The Design Of A Correlation Electron Cyclotron Emission System On J-TEXT

Anomalous transport is a key issue to affect the confinement properties of plasma. It is generally believed that the anomalous transport is caused by plasma turbulence, the specific performance of which is a variety of fluctuations in plasma, such as electric field fluctuation, the magnetic field fluctuation, density fluctuation, temperature fluctuation, and the correlation between the different fluctuations. In the past 25 years, a lot of nuclear fusion devices were committed to measure the small amplitude, high frequency electronic temperature fluctuation on the use of correlation electron cyclotron emission. In order to understand the physical mechanism of the anomalous transport, J-TEXT(27cm minor radius, 105 cm major radius) also committed to the fluctuation characteristics of temperature measurement.In this paper, a correlation electron cyclotron emission(CECE) was developed on J-TEXT for electron temperature fluctuation measurement. The essence of CECE is the different correlation length of thermal noise and temperature fluctuation, namely: thermal noise is white noise, the correlation length of which is close to 0; while the correlation length of temperature fluctuations in the micro-turbulence approximation to 1cm. As long as two separate sampling volume is less than the distance between the turbulence, but greater than the thermal noise, the cross power spectrum by the cross-correlation is decided by turbulence, and has nothing to do with the thermal noise.The existed 16-channel ECE system adopts low noise amplifiers and has a good optical design to optimize the poloidal resolution. So the CECE unit can be developed based on the 16-channel system. In the CECE section. The signal is amplified by a 35 d B low noise amplifier first. Then it is split into 4 channels. In each channel, the signal is filtered by a narrow band pass filter. Two filters have the fixed center frequency(8GHz and 8.2GHz), and-3d B bandwidth BIF=100MHz. The other two filters are yttrium iron garnet(YIG) filters, which are current tunable between 6 and 18 GHz. The YIG filters have a-3 d B passband BIF=100-230 MHz across the tunable center frequency range. The power of these signal is measured by a zero-bias schottky diode detector, and then amplified by a video amplifier. Finally, the video signal is digitized with a sampling rate of 2MHz. The center frequency of the YIG filter is controlled by a numerical control tuning device. The polodial resolution is decided by the optical system. The polodial resolution of CECE is kq =2p/2W?1.5rad/cm. The radial resolution is actually the radiation layer thickness, from simulations of the J-TEXT plasma, we can get that the radial spatial resolution of kr?12rad/cm.Before the test, we must ensure that the CECE is working normally. The electron temperature of CECE agrees with the ECE diagnostics at the adjacent position, proving that the CECE can test temperature signal normally. It verifies that the CECE radiometer is working properly. We must make sure that the noise is uncorrelated when separate frequency bands are not overlapped before the turbulence measurement. The correlation coefficient was measured in the laboratory with 2-18 GHz noise source as the input of CECE IF section. This has been verified by changing the central frequency of one YIG filter and fixing the other one by shot-shot. We can get such a conclusion from the experimental results: when the frequency interval of two channels is more than 200 MHz, the correlation coefficient is close to 0, the thermal noise is almost not related. That is, if the frequency interval of two channels is more than 200 MHz, the influence to cross spectral density of thermal noise can be ignored, and the cross spectral density can be considered completely temperature fluctuation. Therefore, through the cross-correlation way, the influence of thermal noise can be removed, and extracting real temperature fluctuation. We measure the CECE singals of which one has a center frequency of 8GHz with-3d B bandwith of 100 MHz, and another has a center frequency of 8.2GHz with-3d B bandwith of 100 MHz. We calculate the cross-correlation frequency spectrum of the two signals. we can find that the cross power spectral before the plasma breakdown is about two orders of magnitude lower than the cross power spectral when plasma is present at high frequency.