NBTVmk3

NBTV Converter for Oscilloscope Mk3 (digital ?)

Here is another NBTV to XYZ decoder design, which uses only 3 chips and a handful of discrete components.

Incoming NBTV (audio) is first buffered/isolated through a 1:18 ratio step-up audio transformer.
I used a 5000:15 Ω speaker transformer (in reverse).
The sync-seperator is an LM339 (quad comparator), which is not an op-amp (as previously utilized)
but a correct long-tailed pair, comparator circuit, the advantage being that:
input common-mode voltage range can go right down to ground potential.
The comparator output changes high > low when a sync pulse occurs.
The comparator input signal is "DC-restored" with a shottky diode & capacitor which
conducts on the bottom tip of each line sync pulse and 'clamps' it to a level of -0.2 V

An adjustable voltage divider pot' allows setting of the voltage trip point for the negative going "sync tips"
Sync-threshold pot' is adjusted to give the best picture.
LM339 output is an open collector transistor, and is used to (line-sync) discharge a simple RC ramp network.
The simple series RC network is the line scan oscillator (Y input to CRO)

Also a CMOS 4040 (5 stage ÷ by 32) ripple counter counts off 32 line sync pulses
and then generates a high output, used for field scan oscillator reset (retrace).
After 32 line sync pulses have "rippled" thru the 4040 counter, output Q6 (÷64) switches high.
Q6 is wired to reset (pin 11) which then returns all 4040 counter o/p's to low state.
Also Q6 high; rapidly discharges the 1uF field scan timing capacitor, giving field scan retrace

Although the missing sync pulse (end of line 32), signifying (field) retrace is not actually detected;
because of this gap in the pulse stream the ripple counter will eventually 'sync' into this gap,
even though it could take up to 31, 80mS duration fields to do so.
Worst case settling time would be 31 x 80mS or 2.48 second.

NBTV picture "floats" across CRO screen (field by field) until 'lock' occurs. (see video)
Transistor/bias-LED diode constant-current charging of 1uF capacitor
gives linear sawtooth raster scan waveform (X input to CRO)
The voltage ramps down (high to low) to allow correct field scan from right to left on CRO screen (see waveforms).
which is in accordance with NBTV association specification:

32 lines raster scan with a field repetition rate of 12 ½ per second.
This means each line lasts for 2.5 mS and the 32 line picture is completed in (32 x 2.5) 80 mS
Line scanning is vertical, from bottom to top, starting bottom right.
Field scanning is from right to left. Video modulation is positive going with negative going sync.
Each picture line (except the last of 32) ends with a line sync pulse of around 100-250uS.
End of a field scan is defined by the absence of a line sync pulse, or 'missing sync' (line 32).

Screenshot: 'Audacity' analysis of NBTV audio stream.

The luminance (Z mod') part of signal processing remains unchanged , as in previous designs.
The composite video signal is fed to a times 20 gain (26dB) inverting op-amp for use as Z mod input to the CRO.
I found I needed an LM741 op-amp running on split plus and minus 15 volt supplies
in order to get sufficient voltage swing (up to 30V) needed to drive my oscilloscope (1960's valve type)
Other more modern CRO's may not need the same magnitude of voltage "drive"

Test oscillator provides stable sync' pulses to test decoder.