The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator

Status
Not open for further replies.
Jump to Latest
#1,921
diyiggy said:
Hello Andrea,


As I have both the 5 M Hz AT & SC cuts, is it posddible to use the new Discroll board for any with the same BOM (I assume the smallest AT cut size case will ask some legs bending though) ?


If they need more than 3.2V from power supply, do the LiPoFe4 still worth it in serie (I assume you use a smd LDO reg?) ?


cheers


PS : Panettone not yet arrived yet here ! 🙂
Click to expand...


Hi,

the new Driscoll is designed to operate with the SC-Cut crystal, you cannot use the AT because it has a very low ESR not suitable for the new circuit.
The AT-Cut crystal at 5.6448MHz works with the old Driscoll circuit only.

About the power supply you can easily use LiFePo4 batteries in series without any problem. BTW, the new Driscoll oscillator will not use LDO, only a little RFI filter.

Andrea
 
#1,931
Hi,

I haven't been following this thread in a very long time, my question might be redundant/simple, but I figured I might ask anyway. The question is about the Clapp Oscillator schematic (as posted in #9 by andrea_mori in this thread).

In the schematics power supply there are decoupling capacitors and ferrite beads used for EMI suppression. This seems like a great idea, but wouldn't they also form a resonance circuit?

If I simulate this in LTspice with typical DCR of a ferrite bead (300mOhm to 600mOhm) I get a Q of about 9 and a resonance frequency of 33.9kHz (L=22uH, C=1uF, R=500mOhm).

I would think that adding an additional RC-snubber network in parallel with the capacitor would help with reducing the Q of the EMI filter.

Also, Merry christmas everyone!
 
#1,933
Mark Tillotson said:
Most of the jitter is due to broadband phase noise, ie the noise floor of the oscillator, not the close-in phase noise. Close-in phase noise is important for some RF applications, and test-gear, but not for driving an ADC or DAC, where rms jitter is what affects performance.


https://www.analog.com/media/en/training-seminars/tutorials/MT-008.pdf
Click to expand...

I understand that the majority of the jitter is due to broadband noise and 1/f noise. However that is random noise. There are also instances where the noise is not random due to magnetic/capacitive coupling or power supply noise.

If one would add an additional sensitivity to power supply noise (by LC style decoupling) wouldn't that be a bad thing? As far as I understand power supply noise (non-random) would cause AM-modulation and represent itself in the phase noise plot at peaks.
 
#1,934
Mark Tillotson said:
Most of the jitter is due to broadband phase noise, ie the noise floor of the oscillator, not the close-in phase noise. Close-in phase noise is important for some RF applications, and test-gear, but not for driving an ADC or DAC, where rms jitter is what affects performance.


https://www.analog.com/media/en/training-seminars/tutorials/MT-008.pdf
Click to expand...

In digital to analog conversion you have to look to the RMS Phase Jitter, that's mostly affected by the close-in phase noise.

Please, take a look at the attached pictures.

The first shows the phase noise to jitter conversion of phase noise plot of the first measured prototype oscillator. As you can see the calculated RMS phase jitter is around 138 femto seconds.

The second picture shows the same plot with some alterations of the close-in phase noise, +20dBC at 1Hz and 10Hz from the carrier. As you can see the RMS jitter rises to 1.282 pS.

Now the third pictures that shows the same plot but with different alterations, +20dbC at 100Hz, 1000Hz and 10000Hz from the carrier. Also in this case the RMS phase jitter rises to 516 fS, but much less than in the previous alteration of the close-in phase noise.
 

Attachments

  • First_Proto_RMS_Phase_Jitter.jpg
    First_Proto_RMS_Phase_Jitter.jpg
    57.7 KB · Views: 985
  • RMS_Phase_Jitter_add_20dbc_1_10_Hz.jpg
    RMS_Phase_Jitter_add_20dbc_1_10_Hz.jpg
    59.4 KB · Views: 951
  • RMS_Phase_Jitter_add_20dbc_100_1000_10000_Hz.jpg
    RMS_Phase_Jitter_add_20dbc_100_1000_10000_Hz.jpg
    56.4 KB · Views: 939
#1,937
andrea_mori said:
No, I only said that the major impact on the RMS Phase Jitter is the close-in phase noise, then all the broadband noise affects the RMS Jitter and should be limited anyway.
Click to expand...

Then if I may ask, why is it that the resonance of the ferrite bead and the decoupling capacitor is not damped? Wouldn't it add sensitivity to noise since the Q is about 9 at ~33.9kHz?

Am I perhaps modelling a ferrite bead wrong? I model it like an ideal inductor with about 500mOhm series resistance and about 1k parallel resistance.
 
Status
Not open for further replies.
Top Bottom