Building a 5MHz Oscillator

Dear: Sprit

I am Mr. Sun Sudkaew , Synchronization Engineer of CATTellecom Thailand .

About your requirement : 5 MHz gen (clock source) : can see on

www.oscilloquartz.com : to be selected product by your self or under comment

of osilloquartz's customer services.

Sincerelly Your

Sun 15/1/2008

Hello Sprite,

One of these ICs used with an external 5 MHz crystal and a shunt resistor will give you a very stable TTL or CMOS frequency source. Caveat: The output is a "square" wave, not a sine wave.

This approach doesn't teach you how to design an oscillator, however, but as you seem to be interested in a specific frequency, I'm guessing you're needing a 5 MHz oscillator for some other reason.

Hartley and Colpitts oscillators are fine as long as you can tolerate frequency drift. What you can tolerate depends, of course, on your application. Whenever I need a stable, fixed-frequency oscillator (I almost always do) I use a crystal oscillator.

Why? Some years ago I built a ULF-to-MF frequency converter which up-converted the 0-20 kHz band to the 1-1.02 MHz band so that I could listen to whistlers and other weird phenomena on my shortwave receiver. At first I used a 1 MHz Hartley oscillator for the frequency reference but soon replaced it with a much more stable 1 MHz crystal oscillator. The Hartley's frequency drifted badly with temperature. So badly, in fact, that a temperature change of only a few degrees caused the frequency to drift several hundred hertz! (I could actually tell from the freq change when the converter box was in direct sunlight and when a cloud was passing overhead. The converter itself was mounted out-of-doors inside a weatherproof box at the base of my antenna.)

You wanted some pros and cons of Hartleys and Colpitts. A cheap crystal oscillator outperforms them both for about the same price (cheap) and complexity (simple).

Hope this helps.

-e

More on crystal oscillators...

If you would like to build your own crystal oscillator, a good place to start is the Pierce oscillator, shown here. A FET is used as the gain element. The crystal shown here has a frequency of 2.000 MHz, but a 5.000 MHz crystal may be used without other modifications. The 3.3 mH RF choke presents a high impedance to RF (thus blocking it) whilst having a fairly low DC resistance (41 Ω). At 5 MHz, the choke impedance is about 104 kΩ.

As the crystal is the only resonant element in the circuit, it determines the frequency of oscillation. The feedback is positive as there is a 180° phase shift at resonance. The amplitude is limited by the max peak-to-peak voltage at the FET drain. The resistor R can be used to reduce the feedback and the crystal drive, but is not required for oscillation. Resistor R does influence the waveshape, however. As you increase R (experiment with different values!), the waveform becomes more and more sinusoidal. Too high a value - around 15k - will prevent the oscillator from starting. Try different values of R, but you should probably stay below this value. But above all, play with it!

Hope this has been helpful.

-e

there are lots of ways to build up a oscillator.

what waveform do you hope to get? sine wave or pulse?

H or C oscillator can produce a good sine waveform with only one transistor. very simple, just select your amp cooeficient and suitable capacitor and inductor to get the number frequency. if your load is capacitance, C is good.

you can also use IC with a ceramic or cystal oscillator to get either sine wave or pulse wave frequency with high stablity. very easy operate, just link cer or cys to IC enough.

Hi Sprite,

all good responses (with the exception of the obvious..).

Just a suggestion in general, 2 books that contain answers to a lot of these sorts of problems:

The Art of Electronics, by Horowitz & Hill

The Circuit Designer's Companion by Tim Williams

Have fun!

RF_G

Thanks for all the great info!! I never even considered a crystal oscillator, automatically assuming it would be too difficult and expensive to assemble. I think it might be the direction I'd like to move in.

Given that, I'll get my hands on those reference books and do some more investigating :)

Thanks again!

Sprite

Hello again Sprite,

Glad to be of help!

I saw this today...

This oven-controlled, 5 MHz crystal oscillator (OCXO) was originally designed for satellite applications but is now available through alltronics.com for about 50 bucks. Sounds like a lot to pay until you consider its original cost was around US $290. If you need a very stable frequency reference (and can't afford a cesium or rubidium freq standard - like who can?) OCXOs are the way to go. Heck, I might just pick up one of these myself.

What is your application, by the way? Just curious.

-e

Dear Sprite, Google Electronickits.com the have quite a selection

Dragon

I'm building an RF ID tag and am using the oscillator as an ID signal. Right now I have a power source connected to the oscillator which is connected to a transistor and then an antenna. The transistor acts as a switch to vary the impedance seen by the incoming signal from the transceiver, modulating the signal to see the ID (which here is just 10101010...). My other group members are building different components of the transceiver.

Ideally I would like to use as little power as possible, but a stable oscillator is more important than what's powering it at the moment. Mostly because we're all building our components for the first time and aren't too sure what we're doing :) Any error we can eliminate will help significantly.

Depending how much time we have left before the end of the school year, I would like to build a passive tag which would be powered by the signal from the transceiver. We'll see how it goes :)

Sprite

Dear Sprite, I wish you the best of luck and the good news: Even if your equipment does not pass the smoke test, most failures are from a single component. Electronic parts now are relatively inexpensive.

Dragon