[RFI] led bulb test

[Jim Brown]

Hi Greg,

Outside of compliance test labs, our concern is NOT to determine 
compliance with FCC Rules for emissions, which are not necessarily a 
good indicator of the extent to which a given device will radiate enough 
noise into our antennas to be problematic. A LISN (Line Impedance 
Simulation Network) simulates the DIFFERENTIAL loading of the noise 
source by the power system, but most noise is radiated either by 1) the 
combination of poor circuit layout and poor shielding, or 2) as a COMMON 
MODE signal on ALL wiring connected to it (that is, not only the power 
line).

As a point of clarification, what EMC rules call "common mode" is 
voltage between neutral and green, while what WE call common mode, 
because it's what causes radiation, is the sum of all current on the 
cable in question. In most products, that turns out to be the current on 
the green wire and on the shields of interconnecting cables. The most 
common cause of common mode noise current is what audio professionals 
call "the Pin One Problem," where the cable shield or the AC green wire 
fail to bond to the shielding enclosure at the point of entry, but 
instead go THROUGH the shielding enclosure to the circuit board. This 
puts noise current on the Green Wire, which goes right past a commercial 
line filter external to the noise source. That filter CAN be effective 
only if it bonds the Green Wire to the shielding enclosure by a "zero 
length" lead, ideally mounted to the shielding enclosure.

As I see it, a key benefit of using an SDR in chasing RFI is in 
identifying the noise -- that is, is it a power system device like an 
SMPS or variable speed motor controller, the clock for some sort of 
digital electronics, or impulse noise.  An SDR can also help by letting 
us monitor more spectrum at once as we switch the potential noise source 
on and off.

Bottom line -- as hams, we want to know if a given device will bother 
us, not whether it complies with FCC Rules.

73, Jim K9YC

On Thu,11/10/2016 5:26 AM, Greg Troxel wrote:
> Jim Brown <jim at audiosystemsgroup.com> writes:
>
>> Thanks for posting this work. Several comments. First, the use of an
>> SDR to look at spectra of the noise is a really good idea. I'm
>> currently working on updates to the ARRL Handbook, and this is
>> something I'm suggesting. The SDRPlay ($130) is a particularly good
>> choice, because it has a very wide tuning range and can display a
>> large chunk of spectrum at one time.
> I have also been trying to make measurements, but so far just looking at
> a PX3 hooked up to either an attic dipole or a proper outside antenna.
> Operationally for me, if I can't hear the difference between on/off on
> any band on either antenna, with a desk lamp in the shack, it's good
> enough.  But I'm starting to try to do this listening on closed bands;
> my 80m noise level is vastly lower midday than evening, which I suspect
> is typical for relatively quiet locations.
>
> It seems that for repeatable quantitative measuremnts, one wants a Line
> Impedance Stabilization Network and a spectrum analyzer, but this is at
> best $2K.  I have borrowed a LISN (which is large and heavy) and intend
> to hook it up (with at least a 20 dB pad) to a KX3/PX3.  That's not a
> proper SA, but should be pretty good within the ham bands.
>
> So I wonder if you are able to suggest (for the Handbook) how to
> approach repeatability without buying or homebrewing a proper LISN.
> Also, it would be good to discuss rough calibration of things like the
> SDrplay to get measurements that are plausibly close to using real lab
> equipment.
>
> I say all this realizing that the audience includes people with varying
> technical backgrounds and budgets.
>
> But overall, with some repeatability and rough calibration, it should
> then be reasonable to build up a web database of results.   It seems
> like the hard part is the LISN equivalent.  Perhaps that would make a
> good kit, or perhaps I'm overestimating the parts cost/hassle.
>
> 73 de n1dam
>