The test result sheet can be downloaded here:
Notes:
We achieve 66dB A-weighted signal to noise ratio on the video, provided we push the output by 3dB. Putting it on 0dB means we still have a very respectable 63dB SNR!
Many of the measurements are very close to 0.0%. They did change a bit when using a proper terminator, most values went closer to 0.0%. I instead connected a Sony Trinitron monitor, which obviously has some influence. The differences are, however, small and negligible.
When switching OFF the 5MHz lowpass filter, all values went a bit closer to the reference values, except the 5.8MHz signal which was down 2.2%. This is due to the SinX/X sampling roll-off (and probably some analogue filtering as well).
The C/L delay is odd, as the TPG20 should be close to 0ns on a direct connection. This could be a setting, I will investigate it at some point in time. The delta measurement, however, clearly shows the influence given by the analogue anti-aliasing filters on the analogue in & outputs. This too is negligible.
The test was conducted without the Rec. ITU-R F.405-1 PAL pre-emphasis. This digital filter was simulated to be flat within +/- 0.1dB, and due to it’s nature there is no phase delay penalty, so it will not affect the C/L Delay measurement.
Conclusion:
All parameters taken into account, the digital baseband is well within broadcast specifications, given that we filter at 5MHz on purpose to ensure we do not cause interference with Nicam at 5.85MHz.