ISI in optical link

[timedate]

Hi, all

I get a question for ISI ( inter symbol interference) in optical link.


In Razavi'S book,<design of integrated circuits for optical communications> P65, Fig4.2 (b)

It says that if the bandwidth of the receiver is 1/4 of the data rate, then voltage decrease due to ISI is roughly 0.2*Vpp.

However, in Saeckinger's Book, < broadband circuits for optical fiber communication>, P59

It says that if the bandwidth of the receiver is 1/3 of the data rate, the voltage decrease due to ISI is 0.5*Vpp.


I was wondering which one is right? Or they are both right and I am wrong.

 

[AdvaRes]

I don't have these books. But, are you sure they are talking about the same receiver ?

 

[timedate]

"the same receiver？" I think so. They all talking about optical receiver for NRZ data.

 

[biff44]

are you talking about the optical link, or just the receiver. I would assume if you tried to turn on/off a laser diode at a data rate 3 times its bandwidth, it would not look pretty

 

[timedate]

are you talking about the optical link, or just the receiver. I would assume if you tried to turn on/off a laser diode at a data rate 3 times its bandwidth, it would not look pretty

just the receiver, assuming the input is perfectly rectangular.

 

[biff44]

So, please enlighten me. Here they are saying that for 10 GBps data rate, you need 7.5 GHz 3dB bandwidth:
https://books.google.com/books?id=y...eiver "3 dB bandwidth" vs "data rate"&f=false

 

[timedate]

So, please enlighten me. Here they are saying that for 10 GBps data rate, you need 7.5 GHz 3dB bandwidth:
https://books.google.com/books?id=y...eiver "3 dB bandwidth" vs "data rate"&f=false

Yes, usually, the bandwidth is 0.7 or 3/4 of the data rate. This is the trade off between ISI and noise.
If your bandwidth is 1/3 of the data rate, for example, 15GHz for 40Gbps. It still works.

I just want to know how does ISI do damage to the signal. this two books give a different answer.