Figure 1. The Photo of the W Band Noise Source
STZ series full waveguide band noise sources are silicon based solid state noise sources. These noise sources implement high performance semiconductor devices and propriety circuit design to offer high Excess Noise Ratio (ENR) with extreme flatness in entire waveguide bandwidth in the frequency range of 26.5 to 320 GHz in millimeter-wave waveguide bands. The standard noise sources have integrated Faraday isolators to improve port VSWR which results in more reliable noise figure measurements. The operating voltage range of the standard noise sources is from +18 to +28 VDC via a BNC (F) connector which offers immediate interface requirements with industry standard noise meters, such as Keysight Model 8970 or N8975A. In addition, the noise sources can work in either CW or pulse AM operation modes. The AM modulation mode is triggered by TTL control signal via an SMA (F) connector.
The outline drawing of the W band waveguide noise source with integrated Faraday isolator is shown in Figure 2.
Figure 2. STZ-10-IT2, W Band Noise Source Outline
The Triggering Port (female SMA connector) of the noise source is provided to turn the noise source on and off via a TTL control signal any time when the Bias is applied. The switching frequency is limited to 1 kHz. While the AM modulation rate of the standard model is at limited at 1 kHz, the higher modulation rate up to 1 MHz is offered under custom models.
The Power/Triggering Inversion Switch of the noise source is provided to manually turn the noise source on and off any time when the Bias is applied. When the switch is in “ON” position, the LED light will be illuminated, which indicates that the noise source is on (Hot).
How is the noise figure of a DUT tested?
To test the noise figure of the mixer, the low noise amplifier or the receiver, the “Y-Factor” method is commonly used. The output noise power of these devices is measured when the input port is terminated with a known “Hot” or a known “Cold” temperature termination. The noise figure of the device under test (DUT) can be calculated with the following equations. Where “Phot” and “Pcold” are the output power of the DUT when the input is terminated with “Hot” and “Cold” terminations, respectively.
The STZ series full waveguide band noise sources are used as “Hot” or “Cold” temperature terminations by switching the noise source “ON” and “OFF” with the known ENR. Based on that, the noise figure of DUT can be obtained by following equation.
How was the noise source calibrated and what is the uncertainty?
The dominant contributor to the noise figure measurement is the noise source’s ENR uncertainty. Here is the calibration method used for the series STZ noise sources.
- A “Hot” matching termination (295K) and a “Cold” liquid matching termination (77K) are used to calibrate a receiver Noise Figure (NFrec) as reference.
- Then, the known noise figure receiver is used to calibrate the noise source by switching it “ON” and “OFF” to obtain the ENR of the noise source ENRns.
- Repeat above steps five times to obtain five independent sets of NFrec and ENRns, enabling a statistical estimate of measurement uncertainty and improved repeatability. Use Student’s t-distribution to compute the confidence interval of the mean and apply the coverage factor k=2 to report the expanded uncertainty.
Eravant’s in-house LN2 noise source calibration procedure provides excellent accuracy. Figure 3 compares the Eravant calibration results with an ISO/IEC 17025-accredited noise source calibration. The difference between Eravant and the ISO/IEC 17025 reference shows a mean of 0.01 dB, a standard deviation of 0.15 dB, and a maximum absolute deviation of 0.39 dB.
Figure 3. Eravant LN2 Calibration vs. ISO-17025 Calibration
To assess calibration uncertainty, Eravant’s LN2-based method demonstrates very good repeatability. The expanded Type A uncertainty is typically below 0.1 dB across most of the band. When systematic Type B contributions are included, the total estimated uncertainty for the W-band noise source calibration is below 0.5 dB. A comparison of the ENR results and associated uncertainties between Eravant’s LN2 calibration and an ISO/IEC 17025-accredited calibration is presented in Figure 4.
Figure 4. ENR Comparison: Eravant LN2 Calibration vs. ISO 17025 Calibration (with Uncertainty bar). Left: Eravant Calibration. Right: ISO 17025 Calibration.
How often do you recommend calibrating the noise source?
The noise source is a test standard. The accuracy of its ENR value affects the noise figure of any DUT directly. It is crucial to make sure that the noise source’s ENR is within the acceptable range of the calibrated standard all time. Therefore, a reference unit, such as a known noise figure low noise amplifier is recommended to be established, kept, and used for the noise source ENR accuracy checking periodically. Such known noise figure devices can be created at the time the noise source was freshly received. In an environment with normal wear and tear, such as test lab environment, annual calibration is recommended. For more extensive or lighter usage scenarios, calibration can be performed over shorter or longer periods, such as every six months to two years, respectively.
Due to the very high millimeterwave frequencies at which Eravant noise sources operate, there are very limited calibration facilities in the industry worldwide. Therefore, the calibration of Eravant noise sources should be performed by Eravant or an Eravant-approved facility.