The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> High-Speed I/O Design >> Precursor ISI cancellation
https://designers-guide.org/forum/YaBB.pl?num=1363445375

Message started by raja.cedt on Mar 16^th, 2013, 7:49am

Title: Precursor ISI cancellation
Post by raja.cedt on Mar 16^th, 2013, 7:49am

Hello everyone,
Can any one tell me is Linear Equalizer(CTLE) in RX side can cancel pre-cursor?? I have seen mixed answers in the web saying both yes and no.

http://www.ieee802.org/3/100GNGOPTX/public/jan12/szczepanek_02_0112_NG100GOPTX.pdf slide9

http://www.altera.com/literature/wp/wp-01130-stxv-transceiver.pdf 10th page ctle introduction 3rd line

Thanks,
Raj.

Title: Re: Precursor ISI cancellation
Post by ywguo on Apr 2^nd, 2013, 3:28am

Hi Raj,

I don't know why it is said that CTLE cannot equalize precursor ISI in INPHI slides. Probably in their specific project, that CTLE is not enough. Anyway, I think it does equalize precursor ISI because a high-pass filter makes the the rising edge more charp in time domain.

And in slide 14 of the link below
http://www.ece.tamu.edu/~spalermo/ecen689/lecture8_ee720_rxeq.pdf,
the 6Gb/s input opens eye with a CTLE. Obviously it equalize precursor ISI because the edge of input data becomes more sharp, otherwise the eye is closed.

Best Regards,
Yawei

Title: Re: Precursor ISI cancellation
Post by bharat on Mar 23^rd, 2015, 6:21am

why is post cursor ISI is always more than pre cursor ISI?
or
why the dispersion of impulse response of the channel is not equal on both sides.
The PCIGen3 specifications for Pre-shoot is 6dB and De-emphasis is 9.5dB, means post cursor correction is more than pre cursor correction.

Thanks
Bharat

Title: Re: Precursor ISI cancellation
Post by BackerShu on Apr 8^th, 2015, 7:46am

@bharat, Interesting question about why there is usually more postcursor than precursor.

First of all, I think mathematically it is always possible to come up a channel which has more precursor. But for a physical cable, I guess one qualitative explanation can be as following:

There are two main loss mechanism in a cable channel: skin effect and dielectric loss. The first one causes only post cursor and the second one causes almost symmetric precursor and post cursor. The overall loss is convolution of the two losses. Therefore, qualitatively, more postcursor exists in physical cable.
A good explanation about these two loss mechanism is covered in this paper:
http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=1610643

Of course, this explanation is still hand-waving. Hope some expert can give more intuitive answer.