In some systems, it can be a real challenge to find channels to use for proof
of performance tests. A lot of systems have a basic tier which is, of course, available to all subscribers and
nearly everything else is scrambled. A typical procedure for running proofs is to select channels that aren't likely
to offend many subscribers if the scrambler is disabled for the duration of the tests and run the tests the same
as for non-scrambled channels.
A better approach is to leave the scramblers in service and run the tests at the
output of a set-top converter. The set-top box descrambles the signal allowing normal gated measurement techniques.
Remember, C/N has always been required "at the subscriber terminal" and in-channel response will also
be required there starting Dec 30, 1999. Of course the necessary test conditions must be in place (really "quiet"
quiet lines and an appropriate test signal for in channel response). This is a very effective technique. However,
it does require some special considerations.
I found that gated C/N test results on scrambled channels were consistently worse
than for non-scrambled channels (all tests done at the output of a set-top converter). In some cases, in-channel
response was also much worse. After seeing this effect in several systems, using scramblers from different manufacturers,
I began to question whether the degradation was just excessive noise in the scrambler or if there was more to it.
After some experiments (and some very valuable correspondence with a gentleman in New Zealand) it became apparent
that C/N measurements during the VBI always showed worse performance than tests made during "active video",
on line 25 for example. Tests with SA and Jerrold scramblers and set-top boxes showed similar degradation in C/N
tests. In general, C/N measurements in the headend were 4 to 5 dB worse when measured in the Vertical Blanking
Interval (VBI) compared to measurements made during active video. Typical numbers being 47 dB in the VBI and, maybe
51 or 52 dB when measured on line 25. Also, the Jerrold components greatly distorted in-channel response test results
when done in the VBI. The results were fine when performed during active video.
The 4100 and 4200 In-Service Test Processors have the ability to insert quiet
lines and multiburst anywhere between line 10 and line 25. For tests on scrambled channels, I recommend inserting
multiburst on line 24 and a quiet line on line 25. I've used this combination with several systems and with consistently
accurate results. Lines 24 and 25 are generally not visible at the top of the television set display. Technically,
active video begins on line 22. So lines 22 and 23 may seem like a good choice to insert multiburst and a quiet
line respectively. However, some scramblers use those lines. So its best to use line 24 for multiburst and line
25 for a quiet line.