The Great Calibration Run of 2024-11-09

[mungewell]

I ran two Pico units against my Sync-IO (clocked by a Sigma BGS26N black burst generator), firstly to calibrate and then to free-run (with average calibration) for over 16 hours.

Both units are 'improved' with a TCXO, but temperature in room was not controlled and has some change overnight.

After the test the units where very close to the reference Timecode:

• Unit T002 - ttyACM0 = 0.02911544 frames fast
• Unit T004 - ttyACM1 = 0.03665596 frames fast

Annotated summaries:

There is obviously some 'weirdness' in the later part of the test. But noting that this did not occur at the same time for the units...

[mungewell]

I was also running my UltraSync One in parallel; it (like the Pico's) started in phase with the reference, but by the end of the test was significantly 'worse' - over one frame fast.

The 1st Evertz is showing info 'black on white' (smaller font) confirming that SyncIO's LTC output matches the VITC ouput.
The 2nd Evertz is showing info 'white on black' (larger font) confirming that UUT's LTC output does not align with the VITC ouput.

I chose this image grab, with bad sync as it shows that the received TC value changes during the interlaced frame - I am not sure how the Evertz actually handles this...

Note: the top 5300 has little arrow pointing to the right.... Timecode is way off!

It is impossible to know whether the calibration of UltaSync One or the BGS26N is the 'bad' one; but we can say that the Pico-Timecode units tracked the one that they were calibrated against.

[mungewell]

With the 'zoom' function turned on, the bargraph shows that both units are 'ahead' of the reference time (ie the SyncIO is reported as 'behind'). The TC values on the OLED do NOT match the real situation, perhaps related to the 'weirdness' seen in the plot, something that needs more investigations.

[mungewell]

For reference the calibration plots for Units were: