Message-ID: <B897953B629B4814B8BDF3B6B47A3620@gnat>
From: "Alan Melia" <alan.melia@btinternet.com>
To: <rsgb_lf_group@blacksheep.org>
References: <876597267.20170406161946@gmail.com> <15b453285ab-1d0d-6408@webprd-m02.mail.aol.com> <1267857963.20170407112913@gmail.com> <416F4FA73B194E51B852DC7A287B46D0@Clemens0811>
Date: Sun, 9 Apr 2017 22:03:56 +0100
MIME-Version: 1.0
Subject: Re: LF: Wilkinson combiners
Content-Type: text/plain;
	format=flowed;
	charset="iso-8859-1";
	reply-type=original
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: RO

Hi Chris, MMmm well is is and it isnt :-)) the classic Wikinson I know has 
two quarter wave sections of (I think) 1.5*Zo transmission line with a 2*Zo 
resistor across. Of course a quarter wave of coax at 136 or 475kHz is not 
practical. So the way round this is to increase the "lumped inductance" of 
the coax with ferrite loading (dropping the velocity factor considerably). 
The result is a higher impedance in the coax, and the impedance at the 
common end becomes greater than 50ohms so the transformer is necessary to 
bring the terminal impedance back to 50 ohms again. (I am not completely 
happy with my explanation because the previous delay-line coax I have played 
with Zo=1kohm had the ferrite under the braid).This is an interesting 
design, but unlike the pure "lumped circuit" design you were using it does 
not need large input capacitors, which Markus indicates may be affecting the 
tune of the individual PAs. The filter provide isolation I dont think the 
pure sine-wave is that important. After all your Wilkinson inputs are 
Pi-section  low-pass filters, and they are causing the problem we think.The 
filter format is more important, "T" which is inductor input.
Of course the "lumped component" line.....the Pi format could be directly 
replaced with a T format. It would (I think) require two 82uH coils in 
series (not coupled) with one capacitor at the junction, rather than a 
separate "T" LP-filter (without doing the calculation I am not sure of the 
cap value it may be half what you are currently using on the Pi-section). I 
think this might meet Markus's criterion.

 Transmission line Wilkinson combiners are used as module combiners on most 
of the cellular base station PAs but at these frequencies the transmission 
line is strip-line.

As Andy when we mentioned this today, at LF you can really just use a 
transformer as the old Decca PA decks do.

Alan
G3NYK

----- Original Message ----- 
From: "Clemens Paul" <cpaul@gmx.net>
To: <rsgb_lf_group@blacksheep.org>
Sent: Sunday, April 09, 2017 7:13 PM
Subject: RE: LF: Wilkinson combiners


> Hi Chris,
>
>>Alan: The ferrites I was on about are the ones in the W1VD combiner,
>>which i believe is NOT a Wilkinson type. The one at
>>http://www.w1vd.com/137-500-500WCombiner.pdf
>
> Jay's design is in fact a Wilkinson type,namely a broadband design.
> A good tutorial on broadband Wilkinson dividers/combiners can be found 
> here:
> http://www.minicircuits.com/app/AN10-006.pdf
> T1 is made as classical Faraday transformer,for T2 Jay uses transmisssion 
> line sections
> loaded with ferrites while Minicircuits uses toroids for both transformers 
> for their low power combiners.
>
>>I am using two very
>>big wire wound ceramic resistors of 50 Ohm each in series for "R"
>>which look very very similar to those Steve VE7SL is using in the
>>photo he publishes on this page.
>
> You might check this resistors with you AIM and see how much undesirable 
> reactance they show
> on your operating band(s).
>
> 73
> Clemens
> DL4RAJ
>
>
>>-----Original Message-----
>>From: owner-rsgb_lf_group@blacksheep.org
>>[mailto:owner-rsgb_lf_group@blacksheep.org] On Behalf Of Chris Wilson
>>Sent: Friday, April 07, 2017 12:29 PM
>>To: Markus Vester
>>Subject: Re: LF: Wilkinson combiners
>>
>
>