| Both sides previous revisionPrevious revision | |
| university:tools:pluto:users:phase_noise [10 Jun 2019 16:17] – Small wording tweaks Travis Collins | university:tools:pluto:users:phase_noise [10 Jun 2019 16:22] (current) – Capitialization changes Travis Collins |
|---|
| The point of interest in the Allan Variance plot is the minimum of the curve. Normally in plots like this, the deviation is high, because of noise. Over longer observations times, the noise averages out until the minimum is reached. The minimum thus corresponds to the point in time when the deviation from the specified frequency is at its lowest. After that, the stability deteriorates due to drift, temperature effects, and aging. In this case, we can see that the time that the Pluto SDR is most stable is over 10ms (100Hz) to 1 second (1Hz). Over this, there can be a slow drift that will affect the "average" output. | The point of interest in the Allan Variance plot is the minimum of the curve. Normally in plots like this, the deviation is high, because of noise. Over longer observations times, the noise averages out until the minimum is reached. The minimum thus corresponds to the point in time when the deviation from the specified frequency is at its lowest. After that, the stability deteriorates due to drift, temperature effects, and aging. In this case, we can see that the time that the Pluto SDR is most stable is over 10ms (100Hz) to 1 second (1Hz). Over this, there can be a slow drift that will affect the "average" output. |
| |
| {{ :university:tools:pluto:users:pluto_allen_variance.png?800 |}} | {{ :university:tools:pluto:users:pluto_allen_variance.png?800 |Pluto LO Allen Variance}} |
| |
| The logarithmic x-axis corresponds to the observation time ("Tau"). Note that Tau is not the measurement time, but the evaluated time - the measurement lasts longer than Tau. Since the Rohde & Schwarz FSWP Phase Noise instrument calculates the Allan variance based on the measurement range of the phase noise measurement (offset frequency), the observation time corresponds to the measurement range and vice versa. | The logarithmic x-axis corresponds to the observation time ("Tau"). Note that Tau is not the measurement time, but the evaluated time - the measurement lasts longer than Tau. Since the Rohde & Schwarz FSWP Phase Noise instrument calculates the Allan variance based on the measurement range of the phase noise measurement (offset frequency), the observation time corresponds to the measurement range and vice versa. |
| |
| An alternative view of this is to look at stability over time with a [[wp>Spectrogram]]. A spectrogram is a visual representation of the spectrum of frequencies of a signal as it varies with time. This is a very intuitive way to look at things, but it is not numerically precise and is only a qualitative way to investigate things. For example, in the spectrogram below, a 502.001 MHz signal was output in the Pluto, we can see the signal drift over time. This drift is a measurement of the difference between the instrument (Tektronix [[https://www.tek.com/spectrum-analyzer/rsa5126b|RSA5126B]] Real Time Signal Analyzer, which is ±1ppm Initial accuracy) and the Pluto (which is ±25ppm Initial accuracy). The Pluto could be drifting, or the instrument could be drifting. The width of the drift is ~ 100 Hz over 60 minutes. For 500,000,000 Hz carrier, a 100 Hz drift is under 0.2 ppm. It should be noted that this is while the ADALM-PLUTO was allowed to warm up (via self-heating) for 20 min before this measurement was recorded. This was approximately 1 hour of measurement. | An alternative view of this is to look at stability over time with a [[wp>Spectrogram]]. A spectrogram is a visual representation of the spectrum of frequencies of a signal as it varies with time. This is a very intuitive way to look at things, but it is not numerically precise and is only a qualitative way to investigate things. For example, in the spectrogram below, a 502.001 MHz signal was output in the Pluto, we can see the signal drift over time. This drift is a measurement of the difference between the instrument (Tektronix [[https://www.tek.com/spectrum-analyzer/rsa5126b|RSA5126B]] Real Time Signal Analyzer, which has ±1ppm initial accuracy) and the Pluto (which is ±25ppm initial accuracy). The Pluto could be drifting, or the instrument could be drifting. The width of the drift is ~ 100 Hz over 60 minutes. For 500,000,000 Hz carrier, a 100 Hz drift is under 0.2 ppm. It should be noted that this is while the ADALM-PLUTO was allowed to warm up (via self-heating) for 20 min before this measurement was recorded. This was approximately 1 hour of measurement. |
| |
| {{ :university:tools:pluto:users:pluto_drift.png?800 |}} | {{ :university:tools:pluto:users:pluto_drift.png?800 |Pluto LO drift from RTS}} |
| |
| |
