Gentlemen,
All of the one-port-to-two-port, passive splitters I have seen (to date)
consist of a three-port asymmetrical un-un, using ferrite cores and one
turn of fine magnet wire.
Of course, the theoretical loss from in to out (or out to in, if used as a
combiner) is 3dB, but the additional losses just due to the F connectors
(and the quality of their installation) probably approaches another 3dB as
frequency approaches 1 GHz (Why do you think we use SMA, SMB, N, TNC, C
and even more exotic connectors as frequency increases?)
Other losses (I**2 R, dielectric, hysteresis and stray coupling) may add
another dB or two. This is why there is usually no circuit board (Cheap
stuff is too lossy.), just components supported by wire leads (Ugh).
Cheap couplers are usually unmarked or marked -7dB; the professional CATV
are usually -4.5dB. This makes me suspect that the connector losses are
customarily ignored.
Any instrument that measures loss through a properly terminated coupler as
less than -3dB should be returned for calibration!
Bill Gilmore, Electromagnetic Compatibility Engineer
Core EMC, Electrical/Electronic Engineering
Powertrain Component EMC / Rear Wheel Drive, SRT & HEV Platform EMC
"Michael Tope" <W4EF@dellroy.com>
Sent by: rfi-bounces@contesting.com
02/02/2005 01:44 AM
To
<w3stu@arrl.net>, <rfi@contesting.com>
cc
Subject
Re: [RFI] TV cable splitters
----- Original Message -----
From: "Stu Benner" <w3stu@myactv.net>
To: <rfi@contesting.com>
Sent: Tuesday, February 01, 2005 6:39 PM
Subject: RE: [RFI] TV cable splitters
> Jim,
>
> I agree with Mike that an insertion loss like you measured (3.5 dB) is
> quite high. If you meant a splitting loss of 3 db (which is the nominal
> value for a two-way splitter since power is divided in half) with an
> insertion loss of 0.5 dB, that would be a decent splitter. If the
> splitter was outside, I would suspect that the cause of the failure was
> its physical environment. Moisture ingress and/or thermal cycling being
> most likely. Electrical overstress due to transient differential
> voltages being much less likely, but possible.
>
> Stu
> W3STU
Loss (over and above the 3dB splitting loss) for a good
CATV 2-way power divider (e.g. one that costs $0.50 to
manufacture) is 0.3 to 0.4dB in the low VHF range
increasing to about 1.2 dB at 1 GHz. Port-to-port Isolation is
typically 15 to 25 dB. Construction is typically a zinc die cast
housing with a small epoxy glass circuit board soldered or
screwed into the housing. The board holds a pair of small
ferrite transformers (one for the magic-tee and one for the input
matching), a compensating capacitor, and a dump resistor
(150 ohms). These things are typically made in the Far East
particulary factories in China (Taiwan used to have a lot of
the business, but I suspect that mainland China has grabbed
a lot of it by now). The covers on the better splitters are tin
plated and solder sealed to the tin plated zince die cast
housing. The cheaper units have a steel cover over the board
held in place under the rolled edge of the die cast housing
and a bead of epoxy (the solder sealed units typically have
better RF leakage performance than the epoxy sealed units).
As I recall, the main failure mechanism on these things are
the solder joints. You can imagine the amount of care that
goes into each one when you think that the manufacturer has
to pay someone to wind two ferrite transformers each barely
the size of a pencil eraser, strip and solder these transformers
to a PC board, install a resistor and capacitor, stuff the PC
board into a die cast housing, and attach a cover all for about
$0.50 including margin.
73 de Mike, W4EF..............................................
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
_______________________________________________
RFI mailing list
RFI@contesting.com
http://lists.contesting.com/mailman/listinfo/rfi
|