Hi Pierre,
did anyone reply to your post ? I would be a bit surprised in that case, given
the high and broad level of EMC experience among us.
I am pretty much interested in how your NSA and antenna calibration measurements
are going and what results are you getting. This subject is widely covered and
discussed in the EMC literature and my understanding is that there are quite a
number of different opinions on how to combine or match NSA with the most
appropriate antenna factors.
I presently work in a 3m/10m OATS, but I also have worked in 3/10m SARs.
First one question: did you actually perform the four different NSA measurements
that you describe in your email ?
You say :
"Depending the method used, the NSA is OK or NOT".
So I guess you did try all of them. If this is the case, which one(s) got the
good NSA and which one(s) didn't ??
By "good NSA" (the one that is OK in your words) I mean one that is within +/-
4dB all over the 30-1000 MHz range as specified by ANSI C63.4 and CISPR22.
Now I give you my personal opinion, based both on my direct experience and on
other people's experience, taken from the EMC literature.
I personally used either choice 1 or 2.
CHOICE 1: the manufacturer's antenna factors IF they are individually calibrated
for the serial numbers of the antennas that you purchased and IF they are
determined following one of ANSI C63.5 methods. This can work for about 1 year
after purchase of the antennas. You should recalibrate your antennas after one
year of frequent use.
CHOICE 2: antenna factors determined with the 3 Antenna Method or Standard Site
Method. The site to be used for antenna calibration should be an OATS different
from the site that you want to qualify and later used for your emissions tests,
be it an OATS or SAR. The reason is that the site imperfections of the Standard
Site (the site that you use for antenna calibration) are embedded in the
calculated AFs of the antennas and would cancel out when
you calculate the NSA of your site (the 3m SAR).
A good reference can be the following article from the 1995 IEEE EMC Symposium
Record (page 327):
W.M.Elliott, J.M.Roman, R.Robles: "Three-Site Study of Variations Introduced by
Standard Site Imperfections Using the ANSI 63.5 - 1988 Standard Site Method for
Antenna Calibration".
If you don't know how to get it, I may fax you a copy. Just let me know.
I pick up this formula from the reference and refer it to your case :
NSA = ThNSA + da - db
where
NSA = VD - VS - AF1 - AF2
ThNSA = theoretical NSA (ref. ANSI C63.4 and CISPR22)
da = SAR deviation from ThNSA
db = Standard Site deviation from ThNSA
AF1, AF2 = antenna factors of transmitting / receiving antennas
If you use the 3 antenna method in your SAR, da = db and the site error would
cancel out giving you an apparent perfect result (NSA = ThNSA) for your SAR,
even if the site is not perfect.
I would not recommend choice 3 (two antenna method) in favor of choice 2 (three
antenna method) that basically follows the same path but is more accurate.
As for choice 4, it's not clear to me if you performed the measurements in the
SAR or in a OATS. The first case would be wrong because of the above shown
formulas (you would not see your SAR's imperfections). In any case it looks to
me similar to choice 3 and so less accurate than the 3 antenna method.
If you perform both Vertical and Horizontal measurements, that would improve
your accuracy, giving you separate Vertical and Horizontal AFs. But, again, I
would stick with the "classical" three-antenna method just adding the Vertical
polarization.
The ThNSA is used in the reference standards (ANSI and CISPR) to calibrate your
site (SAR or OATS) against the reference theoretical site (infinite ground
plane).
One final point: the ANSI standard C63.5 latest edition is 1998. I heard that
free space antenna factors are considered for SARs (I still don't have a copy of
it).
My opinion is that it's better to calibrate your antennas over a reflecting
ground plane (OATS) if you are going to use the antennas over a ground plane
(OATS or SAR), because the ground plane affects antenna factors.
The good side of free space antenna factors is that they average out the AF
variations with antenna height (1-4 m scan of receiving antenna), but - on the
other end - you lose the non-negligible ground plane coupling effects. So I
think it's better to use free space AFs in a Fully Anechoic Room (FAR), rather
than in SARs or OATS.
My opinion is that - given the many variables involved - we should strive for
the most reasonable compromise between measurement accuracy and ease of
reproducibility of measurement procedures among all test laboratories, and that
(as we all know) is far from easy !
Hope this helps. Let me know !!
