reduce ISI in Rx equalizer

[dinosaur078]

Dear every one

i do not understand why ?

"With Linear Equalizer in the receiver to compensate the channel bandwidth,both the pre-cursor lSI and post-cursor lSI can be reduced"

could you please explain ?

thank all you in advance

 

[zorro]

Hi, Dino

Inter-Simbol Interference is due to samples of the pulse before and after the main sample:
before -> pre-cursors
after -> post-cursors

Linear Equalization can reduce both of them.
Conversely, Decision Feedback Equalization is good only for post-cursors.

Regards

Z

 

[dinosaur078]

thank zorro for your words

can we reduce pre-curors ?
and why DFE only can reduce post-cursors?

thank you

 

[zorro]

You're welcome.

can we reduce pre-curors ?

Yes, with a linear equalizer, as stated in your quote.

and why DFE only can reduce post-cursors?

DFE makes a symbol decision (that's the nonlinear process) before generate (and compensate for) the IIS that originated in it. This can affect only subsequent decisions; it's too late for previous decisions, i.e. for symbols already detected. The previous symbols were affected by precursors of the current one, and the subsequent ones are affected by postcursors.
Is it clear?
Regards

Z

 

[dinosaur078]

thank you for your explaination

btw: could you tell me how can calculate equalization coefficient from S-parameter matrix ?

thank you in advance

 

[zorro]

Equalizers are frequently adaptive, so the coefficients ("tap" values) are calculated automatically by the algorithm.
They depends of the impulse response of the total channel. They are not directly related with the S-parameter matrix of a particular part of the chain. In what S-parameter matrix are you thinking?

Z

 

[dinosaur078]

by the way
could y explain about "preshoot" , what is it, why is it used in Pcie ?

thank y in advance

 

[zorro]

Preshoot and De-emphasis in PCIe are effects of the pre-equalizer filter in the transmitter, intended to compensate the channel response.
Regards

Z