The PM4 is a waveguide dry calorimeter designed to be a primary standard for power measurements throughout the mm-submillimeter range. It is constructed with a waveguide load having a 6 second thermal time constant, and an excellent RF match. A thermal feedback circuit makes the sensor much faster (~0.1 sec TC) for most measurements. A calibration heater resistor is mounted on the load at nearly the same location that most of the input power should be dissipated. Very efficient coupling to the load may be made using standard linear tapers to any smaller waveguide band, and the response is fairly insensitive to the mode content. Input loss is minimized through the use of a very short waveguide. Accuracy may not be verified in the submm using any comparison standard since none exists, but the design is expected to be accurate to within 3%. Comparison between meters in the 100 GHz range shows repeatability within ~1%. Stability and noise have been optimized to permit measurements to 1 μW in a typical lab environment and down to ~0.2 μW in a stable environment. The PM4 is similar to its predecessor the PM3, except that all signal processing is digital, and it includes a digital interface. It is much faster in response, and more accurate than earlier models, the PM1 and PM2.

• Input is WR10 waveguide (1.25 x 2.5 mm) with UG387 precision flange. Useful frequency response is 75 GHz through the submillimeter range, extending even to the visible.
• Sensor size is 5.1 x 4.8 x 7.6 cm. 1 m cable connects to readout.
• 1KΩ heater resistor (on the RF load) is used for DC calibration. Internal calibration check on all ranges.
• RF repeatability between meters is better than 1%, and overall calibration better than 3% using DC calibration heater, and known input waveguide loss.
• Maximum VSWR <1.15:1 in 80-110 GHz band. VSWR is expected to be similar or better at frequencies up to 2000 GHz.
• Input loss is <0.15 dB at 90 GHz.
• 41/2 digit LED panel meter readout, with 4 power ranges. Maximum input power is 200 mW average.
• Analog output BNC connector on back panel: 0-10V corresponds to 0-full scale meter reading.
• RS232 data output port.
• Calibration factor adjustment of up to ±29.9 dB using digi switches.
• Temperature drift is compensated to <2 μW/°C.
• The sensor has a thermal time constant (1/e) of 6 seconds. For faster response, the load is heated to a nearly constant temperature using a feedback loop. When input power is applied, the heater power is reduced, and the circuit measures the change, which is equivalent to the input power. The loop gain varies with the power to be measured, changing the response time. For highest sensitivity, no feedback is used on the lowest scale.

*Response time is given as the time from application of an input to a response at the analog output of 90% of the final reading. RMS noise is in the meter reading, with a maximum meter update rate of 3 Hz.

Vertical scale 6 μW/div, horizontal 2.5 s/div.

Vertical scale 200 μW/div, horizontal .5 s/div.

• The figures above show the actual response (from analog output) to a 34 μW input on the 2 mW scale (left) and a 1 mW input on the 20 mW scale (right).
• Changing scales causes a large thermal transient due to the change in bias heat. Switching scales to a lower power (200 mW to 200 µW) requires 15 min for stabilization. Settling time is 2 min when switching scales upward. The 5 digit display, and the very low noise and high stability, eliminate the need to change scales in most measurements.
• Operational temperature range 10-30 °C.
• Required power 90-240 V 50-60 Hz.

• Sensor prepared for use in vacuum (operation in vacuum will alter the calibration and other specifications)
• Transitions to other standard waveguide bands, with estimated calibration: WR8, WR6, WR5, WR4, WR3, WR2, inquire for price.