Hi -

My frequency of interest is about 434 MHz. I have the following equipment:

HP 8405A vector voltmeter
HP Accessory kit for the above
HP App note AN77-3 (compliments of Wes Stewart)
Fluke 6061A signal generator
Mini-Circuits 40-1000 MHz dual directional coupler.

I'm not sure the info above is really required for this question, but I
thought it couldn't hurt.

The question is: I use about an 8-foot piece of coax to go from the
output of the directional coupler to the antenna so as to reduce my body
and instrument influence. When I follow AN77-3, I do so at the antenna
end of the coax. That is, I set up the 8405A with a short at the end of
the coax. I figure the coax is only about .8 dB of loss. Is this a valid
way of doing the measurement, or must I set up the 8405 with a short on
the output of the coupler and characterize the coax and then calculate
the answer?

Thanks,
John

On Dec 19, 6:37*pm, John - KD5YI wrote:
Hi -

My frequency of interest is about 434 MHz. I have the following equipment:

HP 8405A vector voltmeter
HP Accessory kit for the above
HP App note AN77-3 (compliments of Wes Stewart)
Fluke 6061A signal generator
Mini-Circuits 40-1000 MHz dual directional coupler.

I'm not sure the info above is really required for this question, but I
thought it couldn't hurt.

The question is: I use about an 8-foot piece of coax to go from the
output of the directional coupler to the antenna so as to reduce my body
and instrument influence. When I follow AN77-3, I do so at the antenna
end of the coax. That is, I set up the 8405A with a short at the end of
the coax. I figure the coax is only about .8 dB of loss. Is this a valid
way of doing the measurement, or must I set up the 8405 with a short on
the output of the coupler and characterize the coax and then calculate
the answer?

Thanks,
John

Hi John,

It's common to calibrate measurements at the end of a piece of line
with modern vector network analyzers. The analyzer does much of the
work for you: you just put, in sequence, a short, and open and a
precision load at the point you want calibrated, and the analyzer does
the rest. Obviously loss in the line will reduce the accuracy of
measurements: consider the case where the line has infinite
attenuation; in that case the signal at the sending end of the line
won't change at all for an infinite change in load at the other end of
the line. The accuracy degrades monotonically as the line attenuation
increases. Unless you're trying to get the absolute best accuracy you
can, you probably don't need to worry about 0.8dB loss in the line
(1.6dB round trip?). There's an HP ap note (or two or three?) about
this. Maybe do a bit of Googling??

If you calibrate with open, short and load, you have the information
you need to characterize the line's loss versus frequency and its
delay (or phase shift). You don't need to also measure the line
separately. You could view the calibration as having measured the
line, in fact.

Cheers,
Tom

On 12/21/2010 4:27 PM, K7ITM wrote:
On Dec 19, 6:37 pm, John - wrote:
Hi -

My frequency of interest is about 434 MHz. I have the following equipment:

HP 8405A vector voltmeter
HP Accessory kit for the above
HP App note AN77-3 (compliments of Wes Stewart)
Fluke 6061A signal generator
Mini-Circuits 40-1000 MHz dual directional coupler.

I'm not sure the info above is really required for this question, but I
thought it couldn't hurt.

The question is: I use about an 8-foot piece of coax to go from the
output of the directional coupler to the antenna so as to reduce my body
and instrument influence. When I follow AN77-3, I do so at the antenna
end of the coax. That is, I set up the 8405A with a short at the end of
the coax. I figure the coax is only about .8 dB of loss. Is this a valid
way of doing the measurement, or must I set up the 8405 with a short on
the output of the coupler and characterize the coax and then calculate
the answer?

Thanks,
John

Hi John,

It's common to calibrate measurements at the end of a piece of line
with modern vector network analyzers. The analyzer does much of the
work for you: you just put, in sequence, a short, and open and a
precision load at the point you want calibrated, and the analyzer does
the rest. Obviously loss in the line will reduce the accuracy of
measurements: consider the case where the line has infinite
attenuation; in that case the signal at the sending end of the line
won't change at all for an infinite change in load at the other end of
the line. The accuracy degrades monotonically as the line attenuation
increases. Unless you're trying to get the absolute best accuracy you
can, you probably don't need to worry about 0.8dB loss in the line
(1.6dB round trip?). There's an HP ap note (or two or three?) about
this. Maybe do a bit of Googling??

If you calibrate with open, short and load, you have the information
you need to characterize the line's loss versus frequency and its
delay (or phase shift). You don't need to also measure the line
separately. You could view the calibration as having measured the
line, in fact.

Cheers,
Tom


Many thanks, Tom. That's the answer I was hoping for. I appreciate your
help.

Cheers,
John