Setting of T-filter: C0=-1 , other coefficients are 0.
TEST | BERT |
FRAMING | SYNC |
SIMULATE | DTE |
TX RATE | 1200 (BPS) |
RX RATE | EXT |
DEV | 0 % |
PATTERN | 511 PRBS |
CODE | HEX |
BITS/CHAR | 8 |
RCVD TIME | 30 S |
DT-10 gives PRBS-signal only, if we plug the DCD and CTS controlling wires to IN state (V+) on the keypad. Clock signal can access the record to the eye pattern through TTC interface wire on the keypad. We should connect to the oscilloscope with a measuring head!
TEST | BERT |
FRAMING | ASYNC !!!!!!!!! |
SIMULATE | DTE |
TX RATE | 1200 (BAUD) |
RX RATE | 1200 (BAUD) |
DEV | 0 % |
PATTERN | 511 PRBS |
CODE | HEX |
BITS/CHAR | 8 |
PARITY | NONE |
STOP BITS | 2 |
RCVD TIME | 30 S |
Use the EP2-t in measurement modes GAIN and GROUP_DELAY. The measuring signal should be MTTS!
Plug out the 600-Hz low-pass filter from the measuring arrangement, according to Figure M4. Setting of T-filter: C0=-1 , and the other coefficients are 0. After this:
Plug the 600-Hz low-pass filter to the measuring arrangement, according to Figure M4. After it make the measurements according to Task 2. Compare the obtained results with the results of Task 2.
Draw to impulse response the recorded result of the distortionless case! Slipping in time, and multiplication by a constant is allowed for a correct comparison.
Make the linear distortion caused by the 600-Hz low-pass filter with the help of the impulse response, with the iteration reviewed at the design in the time domai.
Set c0 = 1 on the transversal filter, and the other coefficients must be 0. After it, change the value of a c1, c-1, c-2, c-2 ... coefficients in equal steps (0.1), and observe, if the impulse reaction belonging to the distorted case is near to the reaction function of the distortionless case, or it is far from this, or does not have any influence on it! If it is far, then change the sign of the coefficient, and in the third case do not modify the coefficient When we got to the end of the filter, we can begin a new cycle, and if it is necessary, we can reduce the size of the steps. If we do not impair it, after 3 cycles the two reaction functions have to mask each other.
We should verirfy the equalization with bit error rate measurement!
Realize an "ideal" low-pass filter of 600-Hz boundary frequency with a linear phase characteristic with coefficients determined with a Fourier-series method, and then with Hamming-modified coefficients. Measure the amplitude and running time characteristic of both filters!