Low phase noise oscillator & narrow band passive BPF

[Jensen]

multiple phase noise oscillator

In my new project i may need a low phase noise 160 or 180MHz oscillator for AD9856 ( DDS chip). Someone suggest me to use a 10MHz OCXO to catch lower phase in near-band, in this project the phase noise of DDS output must below -70dBc @10Hz. If i use 10MHz ocxo, then i must use on-chip multipler inside AD9856, it cause some degrade of phase noise. I want to disable the on-chip multipler and use a 160 or 180MHz high stability low phase noise oscillator to meet the performance what i need.
Can I?
I don't know how to design the high ferquency high stability low phase noise oscillator which i need. Could somebody help me?

An other question is how to design a narrow band passive BPF?
Realy narrow band 16.8MHz +/- 2000ppm, 18.4MHz +/-2000ppm
33.8MHz +/-2000ppm, how can i do this???????????

Thanks in advance.
Jensen

 

[rfmw]

50 mhz oscillator 3x multiplier 2x

maybe you should build a multiplier circuit yourself. get a stable and accurate crystal (30MHz overtone, not fundamental frequency!), build a transistor based oscillator and then two multiplier stages (30 MHz x3 x2 =180MHz). connect oscillator to the amplifier stage to produce multiple harmonics and just filter out 3. harmonic (90 MHz) at the first stage and 2. harmonic (180MHz) at the second one.

hope this helps,
rfmw

 

[Jensen]

ultra low phase noise dds

Hi rfmw,

Thanks for your reply.

I have some pcs of 10MHz ultra high performance OCXO(HP10811E and some others), if i built the multiplier, it means that the 10MHz sine may be multiplied 16-18 times, hows the degrade of phase noise performance!? Can u recommend some multiplier chip? I'm not work with Mr. RF before, i don't know what is the best choice of multiplier chip. Could you tell me?

Another question, can i build a 7th overtone oscillator to reach 160 or 180Mhz? Is it a good way?

Thanks & Regards

 

[Mazi3]

3x passive frequency multiplier

on the overtone frequency there is less phase noise than on fundamental. you can use 7-th overtone, but it is hard to oscilate on this frequency. multiply of 3-rd overtone is best way.

 

[Maximilian4711]

bpf sc filter carrier

Hi,

have a look at:

**broken link removed**

Max.

 

[FANT]

passive bpf

Usually the 3rd overtone are the crystals with better performance.
Fifth overtone are still good but have some ( big ) problems with temperature ( difficult to compensate ).
Seventh overtone I have never used, but I have seen are not very often used.
If you have a stable 10 MHz reference, you can simply multiply it by 3, then filter it, then again one more time *3 , filter and the last stage multiply by 2 to obtain 180 MHz.
About the noise, consider every time you double the frequency it doubles at the same manner, and you have to add the noise associated with the transistors used.
Anyway, it is still one of the better surces still obtainable at a fair price.

Mandi

 

[webdog]

build noise oscillator

I guess Fant has probably presentetd the best way of acheiving lowest phase noise! You said that you had HP10811 on hand and this is a sc-cut ovencontrolled oscillator which used to be THE standard which everyone else did compare to. I am not sure what the phase noise is at 10 Hz offset but I would guess -130 to -135 dBc/Hz.

Ulrich Rohde shows in his book "Digital PLL Frequency Synthesizers" an HP design realizing a 640 MHz source from a 10 MHz reference (most likely the 10811) Multiplying, or multiplying/filtering gives similar phase noise close to the carrier as the filters even if they are narrowband crystalfilters will not improve at 10 Hz offset. Injection locking a high-frequency oscillator 160 MHz and the mulitplying gives the worst performance as the Q on the resonator will be not as good as the reference.

So I woluld take a look at the passive 3x and 2x diode multipliers, use very low-noise amplifiers, and you should reach a noise level around -100 dBc/Hz at 10 Hz offset. A multiplication factor of 18 will give 26 dB detoriation. You will loose some s/n in both the amplifiers and diode multiplictors. But you should quite easy meet your requirement of -70 dBc/Hz @ 10Hz

A very neat multiplier is found here: https://www.wenzel.com/documents/2diomult.html
I have personally tried the 4-diode 3x multiplier with very good results!

Good Luck
/WD

 

[Jensen]

high overtone oscillator



Going to build a multipler circuit like it