Cross-Spectrum L(ƒ) Workshop
April 15th, 6:30 PM – 9:00 PM
Room 110, Colorado Convention Center
*This event is not included in the Conference registration. A separate registration is required.
Registration
Registration is $50 USD and includes a light meal and electronic copy of the proceedings.
Register early! The event is limited to 40 participants.
Agenda
This is a true workshop, as opposed to a conference or a seminar.
The “Labs and Industrial Issues” session will be defined accounting for people to come.
In the spirit of the workshop, we expect objective and constructive feedback, not driven by Companies' interests.
This is a true workshop, as opposed to a conference or a seminar.
The “Labs and Industrial Issues” session will be defined accounting for people to come.
In the spirit of the workshop, we expect objective and constructive feedback, not driven by Companies' interests.
6:30 Introduction and theory – E. Rubiola, Femto-ST
7:00 Experimental evidence – C.W. Nelson, NIST
7:30 Labs and industrials issues - A few short presentations by qualified players
8:20 Round table
7:00 Experimental evidence – C.W. Nelson, NIST
7:30 Labs and industrials issues - A few short presentations by qualified players
8:20 Round table
9:00 Closing
Should you like to contribute to the “Labs and industrial issues” session, please e-mail us at
[email protected]
[email protected]
The Problem with Cross-Spectrum Measurement
The cross spectrum method is a standard practice in the measurement of oscillator phase noise. It is widely used by manufacturers and customers/users of oscillators, and by academic/public labs as well, including the primary labs.
In short, the instrument has two separate channels – each consisting of a reference oscillators and a phase detector – which measure simultaneously the oscillator under test (DUT). Assuming that the two channels are statistically independent, the average cross spectrum converges to the DUT noise spectrum, rejecting the single-channel noise (reference oscillators and a phase detector). For a treatise of the cross spectrum theory in general terms, thus not focusing on PM noise, see [1].
Inherently, the method is prone to experimental errors due to any uncontrolled phenomenon affecting both the channels (correlated). For example, the DUT AM noise yields errors if the phase detector is sensitive to power [2]. Until recently, this fact was only interpreted as a limitation of the instrument sensitivity, i.e., the lowest PM noise that one can measure. Otherwise said, most people believe that measurements always result in the over-estimation of the DUT noise because the instrument noise adds up. This is untrue, correlated noise or stray signals can result in negative-error terms. When this happens, the instrument under estimates the DUT noise, which may be embarrassing.
The full awareness of noise under estimation comes from Nelson et al. [3], from NIST. A separate warning comes from Rohde and Poddar [4].
At the state of the knowledge, the existence of the problem has been clearly identified, but the experimental conditions where it shows up are still not under control, and we still not have a general method to detect it.
Purpose of the Workshop
At the workshop we wish to inform the community that measurement errors and inconsistencies are around the corner, and we encourage the participants to communicate their experiences and needs.
At the workshop we wish to inform the community that measurement errors and inconsistencies are around the corner, and we encourage the participants to communicate their experiences and needs.
The workshop is also intended as an occasion for a voluntary collaboration between manufacturers of oscillators and instruments, users/customers, and Gov/Int’l labs.
Conclusions are intended to progressively provide guidelines or recommendations, and ultimately as a contribution to updated procedures and standards.
References
1. E. Rubiola, F. Vernotte, “The cross-spectrum experimental method,” February 2010, arXiv:1003.0113 [physics.ins-det]
2. E. Rubiola, R. Boudot, “The effect of AM noise on correlation phase-noise measurements,” IEEE Transact. Ultrason., Ferroelec. Freq. Control 54(5) p. 926-932, May 2007.
3. C. W. Nelson, A. Hati, and D. A. Howe, “A collapse of the cross-spectral function in phase noise metrology,” Rev. Sci. Instrum. 85, 024705 (2014). See also “Collapse of the Cross-spectral Function,”
4. U. L. Rohde, A. K. Poddar, “Phase Noise Measurement Techniques, Associated Uncertainty and Limitations,” Proc. 2013 IFCS EFTF Joint Meeting p.438-441, Prague, Czech Republic, 21-25 July 2013.
2. E. Rubiola, R. Boudot, “The effect of AM noise on correlation phase-noise measurements,” IEEE Transact. Ultrason., Ferroelec. Freq. Control 54(5) p. 926-932, May 2007.
3. C. W. Nelson, A. Hati, and D. A. Howe, “A collapse of the cross-spectral function in phase noise metrology,” Rev. Sci. Instrum. 85, 024705 (2014). See also “Collapse of the Cross-spectral Function,”
4. U. L. Rohde, A. K. Poddar, “Phase Noise Measurement Techniques, Associated Uncertainty and Limitations,” Proc. 2013 IFCS EFTF Joint Meeting p.438-441, Prague, Czech Republic, 21-25 July 2013.