Q1
a)
There are two filters in a Super-Heterodyne filter:
- RF filter
- IF filter
The first filter (RF) is a general bandpass filter that filters out the majority of frequencies that are irrelevant to the system.
The second filter, the IF filter is for further narrowing in on a frequency, specifically the frequencies to be outputted, after the signal gets added to the Local Oscillator
b)
This method is called convolution
typically it the output is defined as
Where is the filter response in the frequency domain
c)
Asynchronous FSK relies on the receiver passing the signal through seperate bandpass filters and then running envelop detection on both signal, before comparing the outputs of the detectors and determining whether a signal is high or low This method distinctly has the benefits of being easier to make and cheaper due to a local oscillator as well as not needing to know the transmitter frequency in the production of the receiver
Synchronous FSK passes the signal into seperate local oscillators which then get passed through bandpass filters before having a decision made on whether the signal is high or low This has the distinct benefit of being more “accurate” in that you can use frequencies that are much closer together for your FSK
d)
Nyquist’s law state that we must have our sampling frequency be at least double our highest frequency in our signal
e)
As stated before
Which means if we are sampling too slow we risk either losing all the higher frequency components of our signal, or getting aliasing in our sampled data
f)
Using synchronous receivers we can have local oscillators with the same frequency but different phases, allowing us to not just differentiate amplitude (Quadrature) in the real axis, but phase shift (In phase), for data modulation. PSK for example doesn’t use any form of amplitude/quadrature modulation, but modulates a signal between a certain frequency and another frequency exactly a phase shift from the first, QPSK takes this idea further, combining both in phase and quadrature modulation to create a binary modulation in either phase or quadrature QAM takes this idea to the logical extreme and modulates data based on the position on the In Phase and Quadrature constellation
Q2
a)
The advantages of optical fibre systems are, generally higher SNR, much higher data throughput, lower latency
The disadvantages are the costs of switches, laying new fibre cable and complexity of switching systems
b)
Passive Optical Network is a concept where instead of using add drop multiplexing with expensive switches for every household, we have a single node on our metro access network where we transmit the data for multiple households (typically up to 32) on one fiber which proceeds to get split into 32 seperate segments that then go onto each household
Router→ONT→OLT→Metro Access → Metro Aggregation → Regional Metro → Backhaul → Core TODO: Diagram
c)
Typically the best networking protocal to use for PONs is ethernet, as each “block” of data has a header that specificies both an IP address (a specific router/ONT), and a MAC address (a specific device), which means you can transmit a lot of data for multiple users on the same line without having to worry about data going to the wrong person
d)
As our channel capacity for fibre can be defined as :
Our capacity and therefore our data throughput is limited by both the SNR of the system which is limited by general noise, as well as the bandwidth which is capped by the nonlinear shannon limit.
e)
Factors that could lead to transmission in both types of fibre are noise, which obvious at some point can cause bit errors, as well as scattering from impurities in the glass which while it generally affects the signal power, can also lead to loss in transmission, which can cause errors.
f)
For interconnects, as well as for fibres at this length, we would use Small Form-factor Pluggables or SFPs, more specifically for this level of data throughput we would most likely be using QSFPs (Quad SFPs) over a multimode fibre cable (relevant based on distance)