General crystal/PLL question

[[email protected]]

In the days of the first color NTSC TVs, a common practice was to drive
the color crystal directly with the incoming colorburst and then let
the crystal coast for the rest of the line.
My questions are...
How exactly did they drive the crystal? Was there something special
about those crystals that let them be driven? (Those cans were HUGE
back then, was it a different material, cut, etc?)
What if I took a modern hc-49 crystal and set it up in the classic
logic inverter oscillator circuit, but instead of an inverter I took a
NOR gate and periodically pulsed the other input of the NOR gate with a
narrow sync signal?
Would the crystal eventually get 'locked' to the phase of the incoming
sync?
TIA

 

[Mark]

My questions are...
How exactly did they drive the crystal?

The color burst signal(burst of 3.58 MHz was applied to the crystal.

Was there something special about those crystals that let >them be driven? (Those cans were HUGE
back then, was it a different material, cut, etc?)

No nothing special.

What if I took a modern hc-49 crystal and set it up in the >classic
logic inverter oscillator circuit, but instead of an inverter I >took a
NOR gate and periodically pulsed the other input of the NOR >gate with a
narrow sync signal?
Would the crystal eventually get 'locked' to the phase of the >incoming
sync?

Not sure what you mean by "sync signal". Yes it will work if you
pulse it with a 3.58 MHz bust., The crystal will "ring" in phase with
the applied burst.

Mark

 

[[email protected]]

Hi Mark,
Thanks for the reply. I was typing out loud that day. I remembered
later that I actually have the books from the 50s about color TV. Turns
out the crystals *were* specially designed to have a lower Q than
typical crystals, and that they were used in a ringing circuit, not a
true oscillator AFAIK. They just whapped it with the colorburst, and
then sent the decaying signal through a limiter, and this was enough to
generate the color reference.

As for the sync, I was referring to a horizontal sync pulse from a
computer. Let's say the computer uses a 16MHz clock and I want to
re-create this pixel clock only having access to the HSYNC. Typically a
PLL system will try to lock a VCO with feedback from a divider,
suitably programmed with the division ratio of 1016 in this case.

I just wanted to avoid using the counter altogether if I could somehow
just use the rising edge of the sync and see if I have a zero-crossing
on my LO. Turns out this is a lot more work that just using a regular
video genlock PLL.

 

[joseph2k]

The color burst signal(burst of 3.58 MHz was applied to the crystal.


No nothing special.


Not sure what you mean by "sync signal". Yes it will work if you
pulse it with a 3.58 MHz bust., The crystal will "ring" in phase with
the applied burst.

Mark

So sorry sirs, the horizontal sync pulse and "back porch", where the color
burst is, was used a "gated" phase comparison in an otherwise standard
phase locked loop, which used a sample and hold and a reactance tube to
tune the "VCXO". (1960's) An early sampling control system if you will.