colpitt oscillator

hi all,

iam looking for a circuit diagram for a colpitt oscillator designed to produce
50 MHZ frequency or more , i found some circuits but with no analysis or description :S
i hope you guys can help me.

Do you understand how the oscillator works at lower frequencies?

The Wiki for Colpitts Oscillator shows a complete circuit with values for a 50Mhz output, around halfway down the page.

If you need a very accurate signal source, try a crystal based oscillator.

Colpitts and Armstrong oscillators both pick a point within the tuned circuit to use as a common reference. This means that on either end of the tank circuit the signal is 180° out of phase. Since many types of amplifiers are inverting this provides the necessary positive feedback for sustained oscillation.

thatoneguy said:

Do you understand how the oscillator works at lower frequencies?

The Wiki for Colpitts Oscillator shows a complete circuit with values for a 50Mhz output, around halfway down the page.

If you need a very accurate signal source, try a crystal based oscillator.

Click to expand...

thanx man for helping, i tried crystal oscillator but i dont know how to model the crystal in a crystal oscillatort iam talking about the values of C1,C2,R1,Rf
if i have a crystal 50MHZ or 60MHZ what will be the values ?

In your Pierce oscillator above with the crystal, Rf should be around 1MΩ, and R1 should be 1k for LS series logic, up to 330k or more for CMOS (HC/HCT) based logic. C1 and C2 should be equal, typically between 20pF and 66pF.

You will want to buffer the output through another inverter so the load doesn't kill the oscillations.

thatoneguy said:

In your Pierce oscillator above with the crystal, Rf should be around 1MΩ, and R1 should be 1k for LS series logic, up to 330k or more for CMOS (HC/HCT) based logic. C1 and C2 should be equal, typically between 20pF and 66pF.

You will want to buffer the output through another inverter so the load doesn't kill the oscillations.

Click to expand...

i found this PDF (the PDF in the attachment) which say that for 1 MHZ crystal the feed back resistance should be from 5 to 10 MΩ and for 10MHZ crystal Rf should be from 1 to 5 MΩ
and for 20MHZ Rf should be from 470 to 5MΩ so iam guessing for 60MHZ crystal i need around 100kΩ for Rf also according to the pdf R1 should equal 80 Ω . need some opinions .

also do i need another inverter at the output ?

Yes.......

thanx for the reply , but could you explain to me plz why i need onther inverter ? cuz all the circuits i found it using 1 inverter only like post #4

madhat said:

thanx for the reply , but could you explain to me plz why i need onther inverter ? cuz all the circuits i found it using 1 inverter only like post #4

Click to expand...

If the clock is driving another IC or a long PCB trace/wire, the extra load may cause the amplitude or frequency of the oscillator to drift or stop completely. Essentially, the same effect as changing the value of R1 drastically.

A buffer amplifier on the output adds isolation for the oscillator. This creates a large fanout ability (can clock many ICs) and the oscillations remain stable.

The buffer is usually another inverter, as the oscillating inverter is either a 7404 or 7414 or a 40106 (the last two having a schmitt trigger integral). There are 6 inverters in these ICs, so using an extra one as an isolation amplifier / buffer requires no extra parts.

thatoneguy said:

In your Pierce oscillator above with the crystal, Rf should be around 1MΩ, and R1 should be 1k for LS series logic, up to 330k or more for CMOS (HC/HCT) based logic. C1 and C2 should be equal, typically between 20pF and 66pF.

You will want to buffer the output through another inverter so the load doesn't kill the oscillations.

Click to expand...

I think 330k, in conjunction with C1 and C2, will cause WAY too much attenuation and phase shift at 60MHz.

thanx for replying.

There is another thread now open on this topic, so I suggest going here - http://forum.allaboutcircuits.com/showthread.php?t=46911