**3. TWT Performance**

The block diagram of the experimental setup is shown in Figure 14, and the test system is shown in Figure 15. A solid-state amplifier-multiplier chain (AMC) is used to provide the input power for the TWT. Two directional couplers are used to sample the input and output power of the TWT, respectively. The input and output power are measured with two THz power meters simultaneously.

**Figure 14.** Block diagram of the experimental setup.

**Figure 15.** Test system.

The TWT operates in continuous wave mode at an optimal voltage of 20 kV and a beam current of 50 mA. The body current is 3 mA without RF and the worst body current with RF is 3.5 mA. The corresponding electron transmission ratio is over 93%. The TWT is conduction cooled through the mounting plate.

The measured output power and gain against input power for the TWT at different frequencies are shown in Figures 16 and 17. The input and output segments of the slowwave structure have the same size in the design, and the AM (amplitude modulation) /AM was not specifically considered. We plan to use anomalous dispersion to improve linearity in the future.

**Figure 16.** Measured output power against input power for the TWT at different frequencies.

**Figure 17.** Measured gain against input power for the TWT at different frequencies.

The measured saturation output power and gain of the TWT are shown in Figures 18 and 19. The saturation output power is over 8 W and the saturation gain is over 30.5 dB in 204–231 GHz. The saturation output power is over 10 W in 205–228 GHz. The maximum output power is 16 W at 218 GHz. The 3 dB bandwidth is greater than 12.3% of fc. The gain ripple is less than 10 dB in band.

**Figure 18.** Measured saturation output power of the TWT.

**Figure 19.** Measured gain of the TWT.

Comparing the measured saturation output power of the TWT with the simulation results, the output power is lower than simulated. One reason for this is that the beam current is set as 50 mA in the simulation, however, electrons of 3.5 mA are intercepted before they research the output port of the TWT, which can depress the beam–wave interaction efficiency. Another reason is that the insertion loose of the RF window is estimated as 1.5 dB in the simulation, which has been verified by the cold test of the RF window.

In addition, the maximum output power of the solid-stage source is less than 8 mW beyond 228 GHz. This is why the output power beyond 228 GHz reduced significantly.

Comparing the measured gain of the TWT with the simulation results, the gain is higher than simulated. A possible reason is that the input beam is thicker in the input part than design. This can let the average coupling impedance in the beam cross section be higher than the design. Stronger beam–wave interaction can occur, and the gain is higher.

The maximum efficiency of the TWT can research 10.2% and the total efficiency is over 5.5% in band, corresponding to a power dissipation of less than 160 W, as shown in Figure 20.

**Figure 20.** Measured total efficiency of the TWT.

Figure 21 is the photo of the packaged G-band TWT. The weight of the packaged TWT is 2.5 kg and the size is 330 mm × 70 mm × 70 mm, respectively.

**Figure 21.** The photo of the packaged G-band TWT.
