[R-390] FL-101 replacement Corcom IEC power Filters

[Charles Steinmetz]

Bill wrote:

>So when someone says a filter is leaky, they are not talking about the
>leakage that comes with age, yes?
>
>They are talking about the reactive leakage current that occurs with
>capacitive reactance at line voltage and frequency, no?

Correct, generally speaking, although it would not be unthinkable to 
find one with excessive ohmic leakage at this age.  That said, I have 
yet to see one with ohmic leakage.

>In other words, a brand new R-390 class filter will trip a GFCI even if
>it shows infinite DC leakage resistance.

Correct, if the GFCI is operating properly.

>In other, other words, a filter that does not trip a GFCI has to be less
>effective than the original because it has less capacitance to ground.
>And it is that capacitive reactance at RF that makes the filter work,
>with the aid of some series inductance.

It's more complicated than that.  Modern line filters use 
"common-mode chokes" -- two equal windings on the same core, each 
placed in series with one leg of the supply, in phase 
opposition.  So, differential mode current flows unimpeded, but 
common-mode current is opposed by the choke's magnetic 
field.  Theoretically, this eliminates the need for capacitors to 
chassis (or at least makes smaller caps do a bigger job than they 
would otherwise do).  Indeed, that is precisely how "medical grade" 
line filters operate -- they rely solely on the common-mode chokes to 
suppress common-mode currents, without the benefit of any capacitors 
to chassis.

Using two common-mode chokes in series makes the medical-grade 
filters somewhat effective, but the best hospital-grade filters do 
not have nearly as much common-mode attenuation as comparable filters 
that have two chokes and DO have capacitors to chassis.  Why is 
this?  Common-mode chokes work because any common-mode voltage 
creates equal currents of opposite phase, which oppose and cancel 
each other.  Therefore, they are very sensitive to any imbalance in 
the magnitude or phase of the currents generated in the common-mode 
choke.  As anyone who has tried to make a wideband RF transformer or 
balun can attest, getting good magnitude and phase matching over a 
wide frequency range is difficult.  The capacitors to chassis make up 
for the limited performance of common-mode chokes by themselves.

Consequently, "medical-grade" line filters are substantially inferior 
to standard line filters in all respects except leakage current, 
which you can easily see if you look at the attenuation vs. frequency 
data.  For example, compare the attenuation of a 3 amp Corcom 
medical-grade filter (3EHQ) to the Corcom 3VAQ (or even the 3EP or 3EMC).

Since standard filters have leakage currents well below 1mA, there is 
absolutely no need for medical grade line filters in the radio room 
and they are clearly not recommended.  Luckily, line filters do not 
do anything useful for the way we operate radios, so they are really 
nothing but decoration in any case and there is no penalty for using 
one that doesn't work very well.  Indeed, as I suggested the other 
day, you can remove it entirely and you will never notice that it is 
gone (except that it won't trip a GFCI).

Best regards,

Charles