Gaussian filter
Gaussian filters are typically used for MSK and FSK signals. Typically MSK
transmission performs all filtering in the transmitter.
++
g(t)=h(t)∗rect
t
T
h(t)=
e
−t
2
2 σ
2
T
2
√2 πσT
σ=
√ln(2)
2 π BT
rect
t
T
=
1
T
when
t
<
T
2
0 otherwise
Low pass filter (for FSK)
FSK measurements often require spans which are much larger than the symbol rate
because of large deviations inherent in many FSK signals. The wide span may
result in a noisy demodulated signal. The low pass filter is a type of narrow,
gaussian filter with a cutoff frequency greater than the symbol rate. Applying the
low pass filter to the measured signal may smooth the result.
User defined filters
You may define your own filter by creating an impulse response and storing it in a
data register. The impulse response must have 20 points per symbol and have an
overall length of less than or equal to 20 symbols (401 points). Also, the impulse
response must be real and should have an odd number of points so that the center
of the impulse is on a symbol.
Accuracy of user defined filters is undefined. The alpha and BT menu choice is not
applicable because user defined filters includes alpha/BT as part of the definition.
You must select reference filtering based on the purpose and type of user defined
filter. In some circumstances you may have to create a special user defined filter
for the reference signal.
Digital Demodulation Concepts (Opt. AYA)
22 - 18
++ Where T is the symbol interval