Hello

I am trying to understand colpitt oscillators.
For example :

http://www.circuitstoday.com/colpitts-oscillator

or this one

http://www.falstad.com/circuit/e-colpitts.html

I can understand that the transistor is an inverting amplifier , that is 180 degree shift , and the LC tank circuit is frequency selector + phase shifter at the resonant frequency that is another 180 degree shift. So that makes it oscillate.

I can understand this much.

But why is the middle of capacitor C1 and C2 , grounded ?
I read that its called tapping and that ratio of C1 and C2 controls the amount of feedback fed back to transistor..
But what is this tapping , and what is the logic behind it ?
I read at some places that it provides a reference , but I still dont understand it.

Can somebody explain it in very simple terms ( I am a kid at this )

Regards
Silver

 

By putting the ground in the middle of the tank circuit, you get the tank circuit to invert the signal for you. This is where the inversion comes from.

 

By putting the ground in the middle of the tank circuit, you get the tank circuit to invert the signal for you. This is where the inversion comes from.

Thanks for the answer.

Does that mean , if the middle of the 2 capacitors is not grounded , the inversion will not happen ?

How do voltage on both sides look if the ground is removed ?

Regards
Silver

 

Hello Silvermoon.

I don't know your standard but here is the beginning of an explanation.
We can take it further if you like.

A Colpitts oscillator is a resonant feedback amplifier. The centre point does not need to be grounded.

In the attachments we are looking at an ac small signal analysis. DC biasing and power supplies are not included.

Look at fig1 in the attachment.

This is a general feedback amplifier using an NPN transistor. Z3 provides feedback from the collector to the base.
There is 180° phase shift from base to collector so if the network provides a further 180°, the total is 360° and the condition for oscillation is met.

Figs 2 and 3 show oscillators based on two inductors and one capacitor - the Hartley oscillator or two capacitors and one inductor.
It is essential for one of the reactances to be of the opposite sign to the others for the feedback to work. This can be proved using the ac equivalent circuit as shown in fig4.

We can do the maths if you like.

go well

 

You have a emitter with no DC path. This is important, as it is usually grounded.

Short answer, with some proviso's, is yes, it needs grounded. You will find exceptions, but in trying to wrap your head around the subject this is a good start. The Hartley oscillator does the exact same thing with the coil, for exactly the same reasons.

There are many cases in electronics where you get the basics down, and have to refine your understanding later. The first explanation lays a groundwork you can build on later, where you will find it has holes.

Theory of AM radio is a classic example of this, the first things you learn are wrong, but they get you past certain concepts that will be more understandable later.

When I was a teenager I build most of these oscillators on a chunk of with with terminal strips to get it down. My Dad had a favorite expression when I trashed his TV reception (we were in deep country), "Turn it off.".

 

You have a emitter with no DC path. This is important, as it is usually grounded.

Yes you are right , when you bias the transistor you will (usually) ground the emitter and therefore the junction of Z1 and Z2.
However this is done for DC bias reasons not for the purposes of the oscillator.

Suppose you had an inverting op amp instead of a transistor?

 

However this is done for DC bias reasons not for the purposes of the oscillator.

I think I get it now, am I correct to assume the following ?

1. the middle of the capacitors is kept at DC ground so that the output that is fed back to it oscillates at both sides of the reference voltage and then is fed back to base.

2. The emitter is also kept at the same DC ground so that the bias is maintained correctly during each part of oscillation.

Regards
Silver

 

Why this obsession with grounding?

Try this alternative view of a colpitts oscillator.
(first sketch)

The frequency selective feedback presented as a two port network with Vin and Vout (=feedback voltage)

The ground terminal is common to both input and output ie the capacitor connection is split and appears both in the input and output terminals.

If one is ground referenced both are ground referenced. It is hard to find a transistor configuration that is not ground referenced.

However the second sketch shows, as I mentioned before, an opamp implementation.

go well

 

In the diagram provided I think the opamp circuit will "work" as it is , isnt it ?

Why this obsession with grounding?
{QUOTE]

Just trying to get my theory right. Are my assumptions correct by the way ?

Can you explain me this particular circuit

It is different from the other LC circuits. I can understand its a common base circuit , but the connection of L and the capacitors looks different.

Regards
Silver

 

Are my assumptions correct by the way

No. I have said this several times in different ways. How about responding to what I said?

If the maths / circuit complexity is too much please say so. I am running blind here.

I did say that it was difficult to configure a transistor Colpitts oscillator without grounding the capacitor junction - but not impossible.

Here is a common base circuit doing just that.

Can you tell me why it is the same (at signal frequencies) as the general oscillator layout I drew before?


In your circuit C1 - L1 forms a resonant tank circuit which is frequency sensitive and has high impedance at resonance. This provides high impedance collector load for the transistor, maximising its gain at this frequency.

This collector signal is 180° out of phase with the base and emitter signals.


