Hello Tom,

My reaction is may be a little offset from your discussion topic, but deals 
with the roots of C63.4

The method in C63.4 is based on half wave tuned dipole antenna's  because they:

- Easy to calculate (mathematically)
- Emulate a monopole radiator / receiver as close as possible

The height variation prescribed deals with the problem of wave interference 
between the path directly from send antenna to receive antenna and the path 
that reflects by earth plane.   The vector addition of these 2 waves has 
dependent of the signal frequency peaks and zero's in the transmission path, 
that can be cancelled out by changing the height of one of the two antenna's. 
For the full frequency range from 30 MHz off this requires a max height of 4 
meters. For higher frequencies less height will be sufficient. The math example 
below this theory of height change works for mono-poles only. For broad band 
antennas the point of reception (especially. at 30 MHz) is not well defined, 
because of the physical size of the antenna. Path lengths are different for 
left tips and center of the antenna, the angles in the left and right wing of 
the antenna create different path lengths and for the LPA antenna the center of 
reception is different for high then for low frequencies.  Anyway !
broad band antennas do not have such clear responses in peak and zero's, as 
dipoles will. 

Even more, most test engineers find that in practice -were the send antenna is 
replaced by a EUT, even more a diffuse radiator then a broad band antenna- this 
variation in height never leads to finding a peak response.
The increase in path length however makes that the values measured decrease 
when the antenna rises.

For calibration purposes however C63.4 prescribes the height variation. One has 
to realize however that the fatter the antenna is compared to a tuned dipole, 
the lesser this method of calibration gives the expected response.

One way of eliminating the problem is reducing ground reflection by rising the 
total set-up to 20 meters above ground. This way no height variation is 
required, as the path via the ground is much longer as the measurement distance 
between the antenna pair. At Holland Signaal in the Netherlands such 
experiments have lead to acceptable results.

The problem as indicated by you about correlation in the vertical plane makes 
it very clear why correlation is difficult. The path lengths for right and left 
tips of the receiving antenna are VERY different when its orientation is 
vertical. This is true also for dipoles, so this method is principally 
incorrect for vertical oriented antenna calibration

Hope this gives the group some thoughts ......  

Regards,

Ing. Gert Gremmen

==
CE-test, qualified testing, 
Consultancy, Compliance tests for EMC and Electrical Safety
15 Great EMC-design tips available !
Visit our site  :  http://www.cetest.nl 
The Dutch Electronics Directory http://www.cetest.nl/electronics.htm
==


-----Oorspronkelijk bericht-----
Van:    TDonnelly [SMTP:[email protected]]
Verzonden:      dinsdag 20 januari 1998 17:00
Aan:    PSTC
Onderwerp:      Antenna Correlation

In ANSI C63.4 the required antenna is a half wave tuned dipole. There is an 
allowance that lets a broadband antenna be used provided the results can be 
correlated to the half wave tuned dipole. When performing radiated emissions 
testing to CFR 47, Part 15, using ANSI C63.4 the requirement is to height scan 
the receiving antenna from 1 to 4 meters. 
My question concerns the instance where vertical measurements are made at the 
lower frequencies (below 80 MHz). If the received antenna is height scanned 
from 1 to 4 meters at these lower frequencies the measured levels will not 
correlate to a half wave tuned dipole as the worst case vertical levels are 
typically found at lower antenna heights. It has been my experience that the 
height scan of the broadband antenna must be limited to the range that the half 
wave tuned dipole can be scanned over to demonstrate correlation.
I would like to have some discussion on this issue to get a feel for how this 
is being addressed by others.
Tom Donnelly
EMC Engineer
Lucent Technologies
[email protected]

Reply via email to