DR Oscillator Design Problems

[Willem]

dr oscillator

Hi there

I am designing a 4GHz DRO for my final year thesis.

Up to this point in time i have had minimal success.

I am using a BJt in common base topology as the amplifier. I match the output (collector) of the device for maximum reflection coeffecient. Then I use ADS to examine the input impedance and admittance (at the emmiter) of the device. I then construct a resonator strip (microstrip) to tune the input of the transistor. This strip is then used to couple the "puck" to the circuit. Another important thing is that the strip must be grounded in order to let DC current flow. Now here is the problems and questions:

I am not sure if my method of examining if the oscillation conditions are satsified are correct. What i do is: I do an open loop simulation where I connect term1 to the emitter, and term2 to the resonator strip, all the while having mu output match with a 50ohm load on the collector. I then check if the imaginarry parts of the device input impedance and resonator impedance add to zero. I then close the loop and run a harmonic balance simulation. Is this correct?

If anyone has designed a DRO before and can give me some pointers,I would really appreciate it.

thank you

Willem

 

[Willem]

dr oscillater

I attach a picture of my simulation technique, but i don't know from here if my steps are correct.

Any help???

 

[E-goe]

dro design pdf

Hi Willem

I found an interesting paper which decribes some design considerations for a DRO oscillator. They also did an linear and nonlinear simulation for the DRO.


**broken link removed**

Greetz E-goe

 

[Willem]

ne76038 amplifier design

Hi E-goe

thank you

I will look at it. I havent forgotten about your s-parameters.

 

[yykk]

Hi, all


I also going to design a DRO, and it works on 10 GHz, but I found that where there have no DR, the circuit is also have a resonant frequency, it is ocrrect ? Also, the output power is very low, there only around 4 dBm.


Please help

 

[Willem]

Yes YYKK that is corrrect, the oscillator should oscillate at a frequency close to the designed one when you start it up without the DR.

I got mine working, except it doesnt oscilate at one freqeucny when the puck is not on the circuit, but when i put it on it oscillates at 4.3 GHz.

My output power is quite good at 11 dBm.

Added after 24 seconds:

Yes YYKK that is corrrect, the oscillator should oscillate at a frequency close to the designed one when you start it up without the DR.

I got mine working, except it doesnt oscilate at one freqeucny when the puck is not on the circuit, but when i put it on it oscillates at 4.3 GHz.

My output power is quite good at 11 dBm.

 

[FANT]

The output power depends on the transistor, polarization and coupling under the puck ( higher the coupling, higher the noise higher the output power ).
On the breadboard lift few mm the puck fron the pirnted board for a better coupling.

Mandi

 

[biff44]

I personally have never been able to get a common base bipolar amplifier to be stable above 1 GHz, even after much effort to do so. If you do not start with a stable amplifer, it will oscillate all on its own, irrespective of if the DR puck is in or out of the circuit. For phase noise/stability reasons, I think you do not want the basic circuit to be able to oscillate without the puck installed. Don't know why, it is just very susceptable to small discontinuities.

My advice: try a common emitter or FET common source amp design.

 

[Willem]

Hi Biff44

Your opinion clears up a big misunderstanding of my circuit that i have. When i start my oscillator up without the puck present in the circuit, it doesn't oscillate , only when i introduce the puck to the circuit does it start to oscillate at 4.3 GHz. The reason i think is due to the fact that i am not using a spacer between the puck and the circuit. In my case the puck is specified at 4GHz with a spacer, and a bit higher without it. Is that then the reason why the circuit doesn't oscillate at exactly 4GHz? What doe you think?

Thanks for the insight.

 

[biff44]

The DR spacer is there to decouple the puck from your transmition lines, and to keep the effects of housing and dielectric substrate at a minimum. If you just lie the puck flat onto the substrate, you run the risk of overcoupling, possibly having more than one transmittion mode thru the puck, having too broad a resonance line, etc. Cheap DRO oscillators do this (like satellite DBS receivers, etc), but elevating the puck even just a little makes for a more repeatable and higher Q resonator.

 

[helio1972]

i am oscillating an oscillator using NE76038 at frequency 2.45 GHz. I have followed the step described in pozar's book ( common gate), but when i run the ads schemtaic i don't get oscillation , can one explain me why?
furthemore, i would like to know if even i use S parameters should i get oscillation or i must use nonlinear model of NEE76038 to get oscillation?
Hints: anyone who has an ADS project of any oscillator using FET not bipolar don't hesitate to send it to me ( microwave_tetuan@yahoo.com)
Thanks in advance

 

[h0op]

you might want to check the gonzales book chapter 5 it's all about oscillatorvdesigntheG.D. VENDELIN book (design of amplifiers and oscillators by the s parameter method) would give more description and details about design procedures and stability considerations(sub-harminc oscillations..)
your circuit might not be oscillating because you might not have satisfied the nyquist- barkhausen cruteria of oscillations you may want to check the following:
1)after obtaining the -ve resistance of the active part of your circuit you should put the resonant's part resistance (freuncy selective cirduit DR in our case) equal to |-Rin|/3 to account for large signal s parameters variation then check that the point 1+j0 is at the desired resonant frequncy(on a nyquist plot).
2)perform a large signal analysis to find max o/p power of your circuit .
3)and it's to all those designing a dro the spacings β(bet. the strip line and dielectric cavity) and l(dist. from the the beginning of the coupling strip line) shpould be obtained by a 3d EM simulator (like anosofts' hfss) to optimize max power coupling and acceptable phase noise levels .


h0op

 

[Willem]

Hi H0op

Where are those papers you promised? I would really like to read some of them.
I cant find them.

Thank You

 

[cbsirio]

Well,
I have tried several DR oscillators, with Ga-As fets or bjt, on several frequencies, but the best and affordable oscillator I have ever found on Internet is that in Infineon Application Note "appli002"

It doesn't oscillate without its DR, and you can change the transistor with BFP420 and you'll have similar performances!
Also, I have adapted it to work to 5 GHz with no problem: I repeat, it's a non-critical but affordable DR Oscillator!

Ciao from Italy!

Sirio

 

[Willem]

Hi there

I finally got my oscillator working fine. I used my own method of design from information gathered from reading lots of books, papers, poll opinions and just playing around with the design. My main problem was the pcb! I eventually had a board made with thru hole plated vias and all my problems were sorted out. the oscillator was designed for 4 GHz, but oscillates at 3.92 GHz, which i am going to tune with the puck's tuner plate. Also it doesnt oscillate without the puck placed on the circuit.

Maybe im just lucky. who knows

good luck to all those out there and thank you for everyone that helped.

Cheers from South Africa

Willem