Narrow-Bandwidth Television Association

Today I looked more carefully into the PCB and found a point at the SSTV input where the real board deviated from the circuit diagram. For whatever reason the component values in the input amplifier deviated quite a lot from the values in the circuit diagram. It looks as if this is originaly done this way. This is the part that Chris ACL just constructed. I have to look back to the circuit diagrams that Steve posted long ago.

I also saw that the "programmed" IC's for 625 / 525 lines do not have clipped off pins, but the pins are bent outside the sockets (all IC's are on sockets). So it is possible to redo that.

I don't know if the demodulator in the Robot 400 is the same as that in Robot 70. The diagrams I have and what I eventually modified are for the Robot 70 monitor. So it is possible that the 400's demodulator may be of the same type but with different component values. I'll scan the Robot 70 input/demodulator originals and post them a bit later.

Steve A.

Done. Robot 70 Demod A feeds Robot 70 Demod B. The demodulator output is U5, Pin 6.

Final version (for the moment) of my implementation of the demodulator follows...gif and pdf versions.

acl wrote:Images show complete decoder section built on strip board....

Chris, if the photos are of the completed demodulator...I don't see the two trim-pots.

Steve A.

Now the boring bit...the power supply. It's simple and here's what I'm using. No rocket science. The currents indicated are 'guesstimates' for the MkI version, the 5V (now 3.3V) current was far less than indicated here. The currents include the Robot 70 type demodulator and a NBTV DC-restorer/clamp. So the dissipation figures for each regulator are somewhat high.

As I had no idea at first how much the MkI version would consume I threw two big caps at the power supply (C801/802). These probably could be changed to C801=2200uF/25V, C802=1000uF/25V). I expect the MkII version to consume less than the MkI.

The +/-12V regulators shouldn't need heatsinks, and as implied in the circuit the -12V regulator could easily be a 79L12. The LM317 may require a small heatsink, no harm in fitting one anyway.

Ignore the "Not For Construction" note in this drawing, but only this drawing for the moment.

Steve A.

Hi there Steve,

You are ahead of me there. I was about to ask what the likely load was to be on the 3.3 volt line then suddenly your power supply circuit schematic appeared.

Regards Chris Lewis

Chris, I estimate that the 3.3V supply current to be less than 50mA. But that is an estimate.

The LM317 could be replaced by a 7833 or a 78M33 if you can source them, it saves two resistors - wow! I did consider the LM1117-3.3 but its maximum input voltage is limited to 20V - a bit too close to the unregulated +18V supply in my opinion. Added later - in fact ST only rate the LD1117 (their designation) to 15V - so it's a non-starter.

Steve A.

Steve, I isolated for you the FM demodulator part of the circuit diagram. See below. But I have seen already that the real resistor values are different from those in the diagram. For the first amplifier the resistors from the inputs to ground are 150 ohm and the feed back resistor is 47k.

It looks also that the second amplifier is added later, watch the position on the diagram and the routing of the signal. And the PCB lay-out drawing is different from what is on the board, e.g. an extra opamp that is obviously added later. However my circuit diagram is not part of the other documentation, so it might not be consistent with the lay-out drawing that I have.

Steve, corrected my error and added missing potentiometers. I have ordered bits for power supply unit from Maplins. today

Steve,

Can you explain what R1 seed 47K means I'm a bit slow on the uptake?

Regards Chris Lewis

The 'R1 SEED' is just a note to myself regarding the initial settings for the filter design software, you can ignore it. Likewise a lot of the stuff in red where values or some other attribute hasn't been fixed yet. This is always a problem when others see WIP stuff (Work In Progress). I can't see any major changes for the demodulator in the future at this point - except for its total elimination!

Steve A.

Klaas Robers wrote:But I have seen already that the real resistor values are different from those in the diagram. For the first amplifier the resistors from the inputs to ground are 150 ohm and the feed back resistor is 47k...

The circuit diagram is far more recent(?) than the Robot 70 schematic I have. Using 1458 dual op-amps is a give away. There is a chance that over the duration of the production run the old uA709s were replaced by the 1458s requiring a new pcb layout. Maybe using dual devices they had one spare so they just shoved it in as a second amplifier. I can't see any reason for it otherwise.

The interesting thing is the actual demodulator is the same, including component values, as the Robot 70, with the exception of the 709s/1458s. It saved all those pesky compensation components. It was a proven design to them, why re-invent the wheel?

Keeping documentation current as minor changes are introduced in a production run is a pain. Re-print manuals, keeping track of serial numbers and so on. Many manufacturers don't bother except for major updates. The discrepancies you've noted are minor and don't really change the function or performance of the device.

I do notice that the output filter is different, I'll run a quick simulation and see what they were trying to do.

A little later...well, whatever they were trying to do didn't make any difference. The output filter is -3db at 770Hz and the overshoot/ringing is still about 20%. Not exactly what I would call an improvement.

Steve A.

There is a chance that I may be able to simplify this up-converter even further, eliminating the eight AND gates (2x74AC08s) used for blanking. But that's a back-burner thing, something I'll look at when the current version is running. But I feel a MkIII version may be on the horizon.

Steve A.

Klaas ,

When Steve has finished the hardware/software development how easy would it be to produce a PCB for this section? I suppose how much interest was shown in the group?

Regards Chris

Chris, I don't know yet. Anyway I do only single sided PCBs with a few wire links at the component side. That gives some restrictions. But also for Karen it worked. We'll see when the design is stabilized.

Hmm, PCBs, something I hadn't considered. In my 'real work' I farm it out to a PCB design house. It's amazing to watch these guys work, or more correctly watch the software work. For this up-converter once the schematic, netlist and component data is loaded it takes literally milliseconds to get a result.

This would normally require a double-sided board with plated-through holes/vias, but it could be done single-sided with links. The advantage with a PCB is the possible provision of a ground plane that is hard to implement on strip-board and the like.

I'm quite keen on the idea, but I realise that the quantities are going to be very small which usually makes the process quite expensive per board.

Steve A.