What is the difference between DRO and BPF-line feedback? What is more cost effective

First case is DRO with parallel feedback.
Second case is Band-Pass-Filter microstrip feedback (DRO replaced with λ/4 microstrip).

First case is well known "cost effective" stable oscillator. Few MHz frequency drift is not a surprice for DRO. But it seems that it is not so cost effective: for desired frequency we need to buy alot of dielectic "pucks" if we want really cost-effective solution. Actually price for one DRO puck is higher than transistor and smds for small production. The only good thing about DRO i see is that it forgives many design errors.

Second case is Band-Pass-Filter microstrip feedback, rarely used. The most known area of usage is AIA (active integrated antennas) oscillators. In my opinion it is really cost effective solution. In this case we can produce oscillator for any desired frequency without "boring pucks". Again, only a Few MHz drift if substrate is stable enough over temperature range. With todays cheap microcontrollers, temperature measurement components in smd packages and varactors we can easily make frequency adjustement over temperature range. Also we get easier tuning, easier achieve repeatable oscillator design, easier to predict results in simulation software. No more "puck" tuning and gluing.

Also usage of BandPassFeedback oscillator gives us incredible strength against extremal temperature ranges, vibration and other unfriendly conditions. If we get more than 100°C, maybe even 150° or less than 30° many pucks will just peel off and say "bye-bye" to PCB. B-P-F Oscillator will just wait for good conditions and will work again nicely.

I want to understand what parameters we can compare? Why DRO is still widely used?

The main issues of the microstrip resonators compared with DROs are:
-the end effects,
-the dielectric dissipation,
-radiation losses,
-necessity of use of low Er subtrates to increase Q-factor, which results in larger component area.

Terminator3 said:

Also usage of BandPassFeedback oscillator gives us incredible strength against extremal temperature ranges, vibration and other unfriendly conditions.

Click to expand...

Not sure if that is true. For DRO design, you can choose from a range of DRO temperate coefficients (slightly different material), to compensate for temperature drift in the circuit. This way, you can build an oscillator that is very stable over temperature changes.

For your band pass as the resonant circuit, you will experience higher drift with temperature. Also, if done properly, the frequency stability and phase noise of the DRO oscillator should be much (orders of magnitude) better, due to the higher Q of a properly mounted DRO resonator.

DRO Q factor = 10.000
"Good" substrate Q factor = 1/tand = 1/0.002 = 500

Terminator3 said:

Also we get easier tuning

Click to expand...

Sure? The DRO can be easily tuned from outside, by a tuning screw that approaches the DRO. For a microstrip band pass filter, with narrow bandwith for high Q, I wonder how that can be easily in test production?

Terminator3 said:

If we get more than 100°C, maybe even 150° or less than 30° many pucks will just peel off and say "bye-bye" to PCB.

Click to expand...

My experience is different, tested up to > 180°F in temperature chamber.

DRO:

+ tighter frequency control
+ can use low cost substrate with poor thickness, dielectric, artwork control
+ better phase noise due to much higher Q
+ less frequency pulling/pushing
- smaller "tuning" range when assemble
- harder to assemble
- probably needs a metal cover
DRO puck costs more, but substrate is cheaper

Printed resonator
- poor frequency control
- need microwave substrate (rogers, alumina, etc)
- worse phase noise
- sloppy frequency pushing/pulling
+ easier to tune center frequency
+ easy to assemble
+ do not need a cover if resonator has "confined" field
- some chance of double frequency oscillation if poorly tuned

You probably choose one of the other depending on what frequency stability and phase noise you need.

Wonderful explanation. I did not knowed about double oscillation. And pros and cons is great. Thank you.

DRO's are being used in TV Sat converters traditionally because DR was only resonator up to X-band and higher with small outlines. LC-tank may work at 10 GHz only in MMIC which are last 10-15-20 years in research. Obviously, MMIC VCO is suitable for applications without outstanding requirements to phase noise such as TV, relativelly low speed communications and so on. For systems more critical to oscillator phase noise (doppler radar, high-end test equipment) one must use high Q resonators fo oscillators such as DR, YIG, "whispering gallery" mode leucosapphire resonators and so on.