Techtrol Cyclonetics, Inc. The Leader in ultra low noise technology
[Home] [Papers and Application Notes]
PRODUCT
DESCRIPTION AND APPLICATION FOR MODEL NAL2010
AM and PM NOISE CALIBRATION REFERENCE
by Perry
C. Bates
Techtrol Cyclonetics, Inc.
May 23, 1996
ABSTRACT:
The NAL2010 Noise Calibration Reference provides greatly improved AM and PM noise measurement accuracies when used in conjunction with any Noise Measurement System. The instrument also provides a method to achieve AM and PM noise measurement traceability to NIST (National Institute of Standards and Technology) and can provide for noise measurement accuracies of ±1 dB or better for any AM or PM noise measurement. The NAL2010 can be used to achieve accurate calibration for both single source (frequency discriminator), or two source (phase detector), measurement methods. In addition, two coherent, correlated source channels are provided to allow measurement of the noise floor of a two source Noise Measurement System or residual noise of active or passive devices, such as amplifiers and mixers. The hardware and associated calibration process is patented by NIST, Patent No. 5,172,064.
TECHNICAL DISCUSSION:
The calibration process begins with a low noise signal source being connected to the NAL2010 which is then internally power split into two signal paths or channels as shown in figure 1. One signal channel, the Reference Channel, contains a phase shifter, which is used to facilitate shifting phase from the Reference Channel signal relative to the second signal channel. During noise measurement, the phase shifter allows the Reference Channel to adjust the phase detector of the Noise Measurement System to quadrature. The Reference Channel is available at the front of the NAL2010.
The second signal channel from the power splitter is fed to one port of a power summer and is used to develop a second output, the Noise Channel. The Noise Channel is also available at the front panel of the NAL2010. Into the other port of the power summer is fed a very accurately measured, broadband, flat response, Gaussian noise source, which contains precisely the same level, of AM and PM noise. The Gaussian noise is bandwidth limited to approximately 10% of the carrier frequency. The power level of the AM and PM components of the Gaussian noise has been measured relative to a 1 Hz bandwidth and is accurate to within ±.75 dB over the 10% bandwidth. A switch is provided so that the noise can be turned on and off as desired.
The output from the power summer contains both the carrier signal and the precisely characterized broadband noise. As the broadband noise is precisely characterized to within ±.75 dB in power in a 1 Hz bandwidth, by accurately measuring the carrier power level, the relationship between the carrier and the noise level below the carrier can be calculated. A relationship of noise relative to carrier power, or dBc, is now established to an accuracy of the carrier power measurement plus ±.75 dB for the broadband noise.
USAGE
The NAL2010 can be used in many different ways to improve noise measurement with three ways to improve measurement accuracy being discussed here.
PM NOISE, DISCRIMINATOR CALIBRATION
 |
For two carrier phase noise measurements, it is required that an initial phase discriminator sensitivity has been established, such as in figure 2A, the NAL2010 is used to measure the measurement accuracy of the phase discriminator. The Noise Measurement System is connected as the block diagram indicates in figure 2B. The phase discriminator is adjusted to quadrature using the mechanical phase shifter in the NAL2010. A noise measurement is made with the noise OFF on the NAL2010. Because the carriers fed into the phase discriminator are correlated, they will cancel and the noise floor of the Noise Measurement System will be measured. |
With the same setup as figure 2B, by turning the noise ON from the NAL2010 and making another noise measurement, the accuracy of the phase discriminator can be verified and/or established. If care is taken to maintain the same carrier levels at the phase discriminator, a noise measurement of two sources can be taken with the results all being plotted for comparison on the same graph. An example of a graph plotted using this technique is shown in figure 2C. This technique allows visual confirmation that the Noise Measurement System limits have not been reached. The noise plot of the noise source allows visual confirmation that the phase discriminator and the low noise amplifier used to bring the measurement signal to the range of the Noise Measurement System have a flat frequency/gain response. Where extremely low noise measurements are being made, additive noise associated with the Noise Measurement System can be accurately determined and the Noise Measurement System pressed to its measurement limit and still maintain a measurement accuracy of ±1 dB.
