[RFI] Resolving RFI to a DSL Modem

[Jim Brown]

On Sat,3/21/2015 8:32 AM, Kim Elmore wrote:
> I have an  ADSL modem in my garage and when I operate with anything 
> more than 1300-1400 W on 40 m only, it loses the ADSL connection. I';; 
> be getting an IPDSL modem on Monday, though not Uverse (because... 
> Well, it's the phone company). I'm assuming I'll have a similar 
> problem with it, if not worse. 

Rule #1 -- avoid making someone else's problem your problem.  Those 
modems SHOULD reject RF. If they don't, they are either badly designed, 
badly installed, or defective. Bad installation includes things like 
choice of interconnecting cables. Call the vendor and tell them to fix 
it. if they don't, cancel your contract, return the product, etc.

If you're past the time frame where you can do that, there are several 
things to do.

> I've ordered five of the ~1.5" ID type 31 ferrite cores, enough for 
> one per wire that enters or exits the device. How many turns should I 
> pack on: as many as I can in a single layer? ANy guess if I can get 
> away with enough turns when I run two wires per core?

This is one of those things. What you're looking for is a common mode 
choke, so you want to wind enough turns of each CABLE to put the 
resonant choking impedance peak in the frequency range where you're 
having problems.  Since DSL is in the low MHz region, I would expect 
160M and 80/75M to cause the greatest problems.  I'd start with 10-15 
turns for those bands.

Equally important, make certain that all cables carrying the DSL signal 
are a high quality twisted pair. Twisting is FAR more important than 
shielding.  Telcos have been getting cheapo about that for the last 
several decades, and that's responsible for a lot of RFI to telco 
systems and equipment.  This reduces the DIFFERENTIAL coupling to the 
modem.

If you've done as much as you can with that, find and install a good low 
pass filter on the DSL line.

For your station, follow good engineering practice to keep RF off of 
feedlines by using common mode chokes at the feedpoint of every 
antenna.  This limits the coupling mechanism to that between the modem 
and your antenna, not the feedline.

>
> There's a twisted pair that enters the modem, a power line, and four 
> CAT-5 cables. DO I need to observe any special precautions? 

See above.  Common mode choke on every cable, low pass filter on the 
telco line only.

Before you start all this, do this simple experiment to assess how bad 
your problem is. On each band that you operate, use the antenna that 
causes the most severe interference, start with a lower power that 
doesn't break the modem, then gradually increase power until it does. 
Note these levels for each band, convert to dB re: 1500W, using the 
equation dB = 10 log (P1/P2).  If things are good at 400W but break at 
500W, you have a 4.7 dB problem.

It's also possible that there's so much RF that the front end is going 
into saturation (often called desense, desensitization) or something 
similar.

There are a few background paragraphs with more suggestions in 
k9yc.com/RFI-Ham.pdf  around pages 19 and 20.

73, Jim K9YC