[mythtv-users] Coax splitters - how painful are they?

Brian Wood beww at beww.org
Sun Sep 30 13:59:13 UTC 2007

sean darcy wrote:
> On 9/25/07, Jay R. Ashworth <jra at baylink.com> wrote:
>> On Tue, Sep 25, 2007 at 08:01:41AM -0600, Brian Wood wrote:
>>> A "perfect" two-way splitter would have 3db. loss. Since nothing is
>>> perfect in this world, most splitters actually lose 3.5db. per leg for a
>>> two-way split, and 7db. for a four-way.
>> The Ideal 1GHz 4-port I just installed says 7.4db.
> What about a 4-port that has only 2 [out] ports connected. Still 7db
> on the out ports, or 3db?

> What if the 2 unused ports have 75ohm terminators?

The output of each port would be a nominal 7db. down, no matter how many
ports are used.

Unused ports should always be terminated in 75ohms (so, in effect, all
ports are "used".

One problem you will see if all ports are not terminated is the normally
fairly flat frequency response will become far less so, so if you
measure individual channel levels you might see some that are only 1 or
2 db. down, and other rhat are perhaps 10 or 12 down. This is due to
reflections, standing waves, impedance mis-match and a lot of other bad
things that happen when all the ports are not properly terminated.

If you want a device that is immune to whether all the output ports are
terminated or not you need a directional coupler. This is the type of
device used by cable companies to tap off subscriber drops. Some cable
companies do terminate unused ports, but this is for security purposes
(they use "locking terminators") to prevent unauthorized connections,
not for electrical performance reasons. Of course such devices have
higher losses than hybrid splitters, as the matching components "waste"
some signal.


> I've got a splitter for the eight cable runs in the house to connect
> to Comcast. Only two are actually connected. This thread is making me
> rethink how I  do this.

More information about the mythtv-users mailing list