[Lowfer] Sheilding power cords with aluminum foil ??

[C. Turner]

Since the vast majority of the QRM is being conducted into the power 
line rather than being radiated by the cord, shielding it will have no 
real effect at LF/MF frequencies.

Remember also that at LF/MF frequencies, clamp-on ferrite cores are 
utterly useless as they simply cannot add enough reactance to the line 
on which they are installed to impede signal flow!  If you look at the 
specs for such devices, you'll see that they are even of dubious utility 
even at HF, where they may offer only a few 10's of ohms of impedance.

In general, at HF frequencies, snap-on ferrites are often more useful to 
prevent ingress rather than egress.  Most susceptible devices do usually 
have *some* amount of intrinsic RF immunity, in many cases a few extra 
10's of ohms of impedance at the interfering frequency may be all that's 
needed to reduce the RF coming in by several-fold and keep them happy.  
If you are trying to keep some noise from being emitted from a device, 
however, those extra 10's of ohms may not really knock things down too 
much - especially (as is often the case) you need to reduce their 
emissions by 10's of dB!

If you wish to add filtering that is effective at LF/MF frequencies, you 
are talking about needing to add 100's of uH (at least) to whatever line 
is conducting these out.  Taking the example of a switching power 
supply, you also need to shunt any RF on that line to the chassis/shield 
ground of that power supply - and at LF/MF frequencies, this will take 
many 1000's of pF to effectively do so.

The practice of a "Pi" or "L" circuit is a good one:  From the power 
supply, you first throw inductance in the line to add (preferably) 100's 
of ohms of reactance at the frequency of interest and then you shunt 
whatever remains to ground.

***

Take an example of a typical computer supply line choke/filter 
combination - about 100 uH and 2200pF.

At 15 MHz, we could see that the 100uH has a reactance of about 9.4k and 
the 2200pF has a reactance of about 5 ohms:  A 9k resistor shunted to 
ground by a 5 ohm pretty much shorts out the signal so if everything is 
working properly, this is pretty effective at HF.

At 150 kHz, that same 100 uH has a reactance of about 94 ohms and the 
2200 pF has a reactance of 500 ohms - a combination that will have 
little effect on the signal coming out of the box.

Considering that a snap-on ferrite core will probably offer well under 
10 uH of inductance to the line (that's being generous!) and that adding 
that ferried so doesn't add any additional shunt reactance to ground 
(like the cap would) you can see that at LF frequencies, adding that is 
ineffectual.  Likewise, if you were to assume that the foil (or even 
coax-braid) shielding would add, maybe, 100pF/foot of capacitance (an 
overly generous amount!) you can also see that that's equally 
ineffective at LF.

If you you need to modify a device that is radiating garbage from the 
power line connection to make it "LF" friendly, do what has already been 
recommended here:

- Take an AC input filter from a dead switching supply.  Often on old PC 
supplies, these are a toroid wound with twisted wire - one wire 
connected to each side of the line.  They can also take the form of a 
toroid in which wire is wound separately on each half of the toroid.  
Finally, they can also look like a split-bobbin transformer, with a 
winding on each bobbin - but these are less-common.

Having this "dual" choke is important as it forces common-mode RF 
currents to be canceled out.  Since there are typically capacitors 
across the two sides of the power line  on the "power supply" side of 
this choke, the two sides of the line are essentially shorted out at RF, 
allowing this choke to do its job.

Unfortunately, these chokes often have only have about a hundred uH of 
inductance per leg, but some have far more:  I've found that the "split 
bobbin" types (which are, unfortunately, a bit harder to find) usually 
have more inductance - sometimes in the mH range.

- The bypass caps are usually a bit too small for effective work at LF.  
Often, these are 1000-2200pF and as the above example shows, these are 
too wimpy at LF.

Good-quality high-voltage (400 WV or higher for a 120 volt AC mains) 
mylar capacitors in the 0.1-0.47 uF range are pretty easy to find and on 
the "line" side of the bifilar choke, one of these can be put between 
each side of the line (both the "hot" and "neutral") and the case 
ground.  It is also a good idea to put one of these on the "power 
supply" side - between the hot and neutral.

Now doing this poses somewhat of a hazard that is, unfortunately, a bit 
hard to avoid:  Anytime you put a capacitor between line/neutral and 
ground, you are "lighting up" the ground with the reactive currents of 
those capacitors.  If you were to measure between the "lifted" "ground" 
pin on a computer and the earth ground you'd see about 60 volts (half 
the line voltage) of AC - but the current would be low, reactively 
limited by the 2200pF caps inside the computer.

If you have, say, an entire workbench of gear, those types of capacitors 
in each piece can really add up, often causing a 60 volt "bite" if, for 
some reason, your ground wasn't really grounded.  You can, therefore, 
imagine that replacing that 2200pF with a 0.22uF capacitor will make 
this even more "bitey"!

Unfortunately, if you *really* need to effectively bypass the power 
supply at LF frequencies, this is hard to avoid - unless you can come up 
with some bifilar chokes that have several 10's of mH of inductance per 
leg which would still allow you to use the smaller (2200pF or so) bypass 
caps!

73,

Clint
KA7OEI


Andy - KU4XR wrote:
> What has been the experience from those who have tried this ??
> The standard 3 prong black computer cord, wrap a small stranded 
> wire around the cord to use as a drain wire, then wrap the cord
> with a layer of aluminum foil, place a clamp on ferrite core at 
> the outlet end, and at the computer end connect the drain wire
> to the case, either directly, thru a capacitor, or inductor
> which ever works, or leave it dis-connected if it makes things
> worse. It would be a simple method, anyone have any sucess or
> failure trying it ??  Of course at MW and LF frequencies.
>
>
>
> Andy - KU4XR - EM75xr - Friendsville, TN.
> LOWfer Beacon " XR " @ 185.29866 KHz ( QRSS-60 )
> Coordinates:  N:  35* 43' 54" - W:  84* 3' 16"
> http://www.myspace.com/beaconxr
> http://webpages.charter.net/ku4xr/
>
>
>       
> ______________________________________________________________
> Lowfer mailing list
> Home: http://mailman.qth.net/mailman/listinfo/lowfer
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:Lowfer at mailman.qth.net
>
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
>
>
>