[TowerTalk] Analyzer and AIM comments to Jim Lux

[Jim Brown]

On Thu, 5 Jun 2008 08:57:12 -0500, Jay Terleski wrote:

>have compared the AIM to HP  equipment (HP8753B, 3 GHz. two port vector network 
>analyzer, with HP85046A S-parameter test set) and found it to be comparable over
>it's frequency range. I don't have equipment like this myself, but I will be 
>glad to hear from anyone who does a comparison study. The AIM compares quite
>favorably to other antenna analyzers in it's price range, according to data 
>obtained by the ARRL lab. These comparisons can be seen on the w5big.com 
>website. 

The AIM analyzer received a very positive review last year from Rudy Severns, 
N6LF. BUT -- Jim Lux's comments are entirely on target. EVERY measurement system 
has limitations, even the most expensive. Most antenna analyers make reflection-
based measurements -- that is, they measure S11. The accuracy of such 
measurements can be quite good when the impedance of the unknown is within about 
a 3:1 (or 1:3) ratio with the characteristic impedance of the measurement system. 
Accuracy degrades rapidly (and wildly) for the reasons Jim Lux cited if the 
unknown is very small or very large. HP has published at least one applications 
note that says exactly this.  

An example of this problem is the use of reflection-based analyzers to measure 
toroidal common mode chokes, where the impedance of the choke is nearly always 
500 ohms or more, and is often 10-50 times greater! ANY reflection-based 
measurement (that is, S11) will yield poor accuracy, whether it's got HP, 
Tektronix, or AIM on the front panel. A VNA CAN be used to measure these high 
impedances, but the measurement must be of S21 with the unknown being connected 
as the series element of a voltage divider, and great care is still required. 

When the impedance of a mismatched antenna is measured at the far end of a 
transmission line, the line transforms the impedance in a complex fashion that 
depend on the length of the line, its loss characteristic, and the impedance of 
the unknown (the antenna). It is not hard to imagine impedance ratios that will 
cause gross inaccuracies in such a measurement. This is one example of the 
measurement conditions that Jim Lux is describing. 

Nearly four years ago, I published measurements of single wire chokes wound on 
various Fair-Rite cores done by my anonymous collaborator. These were reflection-
based measurements, and utilized the same hardware that W0QE used, and the guy 
who did the measurements is a VERY sharp engineer. Subsequent work I've done in 
the last year or so have shown that those data are polluted by roughly 2pF of 
stray capacitance that the analyzer failed to subtract out. For a choke of 14 
turns around Fair-Rite #61, this measurement error caused an error of nearly 80% 
in the resonant frequency and about 150% in the magnitude of the impedance at 
resonance!  About a year ago, the AIM was one of several measurement systems used 
in a "Round Robin" test to measure a similar choke. The inaccuracies in that 
measurement were of similar magnitudes. 

Bottom line -- ALL analyzers have limitations, even the best. And the accuracy 
with which an analyzer is able to subtract out the effects of a filter between it 
and the unknown should be viewed with caution. 

My work is online at http://audiosystemsgroup.com/publish

73,

Jim Brown K9YC