NGAUTO list, click the following link: