Hi. When measuring Phase Noise using PLL, what noise would be contribuded from the PLL into the amplifier? Lets say we disconnect the control voltage from the oscillator, shouldn't the input stage of the PLL contribute some, altough miniscule part to the total noise seen? Would this noise be masked by the noise from the phase detector? The reason for asking is that I'm trying to improve my PN measurement setup. Currently the PLL consist of a unity gain buffer made of TL072 before the PLL filter, not unlike what Wentzel did in his amplifier. The PLL filter is a low noise op-amp (OP27). BR. Thomas. -- With Best regards, Thomas S. Knutsen. Please avoid sending me Word or PowerPoint attachments.

Hi Ok, let’s back up a bit: There are two basic regions when measuring a phase locked oscillator. You can be “inside” the loop bandwidth or “outside”. Yes it’s a bit more complex than that, but go with the idea for now. Inside the bandwidth, everything about the PLL can / may matter. It can and often does contribute noise to the result. That noise might be from the detector, a divider, or even the loop amplifier. Outside the loop bandwidth you just see the oscillator phase noise. (There is a third region in-between the two where things get even more messy). Based on that, you will see some noise from the loop amplifier. Exactly how much (is it 0.001 db?) or how far out will always be a “that depends” sort of thing. Without measurement (and possibly modeling) there is no single answer to the question. There are many noise sources in any phase noise setup so this issue is hardly unique to the loop amp. Bob > On Jun 28, 2017, at 6:29 AM, Thomas S. Knutsen <la3pna@gmail.com> wrote: > > Hi. > > When measuring Phase Noise using PLL, what noise would be contribuded from > the PLL into the amplifier? > Lets say we disconnect the control voltage from the oscillator, shouldn't > the input stage of the PLL contribute some, altough miniscule part to the > total noise seen? > Would this noise be masked by the noise from the phase detector? > > > The reason for asking is that I'm trying to improve my PN measurement > setup. > Currently the PLL consist of a unity gain buffer made of TL072 before the > PLL filter, not unlike what Wentzel did in his amplifier. The PLL filter is > a low noise op-amp (OP27). > > BR. > Thomas. > > -- > With Best regards, Thomas S. Knutsen. > > Please avoid sending me Word or PowerPoint attachments. > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there.

Thank you Bob. I was thinking of the contribution outside the loop, as for most part my PLL loops are around 1Hz and resolving closer than 10Hz isn't easy with a analyzer that stops at 10Hz. I guess what I'm asking is: should I take any care in the unity gain buffers or would any normal garden variety op-amp work? Wentzel used the 310 in his buffers, and those are a bit special, being a bipolar amp with 2nA input bias current. BR. Thomas 2017-06-28 18:08 GMT+02:00 Bob kb8tq <kb8tq@n1k.org>: > Hi > > Ok, let’s back up a bit: > > There are two basic regions when measuring a phase locked oscillator. You > can be “inside” the loop bandwidth or “outside”. Yes it’s a bit more > complex > than that, but go with the idea for now. Inside the bandwidth, everything > about > the PLL can / may matter. It can and often does contribute noise to the > result. > That noise might be from the detector, a divider, or even the loop > amplifier. Outside > the loop bandwidth you just see the oscillator phase noise. (There is a > third region > in-between the two where things get even more messy). > > Based on that, you will see some noise from the loop amplifier. Exactly > how much > (is it 0.001 db?) or how far out will always be a “that depends” sort of > thing. Without > measurement (and possibly modeling) there is no single answer to the > question. There > are many noise sources in any phase noise setup so this issue is hardly > unique to the > loop amp. > > Bob > > > On Jun 28, 2017, at 6:29 AM, Thomas S. Knutsen <la3pna@gmail.com> wrote: > > > > Hi. > > > > When measuring Phase Noise using PLL, what noise would be contribuded > from > > the PLL into the amplifier? > > Lets say we disconnect the control voltage from the oscillator, shouldn't > > the input stage of the PLL contribute some, altough miniscule part to the > > total noise seen? > > Would this noise be masked by the noise from the phase detector? > > > > > > The reason for asking is that I'm trying to improve my PN measurement > > setup. > > Currently the PLL consist of a unity gain buffer made of TL072 before the > > PLL filter, not unlike what Wentzel did in his amplifier. The PLL filter > is > > a low noise op-amp (OP27). > > > > BR. > > Thomas. > > > > -- > > With Best regards, Thomas S. Knutsen. > > > > Please avoid sending me Word or PowerPoint attachments. > > _______________________________________________ > > time-nuts mailing list -- time-nuts@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@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/ > mailman/listinfo/time-nuts > and follow the instructions there. > -- With Best regards, Thomas S. Knutsen. Please avoid sending me Word or PowerPoint attachments.

