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

Re: [802.3_10SPE] Return loss

Thanks for the correction - for some crazy reason I was using 25-5log(f).

As far as "is there anything special about cg" - yes, there is.
Every PHY has return loss requirements related to its own specs.  CG will put more energy into the low frequency band that 1000BASE-T1.  In 1000BASE-T1, only 10 MHz out of  a Nyquist bandwidth of 375 MHz (about 3%) of the power is contained in the range from 10-20 MHz.   In 10BASE-T1S, as discussed with DME coding, something closer to half the transmit power is impacted by this decreased return loss.  Decreasing the required return loss will result in higher level echo and that will increase PHY front end and echo canceller requirements. 

The change you propose is not without consequences. Proponents need to look at the impact of relaxing the return loss on the PHY complexity, performance as well as on how the new requirement fits with the installed base.  This should not to be considered casually for the purpose of 'harmonization'.


-----Original Message-----
From: Dieter Schicketanz [mailto:dieter_schicketanz@xxxxxxxxxxx] 
Sent: Wednesday, October 25, 2017 11:31 AM
To: STDS-802-3-10SPE@xxxxxxxxxxxxxxxxx
Subject: [802.3_10SPE] Return loss

24-5log(f) crosses the 19 dB at 10 MHz.

Without the slope cg has the highest value in return loss compared to Bq Bp and all cabling standards at 20 MHz. The other ones you mentioned without slope, have much lower values at 20 MHz.

Is there anything special with cg at 20 MHz? I do not think so, therefore it should be aligned with the other high one e.g. 1000BASE-T-1 Dieter Schicketanz

Mit freundlichem Gruß
Best regards
Dieter Schicketanz

Dr. Dieter Schicketanz
Consultant / Berater
82194 Gröbenzell
T:0049 (0) 8142 60139
C:0049 (0) 151 23474821