What determines how "stable" a transmitter is?

[LondonPirate]

I've built a couple of FM exciters from schematics online (just sproggy little VFO 2-transistor units) - these aren't intended to be run through an amp or used for any serious broadcasting, it's more to learn the theory of how a transmitter works and how I can apply that to a future build. With the latest build - using this schematic - I notice it's incredibly sensitive to everything around it. I don't have the circuit housed yet, and all I have to do is approach the board while its fired up and the frequency seems to be shifting - is this normal for a non-PLL circuit?

I understand that PLL will lead to a much more reliable circuit, but I'm guessing what I've built isn't the pinnacle of VFO? What sort of designs were pirates using in the VFO days to ensure some reliability?

[XXL]

That design is terrible, there’s not even a variable inductor on it to change the frequency. Your wasting your time with that. Try just making a stable oscillator first. The oscillator shouldn’t move at all once set with a stable setup.

[Bton-FM]

http://www.gareth.net.nz/nrgworkshop/am ... bility.htm

[Albert H]

Frequency stability of a free-running VHF oscillator is determined by the design of the oscillator, the types of components used, the stability of the supply voltage, the layout and housing of the oscillator, mechanical stability, thermal issues..... Too many things to take into account! There's little point in trying to build stable oscillators these days as PLL control is so cheap and easy to add. The VCO that's steered by a PLL shouldn't be inherently unstable - a poor design still won't perform well. The PLL-controlled oscillator is as stable as the reference crystal (put that in a crystal "oven" if you're really concerned).

[yellowbeard]

Albert has it right there, it'd be quicker to say what is wrong with that oscillator. If you look at the circuit - C5, C6 and L1 for the most part determine the frequency, and the antenna is hooked up to the middle of them and is part of that circuit. Any changes to the antenna will change ITS capacitance and shift the frequency, even walking up to it. A buffer amp and an antenna away from people would solve that problem. A metal box with holes drilled in the lid over the trimmers will help, and packing the box with polystyrene beads gives a bit of protection from temperature variations. Even if you do all this it'll still drift though...

[teckniqs]

I see there is a BC108 on the audio input stage, would this be some kind of audio pre-amplifer for a microphone or something?

...It looks like that's just a little listening bug and they've removed the microphone and changed it to an open audio input on the circuit diagram?

[Krakatoa]

This kind of one transistor transmitter is the typical that's seen on every electronics magazine, blog, etc. and it works... but very poorly.

Look up on google for BF900 FM oscillator, that one is based on a dual gate Mos-Fet and is way much better in stability and cleanliness of the output. I think that it also works with other transistors like the BF961 and 964.

[LondonPirate]

Thanks for the replies everyone - this has been educational for me because I've now learned about the all-important buffer stage!

[Albert H]

The idea of the buffer is to isolate the oscillator from the rest of the circuit. I used to use the buffer as a convenient take-off point for the PLL input - use the signal at the collector of the buffer transistor to go to the next stages in the transmitter (usually the frequency doubler(s)), and feed the emitter signal to the input of the first PLL divider IC through a simple, un-tuned amplifier stage, to get the voltage up a bit so that the logic would switch reliably.

A word of advice: Most of the people trying to hear your station will have receivers that are digitally tuned these days, so if your transmitter drifts, they'll get distorted audio, and will tune elsewhere. In the old days, with analogue-tuned receivers, they usually had "AFC" (to compensate for the drift in the receiver local oscillator) which would allow the receiver to "follow" a drifting rig somewhat!

You really can't afford to drift these days. The band is absolutely packed, and you really don't want to end up jamming a BBC or commercial station for a couple of hundred yards! You MUST use a PLL. The circuitry is simple and well-known. It'll add a couple of quid to the price of your exciter, but will ensure that your little signal stays where you set it! Look at some of the simple two-chip solutions - the SAA1057 is still available for under £1, and a PIC 16F84A is also <£1, and a 4MHz crystal should cost you <50p. You'll need to build a simple PIC programmer (two regulator ICs, a couple of transistors, a couple of LEDs and a USB to RS232 adaptor) - there are plenty of circuits for them around the web, or you can buy a Chinese one for a fiver.