[SOLVED] practical oscillator design

[obrien135]

**broken link removed**

http://www.analogzone.com/hft_0102.pdf

The above links show how to design an oscillator for a certain configuration, but I was unable to see an example of how this theory can apply to the design on a practical circuit such as a colptts or a hartley oscillator.

Does anyone know where I can find such info?

George

 

[goldsmith]

Dear George
Hi
As i understood , you want to learn how to design an oscillator with your hands and your self , right ? if yes , i can help you ! but before that , what is your exact problem in design . aren't you familiar with steps of design ? by the way , in which range of frequency you want design ? HF ? VHF ?UHF ? or less than these ?
Best Wishes
Goldsmith

 

[obrien135]

Dear George
Hi
As i understood , you want to learn how to design an oscillator with your hands and your self , right ? if yes , i can help you ! but before that , what is your exact problem in design . aren't you familiar with steps of design ? by the way , in which range of frequency you want design ? HF ? VHF ?UHF ? or less than these ?
Best Wishes
Goldsmith

Yes Goldsmith, I want to build it myself, and I am trying to understand the theory behind the negative resistance approach, but I can't see how to apply that to a colptts or hartley oscilator becasue the equations don't come out in the right form. Can you tell me how to define actual component values using the negative resistance method? I want to design a local oscillator for a 40M superheterodyne transceiver.

George

 

[goldsmith]

Humm ! perhaps you can't believe thing that i'll say ! i designed many many oscillators until now , but not with that way !
of course i know how to analyze an oscillator but when i want design , i have some simple ways . for example , a simple colpittes oscillator , has many configurations ! you will see many circuits that named colpittes oscillator .
I can describe my ways to design ( i collected these ways , after reading many books ) but it depends on what range of frequency you want design ! for example , would you like to start with 2MHZ ? and for another example we can design a 50MHZ . ( with high amplitudes ) .
Wish you the best
Goldsmith

 

[obrien135]

I was originally trying to design a 10.3MHz oscillator for a 40M transceiver, so that I would have an IF frequency of 3MHz. If thats too high of an IF maybe I could choose a lower one such as 455kHZ. Then the LO frequency would be 7.755MHz, or in that vicinity. Can you guide me to do that one?

George

 

[goldsmith]

Well , do you want design a super hetrodyne , transceiver ?
I'm a bit confused , i can suppose that you want design a 10.3MHZ sine wave oscillator . this is one of the circuits that i can suggest you :

Best Luck
Goldsmith

 

[obrien135]

Yes that one would be good. How would I determine the component values?

 

[BradtheRad]

Although the theoretical math explains the raw basics as to why an oscillator works...

The math cannot tell us how to get components to do what we expect.

There are unpredictable factors such as transistor gain, coil resistance, etc.

Parameters that are known must jibe with parameters that are unknown.
Example:
* Supply voltage
* ratio of values of coils and capacitors
* amount of resistance in the resonating loop
* degree to which the resonating loop is isolated
* value of resistor connected to supply rail
* value of resistor to ground rail
* amount of bias through the transistor (or amount of feedback circulating around the op amp).
Etc.

To aid in conceptualizing the concepts of operation, I have put videos on Youtube showing animated simulations of different oscillators (transistor versions). These show the dynamics of coil action, capacitor action, and current bundles flowing through wires.

Colpitts oscillators:

www.youtube.com/watch?v=mnlIifVnHSU

Clapp oscillators:

www.youtube.com/watch?v=wKnarrvynIw

 

[WimRFP]

@Obrien135: BJTs (bipolair junction transistors) are very good for oscillators up to VHF range. You can use general purpose transistors. Also low frequency transistors as BD 137 can supply > 1W at 10 MHz. The 2N2219A as suggested by Goldsmith and its (smd) deratives like PZT2222A are very useful, you don't need special expensive RF BJT.

Many references on oscillators don't address the RF behavior of the BJT. Make sure that you are familiar with the capacitances and resistors inside the transistor. Especially the total BE capacitance of the forward conducting BJT is frequently forgotten. (it can be in the 100 or more pF range, depending on ft and collector current).

Try to get in contact with local radio amateurs (HAM). If you find somebody who likes home brewing, it can be very useful in gaining knowledge and results.

The parasitic capacitance can be helpful in certain topologies, but may require modification for circuits that assume that the BJT in common emitter gives 180 degrees phase shift. I hope you have access to some measuring instruments to speed-up practical design.

 

[goldsmith]

Dear George
Hi again
I'm completely agree with BradtheRad ( i learned many things for simplifying each design from him )( such as some of oscillators ) .
And about the oscillator that i introduced ( i learned this one from one other of my best friends )
As you probably understood it is a common base amplifier . ( i designed it for frequencies around 180MHZ successfully )
Thus , because it is a CB amplifier the positive feed back will produce by a capacitor between emitter and collector .
and as you probably understood , the out put frequency will select by that LC network in collector (L1 and C3)
So Fr will given by Fr=1/(2*pi*sqrt(LC)) and about the feed back capacitor = it's value will be nearby 1/2*c3 ( don't forget that the internally capacitor of collector emitter junction will have effect on it's value ( those are in parallel together ) and you should consider it's value at high frequencies as well .
and about R3 and R2 , you can use a simple potentiometer with a series resistance to prevent short circuiting , and then you can calibrate it for highest out put amplitude . ( it will be the best way that you can use , but if you don't want use a potentiometer , you should design a CB amplifier , independently , and i won't prefer it for these circuits ! )
and the out put will be from collector that you should use a coupling capacitor to remove DC signal from your out put . in this oscillator you can achieve the out puts , that are multiplied by twice the vcc ( vpp) , which is good ! ( i earned 80vp at 5MHZ) ! and about bypass capacitor of base it should has low impedance instead of frequency of oscillation ( about 10 ohms , instead frequency of operation ( approx )) . and about emitter resistor , you'd better to select it with low values ! ( in practice it should be between 20 ohms and 47 ohms ) and be sure that these considerations will work for all of the frequencies that you need . and about the input potentiometer that i talked about it , at past , it's value can be between 20 k ohms until 1 mega ohms .
and about the duty of the bypass capacitor in emitter , it should has an impedance around 75 ohms instead of operating frequency . as you probably know the Av at CB amplifier will be : Zc/ZE . so if you decrease the ZE the out put voltage will increase but you can't decrease it lower than 75 ohms because thus the feed back voltage will connect to the ground . it will bypass the high frequency components that we want them ( those components maybe create because of a bit nonlinearity ! ) ( each oscillator should have terms of linearity and can be a bit nonlinear )
By the way , the out put impedance of this oscillator will be as high as 200 K ohms and if you need to get current from it's out put you'll need a buffer with transistor or ....... . and about inductors : if you create enough gap space in the core of them , or select them with air core it will be very good , because with gap space your inductor will be pretty linear . ( the nonlinearity of inductor can be cause of damped oscillations and then oscillation will stop !) ( i tested it for many times ) .
And at practice and in board ( not bread board ) and at high frequencies you'd better to use a capacitor in parallel with supply ( pretty nearby your circuit ) , because at high frequencies the current will be high frequency and it will destroy on equivalent circuit of wire ( low value capacitor )
( i have same and good ways for other oscillators too ! )
Try to design , and ask again if you couldn't earn success .
Good luck
Goldsmith

---------- Post added at 03:14 ---------- Previous post was at 03:11 ----------

( by the way , i forgot to say ! try to design , for 10 MHZ , if you couldn't , i can design one of them and bring you the design example ! )

 

[obrien135]

Thank you very much fellows. Its been very helpful

George