why optical link can not transmit data with long strings of 0s or 1s?
it says, this low " transition density" will cause 1. ac coupling 2. offset cancellation 3. clock recovery
I can understand " Clock recovery". That is because the "data and clock recovery circuit" can not recover the clock from this low transition density data.
therefore, without the clock, the data can not be correct sampled. But this problem can be solved by a synchronous clock, am I right?
But what does " ac coupling" and " offset cancellation" mean? or someone introduce some book explain this problem in detail.
I want to know the difficulties about receiving this "low transition density" just Because due to some reason, I do want transmit the signal like this...
by the way, I am doing free space optical link... not fiber.
Please read a basic textbook on data transmission! The answer is easy- an optical fiber cannot carry DC! You can transmit long series of zeros or ones only through a metal cable, two wires or coax, but not through a medium that cannot carry DC! If you fail to understand what "ac coupling" means, take the basic-school course of physics, please.
Fiber optics can carry "DC" just fine, but often the receivers must be AC coupled in order to maintain high bandwidth. Low bandwidth links can in fact work with indefinite lengths of 0s or 1s (I've implemented UART over fiber a number of times, up to 3Mbaud).
AC coupling basically means the signal is put through a high pass filter, and the low frequency part is lost. So if the average value of your signal changes abruptly, the receiver output will "drift" for a while until it reaches steady state. So things like bit stuffing, preambles, and manchester encoding help ensure that this doesn't happen severely enough to cause data to be missed.
Yes, I agree, fiberoptics can easily send a string of zeros or ones. But you have to do it with the proper hardware. If you are using a cheap AC coupled ON/OFF led or laser, then obviously after a while you will lose the DC level.
If you have a better system, like a phase modulated laser, you have absolutely no trouble.
Also, if you manchester encode the data, or use a bit stuffer, you also would have no problem.
... i did not make it clear, in my application , it is a free space optical link, not a fiber link.
so , in that case, will this DC still cause problem?
---------- Post added at 20:17 ---------- Previous post was at 20:15 ----------
my input is 50Mbps, But the signal is DC balanced .
where is this high pass filter usually put ? after TIA ?
---------- Post added at 20:21 ---------- Previous post was at 20:17 ----------
can you explain this in detail ? why the dc will be lost ? or why I should care about this DC, since there is a HP filter in the receiver
and in my application , I can not use mancherster or 8B10B code to limit this long run....