Hi, The PLL acts like a combined lowpass and highpass filter. Noise in the loop is high-pass filtered, thus, the further down from the loop bandwidth, the better suppression of the noise. Noise on the reference input is low-pass filtered, so that it passes through DC and a bit more, but high frequency jitter gets filtered out more and more as the frequency goes up from the loop bandwidth. While noise in the loop gets suppressed, it is wise not to make it too large. However, no need to go overboard. Getting the op-amp to support the loop bandwidth properly makes good sense. Too much leakage due to low load impedance is a bad thing. Cheers, Magnus On 06/28/2017 09:23 PM, Thomas S. Knutsen wrote: > Thank you Bob. > > I was thinking of the contribution outside the loop, as for most part my > PLL loops are around 1Hz and resolving closer than 10Hz isn't easy with a > analyzer that stops at 10Hz. > > I guess what I'm asking is: should I take any care in the unity gain > buffers or would any normal garden variety op-amp work? > Wentzel used the 310 in his buffers, and those are a bit special, being a > bipolar amp with 2nA input bias current. > > > BR. > Thomas > > > > 2017-06-28 18:08 GMT+02:00 Bob kb8tq <kb8tq@n1k.org>: > >> Hi >> >> Ok, let’s back up a bit: >> >> There are two basic regions when measuring a phase locked oscillator. You >> can be “inside” the loop bandwidth or “outside”. Yes it’s a bit more >> complex >> than that, but go with the idea for now. Inside the bandwidth, everything >> about >> the PLL can / may matter. It can and often does contribute noise to the >> result. >> That noise might be from the detector, a divider, or even the loop >> amplifier. Outside >> the loop bandwidth you just see the oscillator phase noise. (There is a >> third region >> in-between the two where things get even more messy). >> >> Based on that, you will see some noise from the loop amplifier. Exactly >> how much >> (is it 0.001 db?) or how far out will always be a “that depends” sort of >> thing. Without >> measurement (and possibly modeling) there is no single answer to the >> question. There >> are many noise sources in any phase noise setup so this issue is hardly >> unique to the >> loop amp. >> >> Bob >> >>> On Jun 28, 2017, at 6:29 AM, Thomas S. Knutsen <la3pna@gmail.com> wrote: >>> >>> Hi. >>> >>> When measuring Phase Noise using PLL, what noise would be contribuded >> from >>> the PLL into the amplifier? >>> Lets say we disconnect the control voltage from the oscillator, shouldn't >>> the input stage of the PLL contribute some, altough miniscule part to the >>> total noise seen? >>> Would this noise be masked by the noise from the phase detector? >>> >>> >>> The reason for asking is that I'm trying to improve my PN measurement >>> setup. >>> Currently the PLL consist of a unity gain buffer made of TL072 before the >>> PLL filter, not unlike what Wentzel did in his amplifier. The PLL filter >> is >>> a low noise op-amp (OP27). >>> >>> BR. >>> Thomas. >>> >>> -- >>> With Best regards, Thomas S. Knutsen. >>> >>> Please avoid sending me Word or PowerPoint attachments. >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@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@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/ >> mailman/listinfo/time-nuts >> and follow the instructions there. >> > > >

Hi Your op-amp noise is either 1/f sloped or flat spectrally. The EFC transfer function has a 1/f slope due to FM to PM conversion (you modulate FM and you look at PM). The noise out of the amp will roll off pretty fast. The gotcha is that your oscillator noise may also be rolling of pretty fast close in. If you are looking at a VCO, “close in” may go quite a far ways out. Your modulation sensitivity may also be quite high. For normal OCXO’s and normal op amps (at gain of 1) it’s very unusual for it to be a problem. Bob > On Jun 28, 2017, at 3:23 PM, Thomas S. Knutsen <la3pna@gmail.com> wrote: > > Thank you Bob. > > I was thinking of the contribution outside the loop, as for most part my > PLL loops are around 1Hz and resolving closer than 10Hz isn't easy with a > analyzer that stops at 10Hz. > > I guess what I'm asking is: should I take any care in the unity gain > buffers or would any normal garden variety op-amp work? > Wentzel used the 310 in his buffers, and those are a bit special, being a > bipolar amp with 2nA input bias current. > > > BR. > Thomas > > > > 2017-06-28 18:08 GMT+02:00 Bob kb8tq <kb8tq@n1k.org>: > >> Hi >> >> Ok, let’s back up a bit: >> >> There are two basic regions when measuring a phase locked oscillator. You >> can be “inside” the loop bandwidth or “outside”. Yes it’s a bit more >> complex >> than that, but go with the idea for now. Inside the bandwidth, everything >> about >> the PLL can / may matter. It can and often does contribute noise to the >> result. >> That noise might be from the detector, a divider, or even the loop >> amplifier. Outside >> the loop bandwidth you just see the oscillator phase noise. (There is a >> third region >> in-between the two where things get even more messy). >> >> Based on that, you will see some noise from the loop amplifier. Exactly >> how much >> (is it 0.001 db?) or how far out will always be a “that depends” sort of >> thing. Without >> measurement (and possibly modeling) there is no single answer to the >> question. There >> are many noise sources in any phase noise setup so this issue is hardly >> unique to the >> loop amp. >> >> Bob >> >>> On Jun 28, 2017, at 6:29 AM, Thomas S. Knutsen <la3pna@gmail.com> wrote: >>> >>> Hi. >>> >>> When measuring Phase Noise using PLL, what noise would be contribuded >> from >>> the PLL into the amplifier? >>> Lets say we disconnect the control voltage from the oscillator, shouldn't >>> the input stage of the PLL contribute some, altough miniscule part to the >>> total noise seen? >>> Would this noise be masked by the noise from the phase detector? >>> >>> >>> The reason for asking is that I'm trying to improve my PN measurement >>> setup. >>> Currently the PLL consist of a unity gain buffer made of TL072 before the >>> PLL filter, not unlike what Wentzel did in his amplifier. The PLL filter >> is >>> a low noise op-amp (OP27). >>> >>> BR. >>> Thomas. >>> >>> -- >>> With Best regards, Thomas S. Knutsen. >>> >>> Please avoid sending me Word or PowerPoint attachments. >>> _______________________________________________ >>> time-nuts mailing list -- time-nuts@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@febo.com >> To unsubscribe, go to https://www.febo.com/cgi-bin/ >> mailman/listinfo/time-nuts >> and follow the instructions there. >> > > > > -- > With Best regards, Thomas S. Knutsen. > > Please avoid sending me Word or PowerPoint attachments. > _______________________________________________ > time-nuts mailing list -- time-nuts@febo.com > To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts > and follow the instructions there.