SloppyMagic
Newbie level 3
Hello everyone,
I'm new here so I hope I'm posting this in an appropriate forum
Here is the question
I need to know if measuring a 40 dB loss ( a ku band 40 db W/G coupler for example) would be more accurate using a suitable VNA (Agilent 8722 for example) or would making the measurement using two power meters be more accurate? Freq range is 13750 MHz- 14500 MHz
what I am interested in here is amplitude only, no phase requirement.
I contend that power meters are the most accurate method of measuring power. However, some of my engineering colleagues just love their VNAs.
In order to do this calibration measurement I would employ an Agilent 8481A sensor to measure an incident power of aprox 5-10 dBm and then use an Agilent 8481D to measure the DUT output, in this case the coupling arm of the 40 dB coupler. The 8481A is suitable for measuring levels from +20 to - 30 dBm, but you don't really want to operate in the top 10dB of any of the Agilent 848X series detector's dynamic range (where they suffer some linearity error).
Likewise, the output signal of the secondary arm of the coupler would be measured using a second power meter employing the hi sensitivity 8481D sensor (-20 to -70 dBm).
if we assume an incident signal level of +5 dBm on the 8481A this would produce a signal level of aprox -35 dBm on the 8481D sensor/power meter (both power meters are Agilent 437B). This signal level keeps both sensors out of the top 10 dB of their respective dynamic ranges for max accuracy.
The measurement is done in two steps. software divides the freq band into 401 points and steps the synthesized source (Agilent 83624A) through 401 CW freqs while making an incident measurement at each CW step.
The DUT (the coupler in this example) is then inserted and, with the 8481D sensor on the secondary arm, the 401 CW freq step is repeated while measuring the signal from the coupler secondary arm using the 8481D and a second 437B. After that a little bit of arithmetic will provide the coupling loss.
The measurement using the 8722 would just be a simple through calibration followed by insertion of the DUT coupler.
I understand the VNA would be much faster, but speed is not a consideration here, while accuracy is paramount. i also understand that any change in signal level between the incident and DUT measurement with the power meters constitutes direct error so GREAT care is exercised during the process to ensure a stable incident signal.
You can assume all measurements are made using correctly configured instruments with valid and traceable calibration. Coax to waveguide adapters are Maury Microwave low VSWR adapters (VSWR <1.05, insertion loss <0.1 dB) and the mainline of the coupler is terminated in a low vswr waveguide load (again, less than 1.05:1 over the freq band of interest)
So which is likely to be more accurate in this scenario, the twin power meter measurement or the VNA?
thanks for any inputs
regards
Ed.
I'm new here so I hope I'm posting this in an appropriate forum
Here is the question
I need to know if measuring a 40 dB loss ( a ku band 40 db W/G coupler for example) would be more accurate using a suitable VNA (Agilent 8722 for example) or would making the measurement using two power meters be more accurate? Freq range is 13750 MHz- 14500 MHz
what I am interested in here is amplitude only, no phase requirement.
I contend that power meters are the most accurate method of measuring power. However, some of my engineering colleagues just love their VNAs.
In order to do this calibration measurement I would employ an Agilent 8481A sensor to measure an incident power of aprox 5-10 dBm and then use an Agilent 8481D to measure the DUT output, in this case the coupling arm of the 40 dB coupler. The 8481A is suitable for measuring levels from +20 to - 30 dBm, but you don't really want to operate in the top 10dB of any of the Agilent 848X series detector's dynamic range (where they suffer some linearity error).
Likewise, the output signal of the secondary arm of the coupler would be measured using a second power meter employing the hi sensitivity 8481D sensor (-20 to -70 dBm).
if we assume an incident signal level of +5 dBm on the 8481A this would produce a signal level of aprox -35 dBm on the 8481D sensor/power meter (both power meters are Agilent 437B). This signal level keeps both sensors out of the top 10 dB of their respective dynamic ranges for max accuracy.
The measurement is done in two steps. software divides the freq band into 401 points and steps the synthesized source (Agilent 83624A) through 401 CW freqs while making an incident measurement at each CW step.
The DUT (the coupler in this example) is then inserted and, with the 8481D sensor on the secondary arm, the 401 CW freq step is repeated while measuring the signal from the coupler secondary arm using the 8481D and a second 437B. After that a little bit of arithmetic will provide the coupling loss.
The measurement using the 8722 would just be a simple through calibration followed by insertion of the DUT coupler.
I understand the VNA would be much faster, but speed is not a consideration here, while accuracy is paramount. i also understand that any change in signal level between the incident and DUT measurement with the power meters constitutes direct error so GREAT care is exercised during the process to ensure a stable incident signal.
You can assume all measurements are made using correctly configured instruments with valid and traceable calibration. Coax to waveguide adapters are Maury Microwave low VSWR adapters (VSWR <1.05, insertion loss <0.1 dB) and the mainline of the coupler is terminated in a low vswr waveguide load (again, less than 1.05:1 over the freq band of interest)
So which is likely to be more accurate in this scenario, the twin power meter measurement or the VNA?
thanks for any inputs
regards
Ed.
Last edited:
