[time-nuts] Low phase noise VCO

Wed Feb 10 12:38:38 UTC 2010

Hi

A good 64 MHz VCXO should have *better* phase noise at 100 Hz than the multiplied 10 MHz out of a GPSDO. Even if you don't hit "state of the art" at 100 Hz, it will be better past a few hundred Hz. 

A narrow loop is just fine. 

If the bandwidth gets to wide, you have to start worrying about the noise floor of the dividers and the phase detectors. With a narrow loop, a $1 or $2 all in one PLL / divider chip will do just fine. If you try to lock up something that's got -160 dbc/hz on a 10 MHz reference, that's not going to be the case.

If you are building a VCXO from scratch, build more than one. Set up a simple quadrature / mixer / sound card phase noise tester. If you already have a good sound card, the rest of it is a sub $100 investment. Then you will know what you have.

Bob 

On Feb 10, 2010, at 12:36 AM, Nick Foster wrote:

> 
> You're right, it's for a USRP. I just got annoyed with the constant frequency offset, so I'm rolling my own. Turns out there isn't much available for good off-the-shelf 64MHz VCXOs. The USRP2 has built-in support for 10MHz sync, but not having one, I'm left to what I do have. Can't injection lock the oscillator on board, as it's a self-contained square-wave clock. So it looks like I'm going to try my hand at a Butler VCXO.
> 
> That said, I know that for PLLs, the maximum control loop bandwidth you can use is limited by the pullability of your oscillator: if you use a VCXO with very low Kv, you might end up with a maximum useful loop bandwidth of 10Hz. No sense in using a 10544A to tune that! The phase noise performance would be pretty awful, since you can't tightly lock the reference oscillator to it. So does it make more sense to use a high-Q LC oscillator instead, with a much higher Kv? I'd be able to phase-lock it much more tightly to the reference oscillator. Plus, a 64MHz LC oscillator is pretty easy to build.
> 
> 
> --n
> 
>> Date: Tue, 9 Feb 2010 21:50:57 -0700
>> From: djl at montana.com
>> To: time-nuts at febo.com
>> Subject: Re: [time-nuts] Low phase noise VCO
>> 
>> Hi Bob and all:
>> This is interesting, because I suspect this frequency source is for an
>> Ettus Research USRP. A little further downtimeline I will be faced with
>> this problem as well. The SDR is designed already and requires a 64 MHz
>> clock, especially as this clock is used for the microwave transmit and
>> receive front ends. I had planned a synthesizer filtered...
>> The oscillator that is in the USRP that I have is good only to about 1
>> part in 10^8 or so for accuracy, better in stability. maybe the existing
>> osc. can be injection locked, or temp controlled in place.
>> Don
>> 
>> Bob Camp
>>> Hi
>>> 
>>> With most SDR's a spur on the clock creates a spur in the radio. No matter
>>> how you do your multiply, you will wind up with some sub-harmonics running
>>> around. Much better / easier / quicker to start at 64 or 65 MHz.
>>> 
>>> Bob
>>> 
>>> 
>>> On Feb 9, 2010, at 8:32 PM, Bruce Griffiths wrote:
>>> 
>>>> Thats not very useful when you want the 4th harmonic as its amplitude is
>>>> zero fro a 25% duty cycle.
>>>> Using a duty cycle of 1/8, 3/8 or 5/8 will maximise the amplitude of the
>>>> 4th harmonic.
>>>> 
>>>> see:
>>>> http://www.wenzel.com/pdffiles1/pdfs/choose.pdf
>>>> 
>>>> Bruce
>>>> 
>>>> Max Robinson wrote:
>>>>> If you start with a square wave odd order is all you can get but if you
>>>>> start with a pulse with a 25% duty cycle you can get even order.  It's
>>>>> best to optimize the pulse width for the harmonic you want.
>>>>> 
>>>>> Regards.
>>>>> 
>>>>> Max.  K 4 O D S.
>>>>> 
>>>>> Email: max at maxsmusicplace.com
>>>>> 
>>>>> Transistor site http://www.funwithtransistors.net
>>>>> Vacuum tube site: http://www.funwithtubes.net
>>>>> Music site: http://www.maxsmusicplace.com
>>>>> 
>>>>> To subscribe to the fun with transistors group send an email to.
>>>>> funwithtransistors-subscribe at yahoogroups.com
>>>>> 
>>>>> To subscribe to the fun with tubes group send an email to,
>>>>> funwithtubes-subscribe at yahoogroups.com
>>>>> 
>>>>> ----- Original Message ----- From: "Nick Foster"
>>>>> <bistromat at hotmail.com>
>>>>> To: <time-nuts at febo.com>
>>>>> Sent: Tuesday, February 09, 2010 6:35 PM
>>>>> Subject: Re: [time-nuts] Low phase noise VCO
>>>>> 
>>>>> 
>>>>>> 
>>>>>>> From: bill at iaxs.net
>>>>>>> To: time-nuts at febo.com
>>>>>>> Date: Tue, 9 Feb 2010 18:24:39 -0600
>>>>>>> Subject: Re: [time-nuts] Low phase noise VCO
>>>>>>> 
>>>>>>> Which leads me to ask a novice question:
>>>>>>> 
>>>>>>> Why not pull a 16 MHz crystal and multiply to 64 MHz?
>>>>>>> 
>>>>>>> If you count down from 64 to 10 MHz, isn't the multiplication inside
>>>>>>> the
>>>>>>> PLL?
>>>>>>> 
>>>>>>> Perhaps the noise is multiplied by 4, but would it work for the
>>>>>>> intended
>>>>>>> purpose?
>>>>>>> 
>>>>>>> Bill Hawkins
>>>>>> 
>>>>>> Can you do x4 multipliers? I thought odd-order harmonics were usually
>>>>>> used for multipliers. I'd be happy to be wrong!
>>>>>> 
>>>>>> Nick
>>>>>> 
>>>> 
>>>> 
>>>> 
>>>> _______________________________________________
>>>> time-nuts mailing list -- time-nuts at febo.com
>>>> To unsubscribe, go to
>>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>>> and follow the instructions there.
>>>> 
>>> 
>>> 
>>> _______________________________________________
>>> time-nuts mailing list -- time-nuts at febo.com
>>> To unsubscribe, go to
>>> https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>>> and follow the instructions there.
>>> 
>> 
>> 
>> -- 
>> Dr. Don Latham AJ7LL
>> Six Mile Systems LLP
>> 17850 Six Mile Road
>> POB 134
>> Huson, MT, 59846
>> VOX 406-626-4304
>> www.lightningforensics.com
>> www.sixmilesystems.com
>> 
>> 
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at febo.com
>> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
>> and follow the instructions there.
> 		 	   		  
> _________________________________________________________________
> Hotmail: Free, trusted and rich email service.
> http://clk.atdmt.com/GBL/go/201469228/direct/01/
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.
> 