Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Question about 2.2 GHz low noise vco...????

Not open for further replies.


Member level 1
Apr 28, 2001
Reaction score
Trophy points
Activity points
I guys,

I need to develop a 2.2 - 2.4 GHz vco with very low noise caracteristics,
about -110 dBc at 10KHz and -130 dBc at 100 KHz...
It is very difficult to find a right schematics for the large tuning range !!!
Have you any good idea ??????

Thank you

Mission Impossible :!:

It's really difficult to obtain -110dBc/Hz @ 10kHz offset. Generally, there is a trade-off between VCO bandwith and phase noise. In fact , you should find a bipolar BJT which will work at that frequency having extremely low
1/f noise.
For the time being , I can not imagine a VCO with your specifications.

Good Luck

If this VCO is not for a fast sweep application, perhaps helps splitting the tuning range up to narrower subbands, and subband switching can be done by PIN switches, so effective coupling of varactor diode (and so its loss) may be made smaller.

Why not use a colpitts oscillator at about 700MHz (ours gave about 121dBc/Hz at 10KHz) then add a X3 multiplier to get to 2.5GHz.

You only need a bandwidth of 400MHz/3 and the added phase noise from the multiplier will be 20Log(N) ie 10log3 = 9.5dB so it would be just in ?

But it will all depend on the loaded Q of the resonator and coupling the varactor more to get the bandwidth will degrade the Ql and phase noise.

So while not impossible very difficult - but do you want such good phase noise anyway ??

If the VCO is going to used as part of a PLL why not have a wide loop filter so that the phase noise at 10KHz is a function of the reference noise ?

For our last application we needed a 600MHz B/W VCO at 7.5GHz Phase noise was pants at -50dBc/HZ but I have a wide loop filter (Fc at 50KHz ) so it doesn't matter that the VCO is crap !


You have to start with a VCO with low phase noise. A YIG resonator has the lowes phase noisse oscillators and can have the frequency changed by varying the current producing the magnetic field. An alternate is to use a transmission line for the resonator. Use it as a one port.

Another alternative is to use a hetrodyning method so that most of the frequency is caused by a crystal oscillator and the lesser part by a VCO. Mix them together.


Try to see if it is applicable the local oscillator used on the spectrum analyzer from matiaz vidmar published here some days ago 8 about one week-10 days ).
Its phase noise is excellent and the working frequency is about what you are looking for.


odyseus said:
Why not use a colpitts oscillator at about 700MHz (ours gave about 121dBc/Hz at 10KHz)

Hi Odyseus,
can you upload a schematic of your VCO? I never seen such a low noise VCO before! Whats your tuning range, power consumption and what varactors do you use?

There are several factors that determine the phase noise of the VCO, Q of resonator, type of transistor, collector current, ft of the transistor etc. -110dBc/Hz at 10kHz is not impossible but a little tricky, the idea of using a lower frequency at a x3 multipler is a good one

This is *very* hard to acheive!
A quick simulation using Leesons equation shows that you need very high loaded Q (50-75), very low flicker noise, low noise-figure, and high power.

This seems almost like an impossible feat! Someone mentioned this article: h**p:// but it will not give you what you need. He states it is worse than a YIG by 20 dB. Building your own YIG-oscillatro would really be a challange to anyone.

So whats your option? Why not letting us know more about your application, maybe there is a way of getting what you need without this very high performance VCO?

Do you really need to have the 2200-2400 MHz in one band momentarely or would it be possible to tune in sub-bands? If so it would be much easier, i.e. using the mixing type, or tune in sub-bands.


Not open for further replies.

Part and Inventory Search

Welcome to