Clapp Oscillator Help

[geex]

clapp oscillator

Hello,

I am predominantly a digital design guy. However, for a recent work, I am required to design a clapp oscillaotr to operate at around 50MHz. Given below is a snapshot of the circuit. I am confused on how to simulate this circuit and I suppose this circuit doesnt wrk as I assumed it would also. Is there something I am missing. I did a transient simulation and ran it for 200ns and I dont know what the output waveforms indicate.

Could some one throw some light in tis regard.

Thanks
Geex

 

[LvW]

clapp oscillator requirements

The source node of your transistor has no dc connection to ground (via resistor).
Therefore, no bias drain current is possible.

 

[geex]

clapp oscillator analysis

Actually it did not matter. This is what I had, just a 1K res to ground. It just induces a small ( in term on zV ) in the net6 node. Is there something else I am missing and also, it would be great if someone could tell me how the simulations to find the oscillations and the frequency are to be done. this is what I had.

 

[LvW]

clapp oszillator

Question: At which point do you measure/simulate the signal ?

Hint: To start oscillations it might be necessary either to inject a short current pulse into the circuit or - alternatively - it may be sufficient to allocate an initial voltage of some mV to one of the capacitors.

If this does not help you should check the loop gain via ac analyses.

Added after 7 minutes:

Added after 24 minutes:

I forgot to mention that a large resistor is necessary between gate and ground in order to define a clean gate potential.

 

[geex]

clapp circuit

Hi,

Thanks for the reply I measured it at net 06 and net030 ( if you can refer back to the schematic ). I think I am missing on getting the measurements for the parameters right. I know you do that by taking into account the gm of the transistors. But I am not to sure, if I am doing it the right way. Are you aware of a equiation relating to the gm directly, so that I can plug in values and see if it works ?

Thanks

 

[LvW]

clapp oscillator how it works

Recommendation: Connect the gate with a 1Megohm resistor to ground and perform a DC analysis by varying the source resistor Rs. As a result you can select a proper drain current and the corresponding value for Rs.

Then, you can simulate the transfer curve of the MOSFET (with a tuning voltage at the gate) and check the transconductance g at the proper drain current.

 

[E-design]

design clapp oscillator

This should help you

 

[E-design]

clapp circuit analysis

I personally prefer this variant which allow smaller L with higher Q to be used. This will produce lower phase noise.

L2 is not necessary but it gives a cleaner sine wave.
You may need to lower R2 to get it to start up reliably.

 

[geex]

analysis+clapp+oscillator

wow. Thanks ! This should help. As of now, this is what I have. I got the oscillations going, but the amplitude is the one which kind of bothers me. Anyways, I am planing on replacing the current source ( Idc ) with a FET. So that its can be built.

Thanks for your help.[/img]

 

[geex]

how does a clapp oscillator oscillate?

Thanks for your help guys. I got the oscillations running. I have a doubt regarding the AC simulations. When I run a AC simulation through a good range and observe the nets between the inductor n capacitor and also the net which forms the voltage divider (the net tapping the capacitor to the source of the transistor ), what should be observing. When I observe them, I look for the AC magnitude and Phase. Not very familiar with oscillator theory, I am not sure what to expect when I see these simulations in these nets specifically.
Any ideas?

Thanks
-
Geex

 

[LvW]

how does a clapp oscillator work

When you perform an ac simulation - what is your motivation to do that ?
It only makes sense if you expect something and if the results can confirm theoretical calculations. Otherwise it is a waste of time.

 

[geex]

design a clapp oscillator

I would want to know and see the phase changes that a net undergoes while the oscillator is in stable state. It probably makes sense to say the phase is proper. else we would not be getting a oscillation in the first place at all. But just for a better clarity and understanding it would be great to know the various phase and magnitude changes that a various nets in the circuit undergoes.

-
Geex

 

[LvW]

clapp oscillator circuit analysis

I would want to know and see the phase changes that a net undergoes while the oscillator is in stable state. It probably makes sense to say the phase is proper. else we would not be getting a oscillation in the first place at all. But just for a better clarity and understanding it would be great to know the various phase and magnitude changes that a various nets in the circuit undergoes.
-
Geex

OK, and what is the problem? As far as I´ve understood you have performed some ac analyses. And you have seen some "magnitude and phase changes". And now ?
Did you gain a better understanding? I´m afraid: Not yet.

For my opinion, the only ac analysis which make sense is an analysis which can proove the oscillation criterion (Barkhausen). For this purpose you have to simulate the loop gain - and that´s not easy for the Clapp oscillator as there is no single loop which could be analyzed, because in principle it is a negative resistance type oscillator.
However, you can redraw the circuit with the aim to convert it to a 4-pol oscillator which resambles a clear loop.
If you are interested in the procedure you should give notice.
Regards.

 

[geex]

+clapp +oscillator

Ah ! Sorry for being vague. I rather have to see some "magnitude n phase changes", but am not sure how to see them. Typically, when I build this oscillator (in hardware), I can test it by hooking it up to a spectrum analyzer. Is there a way I can see the waveforms coming out from a spectrum analyzer ( the mag n phase probably with some noise characteristcs too) by running some simulations ( I am using the DF 2 tool kit in cadence ).

