Topband: 2Wire, Inc. 3800HGV-B Gateway. RFI --Problem fixed for now….

[Guy Olinger K2AV]

My AT&T field guys had nothing particularly good to say about CAT5.
And they also were very negative about putting ferrites.  But they
***WERE*** correct about the ferrites causing additional errors when
the CAT5 run was "marginal".  I personally observed this myself.  I
will not attempt an explanation, because I don't have one, but AT&T's
particular filter WAS better than the ferrites.  The specs are not out
there anywhere, but I think (don't know) that it's an isolation
transformer that has been somehow broadbanded enough to cover .2 to 8
MHz reasonably.  It's an inventoried item or I would have gotten one
from the guy to tear apart and reverse engineer.

Much improvement was had by removing the approx 100 feet of CAT five
on the input to the gateway and making a place for the equipment
inside the house right next to the external box on the outside wall.
But not enough. The guy thinks that bending and semi crushing the
CAT5 twisted pairs, messes up the twists.  He also said there are good
lots and bad lots.

Frank will do much better with the fiber optic.  Verizon FIOS is not
available here, or I would be on it.

On Fri, Feb 3, 2012 at 11:47 PM, Jim Brown <jim at audiosystemsgroup.com> wrote:
> On 2/3/2012 3:29 PM, Frank Davis wrote:
>> My own mitigation techniques involved placing a number of 2.4" #31 toroids on the power cable for the modem and also the VDSL CAT5 cable that feeds the modem .... I was successful in stopping my line from crashing but only if I limited my ACOM1010 to 500W out on 160/80/40m.
>
> The difference between 500W and 700W is only 1.5dB, so if it's clean at
> 500W, it should not take much to make it clean at 700W!
>
>> The BellAliant technician told me that placing the ferrites on the CAT5 feeding the modem caused  a significant number ( hundreds of '000's on a continual basis.) of FEC (forward error correction) events to begin happening on my line.
>
> That does not make sense unless the CAT5 was mechanically distorted by
> the winding.  That would disturb the impedance at bit, but a LOT of
> errors doesn't make sense to me. The ferrites form a common mode choke,
> which the differential circuit should not see.
>
>>    I was not aware of these errors as they did not present themselves on the TV screens.   The service provider did not like these errors on the line and wanted to eliminate them.  Today the techs visited again and placed a new 3 pair shielded drop wire from the pole to my house.
>
> Was it twisted pairs? Twisting is what matters.  CAT5/6/7 is very good
> twisted pair.
>
>>   After much investigation and trial and error he installed a TII Network Technologies filter/splitter (http://www.tiinetworktechnologies.com/repository/datasheetlibrary/NYMDS086-0411.pdf) at the point where the drop wire enters the wiring closet a few meters from the shack area.
>>
>> He referred to it as a "business filter" meaning it is generally used in business locations were internet access maybe required at a specific workstation and not others.  It splits the line between POTS and Internet .   Only after this filter was placed did the RFI stop and the line would stay up when I run 700W from the amp.  There are still ferrites on the power cord to the modem but not on the CAT5 cable to the modem. The techs explanation was that this robust filter was "blocking" stray RF induced into the house telephone wiring and CAT5 drops coming back from various parts of the house  to the common terminal block in the wiring closet.  Deduced from a couple of hours of isolating various feeds etc.   There is a lot of wire as each room in the house has a CAT5 cable and a telephone drop and RG6 cable.
>
> Good move.
>
> 73, Jim K9YC
> _______________________________________________
> UR RST IS ... ... ..9 QSB QSB - hw? BK