A brief update to this.
I've sought advice from a friend who is an ex BBC/Thames broadcast engineer, and have been looking to see if there is anything amiss with the teletext generation side, or the arrival of the signal into the adapter. He does not feel the 'Teletext Eye' will be the problem.
The below attachment shows a side-by-side comparison and clearly demonstrates the problem. It shows a TVs internal decoder handling and presenting the decoded teletext signal correctly and the BBC B adapters decoder, mashing it up!
This attachment is the non demodulated teletext line captured direct from the generator. This was compared to the demodulated signal at pin16 of the SAA5030 to see if it was close to the generated signal. The signal amplitude was expected to be less at this point than when captured direct from the source, and it was. Further testing is being conducted.
I'm currently awaiting a PCB kindly offered by a fellow Stardot member, once that is here, it will allow me to continue further investigations and perform comparision checks.
I've sought advice from a friend who is an ex BBC/Thames broadcast engineer, and have been looking to see if there is anything amiss with the teletext generation side, or the arrival of the signal into the adapter. He does not feel the 'Teletext Eye' will be the problem.
The below attachment shows a side-by-side comparison and clearly demonstrates the problem. It shows a TVs internal decoder handling and presenting the decoded teletext signal correctly and the BBC B adapters decoder, mashing it up!
Under his advice, I used my I scope to capture a line of Teletext (non demodulated direct from my Teletext generator) waveform to see if it looks correct. He agreed it does appear to be OK. It suggests there is possibly a problem within the adapter.I assume the BBC B is doing the final decode and display of the Teletext from software. Over the years I have had problems with the order of the control codes that are sent like "Block" then "Background Colour" then "Foreground Colour" and then characters. The last problem I had was with subtitles and cutting out the black box to put the subtitles in. The box codes can be sent after the text, usually done before. It is not incorrect to do it either way, but it did upset some decoders, so had to issue instructions to the subtitle providers to issue the commands in a different order. The reason I mention this is, as well as your top of the picture colour block, your text does not look correct. The foreground colour instruction looks to be white and then changes to light blue. The lines (columns) are not starting in the same place, like some control code is being incorrectly read?
This attachment is the non demodulated teletext line captured direct from the generator. This was compared to the demodulated signal at pin16 of the SAA5030 to see if it was close to the generated signal. The signal amplitude was expected to be less at this point than when captured direct from the source, and it was. Further testing is being conducted.
I'm currently awaiting a PCB kindly offered by a fellow Stardot member, once that is here, it will allow me to continue further investigations and perform comparision checks.
Statistics: Posted by crustytv — Tue Oct 29, 2024 5:25 am