I was trying to find some help online and found that doing a pss (periodic steady state) and pnoise simulation will help us to find the noise characteristics of a oscillator. Is there anything else???

Reg.
G

 

[LvW]

clapp oscillator values fet

... I rather have to see some "magnitude n phase changes", but am not sure how to see them. ....
Is there a way I can see the waveforms coming out from a spectrum analyzer ( the mag n phase probably with some noise characteristcs too) by running some simulations .
I was trying to find some help online and found that doing a pss (periodic steady state) and pnoise simulation will help us to find the noise characteristics of a oscillator. Is there anything else???

What is your wish ? To see something in the time or frequency domain ?
Do you expect a "waveform coming out from a spectrum analyzer" ??? A waveform?
And suddenly, you speak about noise?
Now, I am confused - and I really don´t know what your problem is.
Sorry.

 

[geex]

design clapp osc

Yes, I want to see something in the frequency domain.

I spoke about noise because, I am doing some write up on this and as a part of noise analysis I had to do some pss n pnoise sims.

And actually I am confused if I am getting it wrong here. Dont we see a bell-shaped curve ( with freq. on the X axis ) when we connect say an osc to the spectral analyzer.

Something like this... See attached img

 

[LvW]

clapp oscillator circuit

Yes, I want to see something in the frequency domain.
.........................
And actually I am confused if I am getting it wrong here. Dont we see a bell-shaped curve ( with freq. on the X axis ) when we connect say an osc to the spectral analyzer.

1.) WHAT do you want or expect to see ? What is your "something" ???
(I have already mentioned the only function which makes sense: loop gain)
2.) Regarding spectrum analyzer: Do you speak about measurements of a real oscillator which is already built ?

 

[geex]

clapp oscillator start up

Sorry for the delay. I was held up in finishing another project. Coming to what I wanted to see. I had attached a picture of the waveform I am typically looking to see.

So this is hwere I stand as of now. I have to design this oscillator. Just to make sure it works, try it out by ordering parts and implementing in PCB. and if things go good, then send the designed circuit for fab.

But initially to get the simulation part right. Though I have got the oscillator to oscillate, I would like to see the noise characteristics of the circuit ( thats why the spectrum analyzer, correct me if I am wrong here ).

I had a colleague suggest to me that I use a tool called multisim for doing it ( I believe multisim will make it easier for the design to PCB board sim ), but the prob with that is I am not sure what type of FET I should be using for the simualtion.

And yes @above: Thanks for your input regarding the loop gain sim..

Reg.

 

[LvW]

designing clapp oscillator

I had attached a picture of the waveform I am typically looking to see.

Very often the source of misunderstanding is an incorrect wording. Your picture does NOT show any waveform ! It is an enveloppe of a spectral distribution - right ?

I would like to see the noise characteristics of the circuit ( thats why the spectrum analyzer, correct me if I am wrong here ).

Where do you want to see the noise? On an analyzer or on a pc screen (simulation) ?

I had a colleague suggest to me that I use a tool called multisim for doing it ( I believe multisim will make it easier for the design to PCB board sim ), but the prob with that is I am not sure what type of FET I should be using for the simualtion.

I suppose it makes sense to use the same FET type for simulation as is contained in your circuit. Or didn´t you select any type up to now ?

Question: Is "noise" the last remaining design problem for you ? If not, I would concentrate first to other design issues.

 

[geex]

oszillator mit multisim funktioniert nicht

I had attached a picture of the waveform I am typically looking to see.

Very often the source of misunderstanding is an incorrect wording. Your picture does NOT show any waveform ! It is an enveloppe of a spectral distribution - right ?

I would like to see the noise characteristics of the circuit ( thats why the spectrum analyzer, correct me if I am wrong here ).

Where do you want to see the noise? On an analyzer or on a pc screen (simulation) ?

I had a colleague suggest to me that I use a tool called multisim for doing it ( I believe multisim will make it easier for the design to PCB board sim ), but the prob with that is I am not sure what type of FET I should be using for the simualtion.

I suppose it makes sense to use the same FET type for simulation as is contained in your circuit. Or didn´t you select any type up to now ?

Question: Is "noise" the last remaining design problem for you ? If not, I would concentrate first to other design issues.

My bad. It is indeed a spectral distribution. Which is what I have wanted to see.

And I wanted to see the noise in a simulation. I did some pss and pnoise simulations and this is what I have come up with.

There are 2 images here. The first one is some noise charac. and the second one is the spectral envelope that I have wanted to see ( hopefully ).

I guess this forum has been great in helping me wok out through the designs n sims...

Do let me know if this your comment on these images.

I also have some questions in these sims.

1. Why do we see a decrease in the dB before it peaks at the desired freq.
2. Do we observer peaks at all harmonics ( as seen with those smaller spikes ar 2f0, 3f0 and so on. )
3. Is it common to see a sharp peak in the spectral envelope.

Thanks

Reg.