Variable frequency oscillator design for HF

[Radiotech]

I have only used crystal oscillators for my builds, but crystals can be hard to find that match the frequency you want to transmitt on. Is there any good and simple VFO altnernatives for HF using a colpitts oscillator for example?

I am looking for something like the lower example in this image, from the ARRL handbook. Is this something i should go for, or is there better and prooven designs?

[Stretchyman]

The AD9833 is $2 on a PCB wi a CLK from CH. Sure it needs SW to drive it but I can send you that for whatever frequency you want. You'll need to use an attiny45 and a transistor or 2 to amplify the signal but then you can have any frequency you want (1-10MHz) with Xtal stability.

It's an option for you.


Str.

[Teotwaki]

The AD9833 is one easy way to break free from fixed crystals. I have one on the bench and use an Arduino "Nano" board to control it. There's lots of software already available. It's easier to use the square wave output and filter it as the board's  sine wave out is fairly poor quality.

If you want something you can just use without software and a microcontroller there are modules with everything on the board to allow it to act as a VFO. Depends on your skills and budget as well as what you define as simple

[Radiotech]

Looks interesting, i will think about it!

I tested a few crystals i got on line, but they didnt work. Some of them(the one on the right) worked very well, but the one on the left didnt work at all, both the same frequency though. Is there any difference between them that is important(besides the size)?

[Teotwaki]

I don't know how you are testing them but crystals can be designed to operate properly in either a series or parallel resonance mode

[Brian]

The AD9833 is one easy way to break free from fixed crystals. I have one on the bench and use an Arduino "Nano" board to control it. There's lots of software already available. It's easier to use the square wave output and filter it as the board's  sine wave out is fairly poor quality.

If you want something you can just use without software and a microcontroller there are modules with everything on the board to allow it to act as a VFO. Depends on your skills and budget as well as what you define as simple

Thats what I'm using.

[redhat]

The AD9833 is one easy way to break free from fixed crystals. I have one on the bench and use an Arduino "Nano" board to control it. There's lots of software already available. It's easier to use the square wave output and filter it as the board's  sine wave out is fairly poor quality.

If you want something you can just use without software and a microcontroller there are modules with everything on the board to allow it to act as a VFO. Depends on your skills and budget as well as what you define as simple

Thats what I'm using.

I ordered a few last month, and hope to play with them more once I get some rotary encoders (which are not included!)

+-RH

[M R I]

I been working with DDS for years. I highly recommend them.

[Albert H]

I often use a simple CMOS PLL, using the 4060 for the reference oscillator (with a standard crystal, usually 4MHz) and reference divider, a 4040 as a preset divider (programmed with diodes), and a 4046 for the phase comparator, loop filter and Voltage Controlled Oscillator.  The 4046 gives a squarewave output, and the mark / space ratio is somewhat proportional to the frequency, so if I want a really accurate 1 : 1 mark / space ratio, I generate twice the output frequency, and then use a 4013 or 74HC74 bistable to guarantee the waveshape.

The whole CMOS PLL fits on a tiny PCB (really tiny if you go for the Surface Mount versions), uses a standard crystal (I frequently use 4 MHz because I have a boxful of them!) and is easy to set up.  It costs just a couple of £ in parts, and will generate any frequency I want. 

There's an even simpler version for mediumwave - a 40103 preset divider, 4060, a 4046 and a 4013.  The crystal is a 4608 kHz (baud rate generator crystal) which is cheaply available from Farnell / Newark and many other suppliers.  This gives 9kHz steps (for Europe.  If I want 10 kHz steps, I have to use a 5120kHz rock.

My current cheap MW rig uses the 4-chip synthesiser, driving an IRL-series FET (logic level input), and is modulated by a really simple series modulator, using a TDA2040 IC as a "power op-amp".  This basic design is capable of ~20 Watts carrier / 80 Watts peak, and sounds really good on the air. The most expensive parts are the heatsink and the mains power supply.   It's not super-efficient, but it's simple and cheap! 

[Pigmeat]

I used to be in contact with a pirate who only used VFO's for his transmitters about 20 years ago. I heard a couple of them on the air, the frequency was solid and the audio was great. Damned if I can remember the name he used? He was a member of the old "Homebrew Pirate Radio Group" on Yahoo, years ago.

[Albert H]

You'll never get a VFO to be as stable as a crystal or a PLL synthesiser.  With a lot of care and effort, and careful component selection, you can get a VFO reasonably stable, but if the temperature changes , or the relative humidity changes, the VFO frequency will change, or even drift!  If you insist on making the rig "frequency agile", another approach to oscillator stability is the "Huff & Puff" frequency correction circuits.  This is actually a great approach, as long as you have a handy freqency counter!  I've experimented with various FLL (frequency-Locked Loops) over the years, and have got a couple of reliable designs that will certainly keep a transmitter in place for the duration of a broadcast, and they only use a handful of components.....

Many years ago, I experimented with various diode types, trying to find cheap parts that would provide a reasonably wide range of capacitance as the bias voltage was varied.  I found that a reverse-biased green LED was just about as good as an expensive "real" varicap diode!

The "minimal" circuits published in the RSGB's Magazine "Radcom"  and in "Sprat" magazine by Hans Summers are excellent and I can recommend them.  The version I build uses a cheap computer crystal for the reference, and allows tuning of the oscillator in 10 Hz steps.  Once the frequency is selected, the rig will stay where it's put for many hours, irrespective of temperature changes, barometric pressure or moon phase......

If anyone is interested in trying out the "Huff & Puff" approach to frequency control, I'll put some circuits up on here.  I also (sometimes) incorporate a simple PIC-based frequency counter, and I'll also be happy to provide that circuit.

[ThaDood]

As an experiment in the mid-1990´s, I keyed an Icom IC-745 HF rig, on FM mode on the 10W setting, and loaded that to a dummyload. With the RG-58 coax, (Not 100% shielded.), that fed the dummyload, I had a few turns of coiled wire around that coax and fed it directly to the FT-243 socket of a Knight T-60 transmitter, also with its output on a dummyload. That worked well. The IC-745 was the signal source, and the T-60 was the AM modulator. And, very FREQ agile. It was a neat experiment at the time.