Oscillator help

[dominicbeesley]

Hello all,

Over the past couple of weeks I've been trying to simplify the design of my colour televisor decoder circuitry to make it more reliable and easier and cheaper to build. I have to say I'm stuck!

The trickiest part is the oscillator - this needs to produce two 15kHz square waves 90 degrees apart. It also needs to be stable to withing a few Hz either side of 15k and to be "pullable" either side of this by roughly +/- 6Hz given a control voltage of 0..12V.

In the current design this is acheived with a fairly conventional 6MHz VCXO beloved of normal tellys and radio amateurs this is pulled using a varicap (or a big LED!) and divided down to 60kHz by two 4 bit counters and then a few flip-flops make nice perfect square waves at 15kHz.

What I'd really like is something running at 15kHz without the need for dividers, preferably analogue, preferably realisable with 1930's valves - tall order probably but I know early colour TVs managed to do this.

So far I've tried:

60kHz "watch" crystal - these are very stable but can't be pulled easily they're designed to be bang on spec and not change - anyone know of a way of extending their pull range, so far I've only managed 1-2Hz

4066 - much beloved PLL's internal VCO - not a chance if left free running it floats around 15kHz but won't hold a note long enough...or am I not building it right. I can't get this to stay within 100Hz of 15k

Op-amp phase-shift oscillator - the best I've managed here is something that will stay close to the right frequency so long as it is kept at exactly the right temperature.

Breathing on it sends it way off into the distance. I'm using pretty standard film capacitors and carbon resistors though - would it be worth trying better components or would I just be throwing good money after bad?

I've an idea to use large inductors in this instead of resistors....would that work? Would they be any more stable?

Any ideas - no matter how outlandish - considered.

Dom

[Klaas Robers]

An LC (inductor and capacitor) oscillator is much more stable than an RC oscillator. Realise yourself that in the past radio transmiters were directly dependent on the stability of an LC circuit. The type of oscillators are called: Colpitts, Hartley, Clapp. Look into old radio magazines and -books and you will find the diagrams, including the radio tubes.

Realise yourself that inductors are made of copper wire and they expand when they get hot. So the frequency will go down when the coil heats up. Heating up is more likely in the neigborhood of a burning radio lamp then in the neighborhood of a cold transistor...... But you might compensate this by using some capacitors with a negative temperature coefficient. Also old and well known technology.

[dominicbeesley]

Thanks Klaas,

the thing that has put me off LC oscillators so far is that it needs quite a big inductor for 15kHz but I suppose I should just give it a go!

I'd also like to try and keep it all made out of readily available off the shelf components and inductors formers are getting more difficult to get hold of but not yet impossible...

I've got a load of old TOKO type trandformers and will have a go at rewinding some tomorrow....

The other idea I've been mulling is some kind of mechanical resonator but not really sure where to start....

Cheers

Dom

[Steve Anderson]

Dom,

Oscillators of the variable kind are not my strong point, but one of our members, Hans, has a page on his web-site devoted to Huff & Puff stabilized VFOs. In one he's using an 8kHz (yes 8kHz) crystal which gets divided down to 10Hz for some form of timebase.

Maybe this could be adapted for your own uses...just a thought...link here...

http://www.hanssummers.com/huffpuff.html

A WIP version uses tubes/valves too...

Steve A.

(WIP = Work In Progress)

[Alan]

Dom

For an inductor, try a toroid.
You can find a calculator for the number of turns vs. material at
http://www.66pacific.com

Regards,

Alan

[dominicbeesley]

Thanks lads,

Yes thanks, I've used Hans's site for loads of tips! Though he seems not to be answering any emails - or at least not my dim questions about my own attempt at a spectrum analyser!

Huff and puff is probably no good for my needs - I need to sync to the NTSC carrier signal but only get a brief burst of 1ms carrier every 80ms to compare with - so if the oscillator is out by anything more than 6.5Hz to start with it will lock to a sideband of the carrier burst i.e. anything at 15kHz +/- n*12.5Hz. Huff and Puff seems to work by always having a good carrier to lock to albeit divided down.

I'd love one of these 8kHz crystals though, even better if I could get a 60kHz one!

Alan, I've been playing with torroids but for this I'll probably need something adjustable ... I will be using them to make filters though. However one of Hans's VCXOs works by having a torroid next to a relay magnet to adjust the inductance...another idea to look at!

Anyway hopefully I should get some play-time tomorrow so I'll try some inductors out...

Cheers

Dom

[Steve Anderson]

...I've used Hans's site for loads of tips! Though he seems not to be answering any emails....Dom

I know that Hans is up to his neck in other non-NBTV matters at the moment, he has very little time for Amateur Radio, his web-site or e-mails right now. It's really best to be patient, in a few weeks things for him may be back to some semblance of normality...such is life!

Steve A.

[dominicbeesley]

Smart, it wasn't a crticism though! Anyway his site is probably the best I've found yet on the net along with vintage-radio, you guys and http://www.sm0vpo.com/ I've learnt more in a few years than I ever did when studying properly!

D