Lowest noise (phase noise and ADEV) method to achieve 10 MHz signal from 5 MHz input
Hi,
I've got a 10 MHz distribution amplifier and am considering purchasing a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal. It seems like a possibly more expensive, but lower noise way
would be to use a PLL with a divider that locks the divided 10 MHz
signal to the 5 MHz input. If the time constant of the loop filter is
set long enough, does this avoid the phase noise multiplication issue?
From what I've gathered, this is a technique HP used in some of their
gear. For example, the 8566 and 8340/1 lock a 100 MHz VCXO to an
external reference with a PLL.
Any other thoughts on this?
Matt
you can use this:
http://www.timeok.it/wp-content/uploads/2015/08/high-performance-frequency-doublerv1-31.pdf
Luciano P. S. Paramithiotti
timeok@timeok.it
www.timeok.it
Da "Matt Huszagh" huszaghmatt@gmail.com
A time-nuts@lists.febo.com
Cc
Data Sun, 28 Nov 2021 19:29:58 -0800
Oggetto [time-nuts] Lowest noise (phase noise and ADEV) method to achieve 10 MHz signal from 5 MHz input
Hi,
I've got a 10 MHz distribution amplifier and am considering purchasing a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal. It seems like a possibly more expensive, but lower noise way
would be to use a PLL with a divider that locks the divided 10 MHz
signal to the 5 MHz input. If the time constant of the loop filter is
set long enough, does this avoid the phase noise multiplication issue?
From what I've gathered, this is a technique HP used in some of their
gear. For example, the 8566 and 8340/1 lock a 100 MHz VCXO to an
external reference with a PLL.
Any other thoughts on this?
Matt
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
For those interested, Dave PA5DOF and your's truly are working to get the doubler implemented that Luciano has on his website. The first PCB prototype is in and has been built and works OK. Some minor tweaks/fixes will be added to a V0.2 PCB. If there is interest I can share the Gerber files.
The doubler is intended to be hooked up to an Efratom MRK-HLN 5MHz Rb. Our shack T&M equipment for the best part requires a 10MHz reference.
Wilko
On 29 Nov 2021, at 10:35, timeok@timeok.it wrote:
you can use this:
http://www.timeok.it/wp-content/uploads/2015/08/high-performance-frequency-doublerv1-31.pdf
Luciano P. S. Paramithiotti
timeok@timeok.it
www.timeok.it
Da "Matt Huszagh" huszaghmatt@gmail.com
A time-nuts@lists.febo.com
Cc
Data Sun, 28 Nov 2021 19:29:58 -0800
Oggetto [time-nuts] Lowest noise (phase noise and ADEV) method to achieve 10 MHz signal from 5 MHz input
Hi,
I've got a 10 MHz distribution amplifier and am considering purchasing a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal. It seems like a possibly more expensive, but lower noise way
would be to use a PLL with a divider that locks the divided 10 MHz
signal to the 5 MHz input. If the time constant of the loop filter is
set long enough, does this avoid the phase noise multiplication issue?
From what I've gathered, this is a technique HP used in some of their
gear. For example, the 8566 and 8340/1 lock a 100 MHz VCXO to an
external reference with a PLL.
Any other thoughts on this?
Matt
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
I see you are using two for the phase noise test setup (a common way of
doing it). I am curious about the phase matching between units. From the
results it appears they must be pretty well matched, but did not see any
actual data. Since it is over a very narrow frequency BW it should be no
problem. Thanks & Regards - Mike
Mike B. Feher, N4FS
89 Arnold Blvd.
Howell NJ 07731
848-245-9115
-----Original Message-----
From: timeok@timeok.it timeok@timeok.it
Sent: Monday, November 29, 2021 4:35 AM
To: time-nuts@lists.febo.com
Subject: [time-nuts] Re: Lowest noise (phase noise and ADEV) method to
achieve 10 MHz signal from 5 MHz input
you can use this:
http://www.timeok.it/wp-content/uploads/2015/08/high-performance-frequency-d
oublerv1-31.pdf
Luciano P. S. Paramithiotti
timeok@timeok.it
www.timeok.it
Da "Matt Huszagh" huszaghmatt@gmail.com
A time-nuts@lists.febo.com
Cc
Data Sun, 28 Nov 2021 19:29:58 -0800
Oggetto [time-nuts] Lowest noise (phase noise and ADEV) method to achieve
10 MHz signal from 5 MHz input
Hi,
I've got a 10 MHz distribution amplifier and am considering purchasing a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal. It seems like a possibly more expensive, but lower noise way
would be to use a PLL with a divider that locks the divided 10 MHz
signal to the 5 MHz input. If the time constant of the loop filter is
set long enough, does this avoid the phase noise multiplication issue?
From what I've gathered, this is a technique HP used in some of their
gear. For example, the 8566 and 8340/1 lock a 100 MHz VCXO to an
external reference with a PLL.
Any other thoughts on this?
Matt
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send
an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an
email to time-nuts-leave@lists.febo.com To unsubscribe, go to and follow the
instructions there.
Hi
The PLL will have the same 20 log N issue plus a number of other
problems. The big question is: what are you after?
If you need something below -175 dbc / Hz far removed, what is
the application? There may well be better ways to do this or that.
If you are after -125 dbc / Hz at 1 Hz, that’s a very different need and
the approach to getting it would be significantly different.
Simple answer is that most equipment is not all that sensitive to phase
noise and that distribution amps have a noise floor. You can go nuts
getting a super quiet reference and it will have no real impact outside
a few very specific applications. ( which likely will need custom “plumbing”).