An alternate
method for calibration of the phase discriminator is provided for in the
block diagram in figure 2B. A phase modulator is shown in the Reference
Channel signal path which can be used to phase modulate the carrier. Typically
a 50 KHz tone with a modulation level at -50 dBc is used to establish
a reference for phase discriminator calibration. The method allows
phase discriminator calibration using only one carrier source.
PM NOISE, FREQUENCY DISCRIMINATOR
In Noise Measurement Systems where only one carrier is required for noise measurement, only the Noise Channel from the NAL2010 is used for calibration. With the noise ON from the NAL2010, the carrier to noise level is accurate within ±1 dB, and therefore can be applied to the input of a single carrier frequency discriminator measurement instrument for its calibration. Both the correct reporting noise level and flat response of the Noise Measurement System can be confirmed. As described above, a PM tone can be inserted as well to confirm instrument sensitivity accuracy. In this case however, the phase modulator is inserted into the Noise Channel as shown in figure 3. The noise floor of a single carrier frequency discriminator Noise Measurement System can not easily be measured. The noise floor of a frequency discriminator can be determined however if the carrier being measured is lower in noise than the noise floor of the instrument measuring it. The Techtrol Cyclonetics, Inc. LNS712C is useful for these measurements.
AM NOISE, DETECTOR CALIBRATION
AM noise can be very elusive to measure accurately and with repeatability. The
NAL2010 is ideal to remove the uncertainty for AM measurement. As
carrier power level is critical in maintaining accuracy during AM noise
measurement, we suggest that a +13 dBm carrier level be used if possible
to assure adequate noise floor capability particularly during noise measurement. 
The AM calibration process begins with applying a carrier to the NAL2010 and attaching a detector to the Noise Channel Port. Typically the carrier is adjusted to +13 dBm as mentioned above. The carrier is modulated with a 50 KHz AM tone at -50 dBc below the carrier. The detector sensitivity is established. The noise of the NAL2010 is then turned ON and an AM measurement is made to confirm the detector sensitivity accuracy and the flatness of the detector and low noise amplifier combination for making noise measurements.
FAST RESIDUAL NOISE MEASUREMENT
Achieving accurate and repeatable data for residual noise for active and passive components can be very frustrating and time consuming. The NAL2010 with other associated instruments can greatly simplify these types of measurements. The first issue in making residual noise measurements for PM noise measurements is the selection of the signal source. Even though a correlated noise measurement technique will be discussed for making residual noise measurements, care must be taken in source selection. To achieve good results closer than 10 KHz for X-Band frequencies, a source with very good phase jitter and phase noise is required to achieve repeatable, accurate results. Low noise crystal oscillators meet the requirement when state-of-the-art residual noise performance is to be measured.
A second issue in residual noise measurement is the calibration method for the phase discriminator. As the noise level from the NAL2010 is very accurately calibrated, a dBc reference line can be established and the residual measurement for the component can be determined relative from that calibration line. As shown in figure 5, the only instruments needed to make an accurate measurement of residual noise is the NAL2010, a signal source such as the Techtrol RM414 series signal source, a phase discriminator, a low noise amplifier and an appropriate spectrum analyzer. Using this method, state of the art residual noise measurements can be made at production level. See the Techtrol LNA40 data sheet for the low noise amplifier.
Measuring the residual AM noise of a component is a different situation in comparison to residual PM noise measurement. During residual phase noise measurement, two correlated carriers are fed to the phase discriminator, which is placed in quadrature. This technique cancels the correlated phase noise and only the additive noise of the component being measured remains to be measured. For AM noise measurement, no such cancellation process is available. In this case, a signal source which has AM noise better than the device to be measured must be used to measure the additive or residual noise of a component. In many cases an analysis of the signature of the noise is used to determine compliant results. Several techniques are used at Techtrol to determine the residual AM noise component of an amplifier in particular. As this is a subject to be covered application by application, Techtrol should be consulted regarding each requirement. It can be said however that the same calibration and data extraction for PM noise measurement is also be used for AM residual noise measurement.
CONCLUSION
The NAL2010 has a variety of applications not discussed in this paper. The instrument has proven to be extremely useful in noise measurements to eliminate confusion, mistakes and accuracy issues. Data can be graphically presented with calibration, measurement system noise floor and measurement results on the same graph, thereby eliminating confusion in interpreting the data from a noise measurement of a component or product.
Back to [TOP]