Paolo Roncone
Compuprint s.p.a. - Italy
"Pierre SELVA (NCE)" <pierre.se...@wanadoo.fr> on 09/03/2000 10.00.27
Please respond to "Pierre SELVA (NCE)" <pierre.se...@wanadoo.fr>
To: "Forum Safety-emc" <emc-p...@majordomo.ieee.org>
cc: (bcc: Paolo Roncone/IT/BULL)
Subject: Antenna factors to be used for NSA measurement
Hello all, In this forum, a lot of discussion have been made on NSA measurement. Today, I would like to have your opinion on antenna factors to be used to performed such measurement. In our laboratory, we have one Open Area test Site (3, 10 and 30m) (OATS) and a semi anechoic room at 3m (SAR). We use CHASE CBL6112A bilog antenna. (30MHz - 2GHz). My main problem is to perform the NSA measurement in the SAR, following ANSI C63.4 and/or EN55022 requirements. Which antenna factor must be used ? I have some possibilities, and the results differ depending the one I choose. 1st choice : Use the antenna manufacturer factor, which are done at 10m. 2nd choice : Use our antenna factors. These ones have been calculated on our OATS at 3m, following the ANSI C63.5 requirements, and using the 3 antenna method. In this case, the antenna are determined in horizontal polarization only. 3rd choice : Use our antenna factors, determined following ANSI C63.5 requirements, but using the 2 antenna method (chapter 5.2 of ANSI C63.5:1998) 4th choice : Use our room manufacturer method. This method consists to perform measures in the exact geometric configuration which will be used in the chamber. For example, antenna factor are determined in Vertical polarization, at 3m distances, and 1m height for the transmit antenna and 1-4m height for the receive antenna. We perform a Vdirect measurement (VD), a VSite measurement (VS) and the antenna factor is calculated with the following : Af = (VD - VS - ThNSA) / 2, where ThNSA = theoritical NSA. Following our room manufacturer, this last method is used by most of the people who are performing NSA measurement. It's not easy for me to understand why we use the theoritical NSA in this case, because our OATS is not the perfect one, and it seems to me that we introduce an error in this calculation. Depending the method used, the NSA is OK or NOT. So, I'm very interesting to know : - which factor are you using ? - does it exist some literature on this matter ? - Have you been facing this problem, and how do you solve it ? Thanks a lot for your contribution, best regards ============================================== Pierre SELVA Safety Lab manager NCE EMC, Radio and safety lab. 19 rue François BLUMET Phone : 33 4 76 27 83 83 ZI de l'Argentière Fax : 33 4 76 27 77 00 38360 SASSENAGE e-mail : pierre.se...@wanadoo.fr ==============================================
|
Hello all,
In this forum, a lot of discussion have been made on NSA measurement.
Today, I would like to have your opinion on antenna factors to be used to
performed such measurement.
In our laboratory, we have one Open Area test Site (3, 10 and 30m) (OATS)
and a semi anechoic room at 3m (SAR).
We use CHASE CBL6112A bilog antenna. (30MHz - 2GHz).
My main problem is to perform the NSA measurement in the SAR, following
ANSI C63.4 and/or EN55022 requirements. Which antenna factor must be used
?
I have some possibilities, and the results differ depending the one I
choose.
1st choice : Use the antenna manufacturer factor, which are done at
10m.
2nd choice : Use our antenna factors. These ones have been calculated on
our OATS at 3m, following the ANSI C63.5 requirements, and using the 3 antenna
method.
In this case, the antenna are determined in horizontal polarization
only.
3rd choice : Use our antenna factors, determined following ANSI C63.5
requirements, but using the 2 antenna method (chapter 5.2 of ANSI
C63.5:1998)
4th choice : Use our room manufacturer method. This method consists to
perform measures in the exact geometric configuration which will be used in the
chamber. For example, antenna factor are determined in Vertical polarization, at
3m distances, and 1m height for the transmit antenna and 1-4m height for the
receive antenna.
We perform a Vdirect measurement (VD), a VSite measurement (VS) and the
antenna factor is calculated with the following :
Af = (VD - VS - ThNSA) / 2, where ThNSA = theoritical NSA.
Following our room manufacturer, this last method is used by most of the
people who are performing NSA measurement. It's not easy for me to understand
why we use the theoritical NSA in this case, because our OATS is not the perfect
one, and it seems to me that we introduce an error in this calculation.
Depending the method used, the NSA is OK or NOT.
So, I'm very interesting to know :
- which factor are you using ?
- does it exist some literature on this matter ?
- Have you been facing this problem, and how do you solve it ?
Thanks a lot for your contribution,
best regards
==============================================
Pierre SELVA Safety Lab manager NCE EMC, Radio and safety lab. 19 rue François BLUMET Phone : 33 4 76 27 83 83 ZI de l'Argentière Fax : 33 4 76 27 77 00 38360 SASSENAGE e-mail : pierre.se...@wanadoo.fr ============================================== |
