How do you multiply the frequency of an oscillator?

Join the discussion
Registration is free. Start your own thread to ask a follow-up.
38 replies · 12K views
the transistor in the doubler cct will produce some amplification

the cct shown is a pretty standard way of what is used in a multi stage multiplier
when starting with a low freq crystal oscillator.
for the 144MHz ham band the crystal freq was commonly 12MHz tripled to 36MHz doubled to 72MHz and finally doubled again to 144MHz

Dave
 
Engineering news on Phys.org
So for that circuit, you put in a big enough signal to get the transistor to conduct on the + half cycles, and it's cut off for the - half cycle? And that drives the resonant circuit to ring at its tuned 2x frequency? Interesting.
 
That's it. The circuit exploits the transistor's nonlinearity.

Vbe's threshold would reduce the conduction angle below 180°, making the circuit more efficient at producing 2nd and 3rd harmonic, but the transistor's limited speed makes the waveform smoother and less harmonic-rich.

I used it up to the 5th harmonic, but this is exaggerated. The main limit is the selectivity, hence accuracy, of the output filter. A multistage filter gets any desired selectivity, but only if it pick the proper harmonic under any circumstance. As a doubler or tripler, two LC are a good value, then Q=10 uses to be more than enough.

Before chips were available, PLL were too complicated hence uncommon, and this circuit was standard practice. Still nicely simple, and it introduces less phase noise than a PLL.

Symmetric circuits, with two transistors, reduce the odd or even harmonics.
 
The other common way for freq multiplication these days is to use MMIC devices like MAR or ERA etc series

You drive the MMIC hard from the initial osc cct to give a good rich harmonic output
filter this output with a bandpass filter to isolate the wanted harmonic, before going into the next multiplier stage.
A number of amateur transverters use this scheme eg the Downeast Microwave transverters

cheers
Dave
 
Enthalpy said:
That's it. The circuit exploits the transistor's nonlinearity.

Vbe's threshold would reduce the conduction angle below 180°, making the circuit more efficient at producing 2nd and 3rd harmonic, but the transistor's limited speed makes the waveform smoother and less harmonic-rich.

I used it up to the 5th harmonic, but this is exaggerated. The main limit is the selectivity, hence accuracy, of the output filter. A multistage filter gets any desired selectivity, but only if it pick the proper harmonic under any circumstance. As a doubler or tripler, two LC are a good value, then Q=10 uses to be more than enough.

Before chips were available, PLL were too complicated hence uncommon, and this circuit was standard practice. Still nicely simple, and it introduces less phase noise than a PLL.

Symmetric circuits, with two transistors, reduce the odd or even harmonics.

This is sooo cool! So does the transistor's capable running speed have to be at an exact value, or does it just have to be capable of running around that frequency?
 
for low power circuits, the frequency limit of the the transistor is the critical part.
all transistor datasheets show a Ft for the transistor. this is the maximum and generally one would choose a transistor that has a much higher Ft than what their circuit is going to operate at.

Dave
 
davenn said:
for low power circuits, the frequency limit of the the transistor is the critical part.
all transistor datasheets show a Ft for the transistor. this is the maximum and generally one would choose a transistor that has a much higher Ft than what their circuit is going to operate at.

Dave

Yeah, I knew it had to be higher, but does it have to be a certain amount higher other than extra headroom? Doesn't it have to be non-linear or something like that to make harmonics? Or is the LC tank the part that makes the harmonics and the transistor just acts as a buffer?
 
I think most countries allow very low-power FM transmissions for things such as baby monitors, wireless microphones, etc. At a 100MHz operating frequency, you can verify its frequency on your FM broadcast receiver. Example: one transistor transmitter http://www.next.gr/oscillators/rf/RF-oscillator-circuit-2N3904-l6324.html[/color]

First check that this band of operation is permitted in your country.
 
tackyattack said:
Yeah, I knew it had to be higher, but does it have to be a certain amount higher other than extra headroom? Doesn't it have to be non-linear or something like that to make harmonics? Or is the LC tank the part that makes the harmonics and the transistor just acts as a buffer?

The Ft quoted in transistor data sheets is the frequency at which the current gain reaches one.

So, you don't want to be using it at that frequency. At about half of that frequency, the transistor will have substantial gain and will give harmonic output from drive at a lower frequency.

So, for output at 200 MHz, you would need a transistor with an ft of at least 400 MHz.

I have some transistors with an ft of 6000 MHz but these would be capable of output up to 3000 MHz so they have to be used carefully.

For this project, you would need an output of 5 or 10 watts from the transmitter, and this could cause lots of interference unless it is properly designed and tested.