Bob
On Nov 28, 2021, at 10:29 PM, Matt Huszagh huszaghmatt@gmail.com wrote:
Hi,
I've got a 10 MHz distribution amplifier and am considering purchasing a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal. It seems like a possibly more expensive, but lower noise way
would be to use a PLL with a divider that locks the divided 10 MHz
signal to the 5 MHz input. If the time constant of the loop filter is
set long enough, does this avoid the phase noise multiplication issue?
From what I've gathered, this is a technique HP used in some of their
gear. For example, the 8566 and 8340/1 lock a 100 MHz VCXO to an
external reference with a PLL.
Any other thoughts on this?
Matt
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
Your push-pull doubler is much nicer than my full-wave rectifier with
Schottky diodes
http://g4jnt.com/10MHzDist.pdf
Andy
www.g4jnt.com
On Mon, 29 Nov 2021 at 15:42, Wilko Bulte wkb@xs4all.nl wrote:
For those interested, Dave PA5DOF and your's truly are working to get the
doubler implemented that Luciano has on his website. The first PCB
prototype is in and has been built and works OK. Some minor tweaks/fixes
will be added to a V0.2 PCB. If there is interest I can share the Gerber
files.
The doubler is intended to be hooked up to an Efratom MRK-HLN 5MHz Rb. Our
shack T&M equipment for the best part requires a 10MHz reference.
Wilko
Luciano P. S. Paramithiotti
timeok@timeok.it
www.timeok.it
Da "Matt Huszagh" huszaghmatt@gmail.com
A time-nuts@lists.febo.com
Cc
Data Sun, 28 Nov 2021 19:29:58 -0800
Oggetto [time-nuts] Lowest noise (phase noise and ADEV) method to
achieve 10 MHz signal from 5 MHz input
Hi,
I've got a 10 MHz distribution amplifier and am considering purchasing
a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even
an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal. It seems like a possibly more expensive, but lower noise way
would be to use a PLL with a divider that locks the divided 10 MHz
signal to the 5 MHz input. If the time constant of the loop filter is
set long enough, does this avoid the phase noise multiplication issue?
From what I've gathered, this is a technique HP used in some of their
gear. For example, the 8566 and 8340/1 lock a 100 MHz VCXO to an
external reference with a PLL.
Any other thoughts on this?
Matt
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe
send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe
send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send
an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
Hej,
On Monday, 29 November 2021 04:29:58 CET Matt Huszagh wrote:
I've got a 10 MHz distribution amplifier and am considering purchasing a
5 MHz reference. Most (not all) of my equipment accepts a 5 MHz
reference, but I'd like to be able to use the existing distribution
amplifier I have if possible. Therefore, I'm considering ways I might
generate a low-noise 10 MHz signal from the 5 MHz reference.
An obvious way is to use a doubler. However, as I understand it, even an
ideal doubler will add 20log(2)=6 dB of phase noise to the 10 MHz
signal.
Yes, this is a fundamental physical property which cannot be avoided.
You can see it that way: If there is a given fluctuating time delay in your
otherwise perfect 5 MHz signal, the amount of phase shift this time
fluctuation corresponds to in a 10 MHz signal is simply twice as big as in the
5 MHz signal, simply because the phase evolves twice as fast at 10 MHz.
This factor of two makes up the 6dB in noise power.
If you divide the 6 dB noisier 10 MHz signal down again, (neglecting
additional technical noise), you get the original performance of your 5 MHz
signal back.
Cheers,
Jürgen
Jürgen Appel jap@dfm.dk writes:
Yes, this is a fundamental physical property which cannot be avoided.
You can see it that way: If there is a given fluctuating time delay in your
otherwise perfect 5 MHz signal, the amount of phase shift this time
fluctuation corresponds to in a 10 MHz signal is simply twice as big as in the
5 MHz signal, simply because the phase evolves twice as fast at 10 MHz.
This factor of two makes up the 6dB in noise power.
If you divide the 6 dB noisier 10 MHz signal down again, (neglecting
additional technical noise), you get the original performance of your 5 MHz
signal back.
Doesn't this assume that you inherit all the noise of the reference? For
example, the short-term stability of a GPS 1PPS signal is poor, but you
can still use it to discipline a VCO with much better short-term
stability and retain most of the short-term stability of the VCO.
Matt
Hi
At some point you must go “inside” the loop (or your OCXO isn’t locked …). At that
point, you run into the basic physics of frequency multiplication.
Bob
On Nov 30, 2021, at 7:16 PM, Matt Huszagh huszaghmatt@gmail.com wrote:
Jürgen Appel jap@dfm.dk writes:
Yes, this is a fundamental physical property which cannot be avoided.
You can see it that way: If there is a given fluctuating time delay in your
otherwise perfect 5 MHz signal, the amount of phase shift this time
fluctuation corresponds to in a 10 MHz signal is simply twice as big as in the
5 MHz signal, simply because the phase evolves twice as fast at 10 MHz.
This factor of two makes up the 6dB in noise power.
If you divide the 6 dB noisier 10 MHz signal down again, (neglecting
additional technical noise), you get the original performance of your 5 MHz
signal back.
Doesn't this assume that you inherit all the noise of the reference? For
example, the short-term stability of a GPS 1PPS signal is poor, but you
can still use it to discipline a VCO with much better short-term
stability and retain most of the short-term stability of the VCO.
Matt
time-nuts mailing list -- time-nuts@lists.febo.com -- To unsubscribe send an email to time-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.