Are you with me so far?

 

Both the above circuits do not invert the signal, it is basically a common base amplifier, which does not invert. You have positive feedback between the emitter (input) and the collector (output) via the cap between them, which is not part of the resonant circuit. It is not a Colpitts oscillator.

While I could be wrong, one of the hallmarks of a Colpitts is the use of the tank circuit to invert the sin wave for an inverting amplifier, which is why I maintain grounding is one of the key steps.

The jammer circuit you show is not allowed on AAC, it is illegal in the USA and a safety hazard if it were ever built. Since this is for the purposes of discussion it is allowed, but if you ever try a real project please be aware of our ToS (Terms of Service).

 

You have positive feedback between the emitter (input) and the collector (output) via the cap between them

Perhaps you would like to explain this statement further, Bill

 

A common base amp is a non inverting amplifier. I almost never use it myself, but it is good for RF amplifiers, as it has pretty decent RF characteristics. Putting a capacitor between the output and input of an amp is a pretty good way of giving positive feedback, with the tank circuit on the output selects the frequency.

Give me a bit and I'll redraw the circuits I saw and repost them.

 

I have seen the lower design called a Colpitts oscillator, but don't really accept it. You will see a variation in the Wikipedia article, among others. I think there should be a emitter resistor in there somewhere.

You will find common variations of this circuit used for transmitters of various types.

******************

I got it a bit wrong, looking at it, but I think I got my idea across. If you want I will redraw it.

 

A common base amp is a non inverting amplifier

You are right I wasn't thinking earlier.


However the capacitor feedback varies with frequency according to the RC constants introduced by the capacitor itelf and the input and other circuit impedances.

In order to promote one frequency above others to form an oscillator you need a tuned feedback network.

I agree that the OP's last circuit is not a Colpitts. You can see that the feedback network is not an LC pi as mine is in post # 8.

The LC pi is the characteristic of the Colpitts, not a connection to earth of the capacitor junction point, as described in post# 10

Here is another non grounded version of your configuration. Sorry my drawing facilities are not up to yours.

 

You lost me, is that a ground on the base? Since it is in my albums I will redraw the other schematic for accuracy, if I botch let me know please.

You could always download my PaintCAD, I give it away after all.

 

OK, but you have grounded the base (or connected it to one power supply rail). Where is the feedback?

There is a scheme that uses the internal capacitance of the transistor, but I don't see that here.

To the OP, do you have any questions? We are a little off topic, but it does relate some.

 

Hello


But why is the middle of capacitor C1 and C2 , grounded ?

Regards
Silver

According to the description given with the first circuit in question, the author went on to say that the voltage developed across C2 is the regenerative phase shifted signal that needs to be fed back to the base emitter junction, to sustain oscilations.

The ground in the middle connects the top end of C2 through CE back to the emitter, while the bottom end of C2 is connected back to the base, completing the feedback voltage circuit between base emitter inputs.

Even though He said through Cc I don't quite know how it would feedback to the base that way,

unless its the inherent junction capacitance cj, between the collector and base, that has an effect at high frequencies. which would make sense then.


If I understood the author correctly.

 

No. I have said this several times in different ways. How about responding to what I said?

If the maths / circuit complexity is too much please say so. I am running blind here.

I did say that it was difficult to configure a transistor Colpitts oscillator without grounding the capacitor junction - but not impossible.

Here is a common base circuit doing just that.

Can you tell me why it is the same (at signal frequencies) as the general oscillator layout I drew before?


In your circuit C1 - L1 forms a resonant tank circuit which is frequency sensitive and has high impedance at resonance. This provides high impedance collector load for the transistor, maximising its gain at this frequency.

This collector signal is 180° out of phase with the base and emitter signals.


Are you with me so far?

1. My question is that what is the purpose of grounding exactly , and what happens if the grounding is not done. (In the common emitter case)

If the middle of C1 and C2 is cut off from -Vcc , what happens , how do the signals on both sides of tank circuit look like ?

2. In your circuit of common base , shouldnt the collector signal be in phase with the emitter signal ? I guess common base is non-inverting amplifier ?

3. Yes , as of now the complexity of the tank circuit and its various configurations , confuses me , so you need to explain me in much simpler language. (thanks in advance)

Regards
Silver

 

Both the above circuits do not invert the signal, it is basically a common base amplifier, which does not invert. You have positive feedback between the emitter (input) and the collector (output) via the cap between them, which is not part of the resonant circuit. It is not a Colpitts oscillator.

While I could be wrong, one of the hallmarks of a Colpitts is the use of the tank circuit to invert the sin wave for an inverting amplifier, which is why I maintain grounding is one of the key steps.

1. Its not a colpitt , then what is it then ?

2. Can you please confirm this :

C2 does not determine the frequence here ?
Only L1 and C1 determine the frequency here ?

Regards
Silver