This time the result will be a flat frequency response, as shown in fig. 3.24. By using a
weighted frequency response with the inverse characteristic of the filter under test we
have measured the deviation from the ideal response and the cursor can now be moved to
any frequency to read the frequency response error. This gives more accurate results than
a flat sweep as it was measured over a small dynamic range with no steep slopes.
The CCIR 468-4 weighting curve specification shown in fig. G.2 (appendix G) demands
that the response at 16kHz should be -11.7dB within ±1.6dB. The display in fig. 3.23
would suggest that the 16kHz response is +0.86dB high, but we know that the accuracy is
dubious, and we haven’t checked the 6.3kHz level which should be used as a reference
level (since its tolerance is 0). Using the weighted sweep (fig. 3.24) allows us to read the
error directly from the graph. By moving the cursor to the 6.3kHz frequency and setting a
test level by pressing
twice we can examine the errors at other frequencies relative
to the assumed 0dB error at 6.3kHz, as required by the specification. As fig. 3.21 shows
the actual error was +0.50dB (less than the +0.86dB measured earlier) and well within the
allowed ±1.6dB tolerance. A manual test using steady tones gives exactly the same
+0.50dB value.
Most filters can be tested in this way. Enter the filter’s response into the LA101
weighting editor and press
to obtain the inverse response. Select the user
weighting in manual mode, by pressing
and run a sequence containing a suitable
sweep segment. Note that the LA102 does not need any special settings (the operations in
the above example were to force it to test its own internal filter for demonstration
purposes − see section 3.24).
Some care is needed with filters which have very deep notches. For example, if a filter
rejects 1kHz by 90dB, then this technique would require a level of +90dBu in order to
obtain a 0dBu output, and this is clearly impractical. In such cases the weighting should
be modified so that the system is not overloaded and the results around that frequency
should be ignored. Weightings can also be switched on from within a sequence definition
- see section 4.29.
3.20 Automatic Results Storage Sequence results − automatic storage
The LA102 can be configured to automatically store sequence results in the next available
memory by setting configuration Y. This is provided for receiving several sequences at a
remote site without user intervention and is especially useful for receiving overnight test
transmissions. It can also be useful where several sets of sequence results need to be
recorded for later examination or simply for logging the last few tests for reference. The
LA102 can be configured to power up in sequence mode, and automatic results printing is
also possible (see section 3.4).
Setting configuration Y1 (
AUTO STORE SEQ RESULTS, IN MEMORY 1 NEXT) will
cause the sequence results to be stored in memory 1 immediately they are received.
59
3. Sequence Testing
To Next Page
Loading...