BPL Comments - Part 2

25 February 2004

John Matz KB9II

 ... More BPL Comments

* IF Resolution Bandwidth *

Someone might ask why I am so insistent about specifying a 
measurement bandwidth for the corresponding emission level.  Well, 
in most cases, where the spurious emission is a single unmodulated 
carrier, the measurement bandwidth doesn't matter.  An unmodulated 
carrier has all its power at one frequency and reads the same power 
with a 1 kHz or 1 MHz IF bandwidth receiver.  The problem is when 
the carrier is modulated, especially with high speed data.  Say we 
have a 1 milliwatt carrier with BPSK modulation of a 100 KBPS data 
stream.  The signal is probably 100 kHz wide now at an almost 
constant level.  If the test receiver has a 1 MHz IF bandwidth, 
practically all the signal is contained in the IF bandwidth and the 
power read by the receiver is 1 milliwatt.  But say the receiver IF 
resolution bandwidth (before detection) is reduced to 10 kHz.  The 
receiver now only sees 10 kHz at a time.  It reads power at less 
than a tenth of a milliwatt now, 0.1 mw.  If it is set to 1 kHz, it 
will only see a hundredth of the power in the modulated signal, 0.01 
mw.  This makes the indicated power level drop by 20 dB or more.  
Obviously the reading, and compliance to an emissions limit, depend 
on the resolution bandwidth used on a modulated carrier in a system 
carrying data traffic. 


* CISPR and Test Receivers *

It is not clear what bandwidth is to be used for Part 15 
measurements.  Part 15 allows the use of compliance by showing that 
the unit meets CISPR limits, usually CISPR 16-1.  CISPR is a 
European standard that is more restrictive than Part 15.  It provides 
limits that decrease with increasing frequency, which makes a lot of 
sense. CISPR receivers usually make measurements in a 10 kHz 
resolution bandwidth too, at least through 30 MHz test frequency.  
A 10 kHz measurement bandwidth gives a 20 dB lower reading than a 
1 MHz bandwidth on high apeed data modulation.  


* Poor Measurements *

As one can see, Part 15 does little to control interference.  The 
limit is set at a high number. It is not changed with frequency.  
No measurement bandwidth is specified.  NPRM Appendix C only 
measures vertically polarized emissions, not horizontally 
polarized ones.  So if one wants to show Part 15 compliance and 
make emissions look 30 dB better, one simply has to be cross-
polarized and turn the test instrument bandwidth down to 1 kHz. 
Then you can be in compliance.  The difficulty is that you have not 
really changed the emissions.  The unit under test could exceed 
CISPR in a 10 kHz bandwidth by 40 dB, and it could still pass 
Part 15.  

What's sad is that CISPR limits are not even tight enough to 
control interference.  There have been even tighter limits in some 
proposals in Europe, but CISPR limits are out there now.  Too bad.  
... and Part 15 limits are even higher.  


* More Recommendations *

Part 15 was created to control interference to and from broadcast 
receivers.  Spurious signals were almost always unmodulated.  In 
this new era of unlicensed transmitters, Part 15 limits are not 
adequate.  It seems that Part 15 should be changed for future 
products to a minimum of CISPR-style limits with measurement 
bandwidth specified.  Unlicensed transmitters should have specific 
radiation limits and should be placed in specific bands for 
unlicensed operation.  That way, they will not interfere with 
licensed higher power stations on other frequencies.  


John Matz KB9II
25 Feb 2004