Re: LF: Wilkinson combiners - More

[Andy Talbot <andy.g4jnt@gmail.com>]

Oops, that should have read 

V <= 4.44 .F.N.A.B     (V RMS)

You want to keep B down below some value where it drives teh core into saturation / non-linearity.    But not so many turns that copper losses start to dominate

'jnt

On 11 April 2017 at 13:01, Andy Talbot <andy.g4jnt@gmail.com> wrote:

If your PAs are the push pull type using a transformer, you'll have to keep those, and take their secondary winding to the combiner.   Not ideal but not a lot you can do if you use that sort of design rather than a bridge or half-bridge

I don't use toroids for LF transformers - never have, never would.  I use SMPSU cores - you know where you are with them and are safe in the assumption they are exactly as advertised .   Its a long time since I bought any, but the last ones I used were 3C90 material IIRC, but there's better stuff out there now.

I might use iron dust toroids as inductive elements for filters and DC chokes, but gapped ferrites work as well, and rather more precise in their calculations

As for determining core size and number of turns Just use the equation V >= 4.44 .F.N.A.B (all in Si units) ,     keeping B below around 0.1 or even below 0.07 Tesla and you're there.   Look at SMPSUs to get a feel for the size of core for a given power.

If your individual modules are designed for 50 ohm load, then  your secondary on the combiner needs to deliver the summed power to 50 ohms.  I only ever saw the idea used on the Deccas with their own individual tanks.  You'll just have to try it and see

Just noticed you sent this message off-list ?

Andy

On 11 April 2017 at 11:50, Chris Wilson <dead.fets@gmail.com> wrote:

Hello Andy,


OK, I can see the elegance in that, but pardon the probably daft
question, but at the combining output transformer, is its input taken
from my existing output transformers' secondary windings, or do I
remove both of those and the combining transformer takes their place?
How would I work out what toroid and winding wire to use as a single
amp already causes some warming, albeit I believe it is derived from
the primary wire, rather than the ferrite itself being heated. Thanks.

Monday, April 10, 2017, 7:49:36 PM, you wrote:

> Why not keep it simple, and do it like the Decca transmitters?  
> Forget complicated splitters/combiners with messy deliberate
> isolation.  You are combining identical signals so that sort of design is not required.


> BEFORE any low pass filters, take each output of each transmitter
> module to separate identical primary windings on one, normal ferrite
> transformer core.  Use a single secondary to take the combined power
> off through a low pass filter to your antenna.   (But do size the
> core and turns for the TOTAL power to be generated)


> As both transmitter modules will be delivering identical outputs,
> the voltage induced in primary 2 coil from Tx module 1 will be
> exactly equal to the that delivered to it from Tx  module 2 (and
> vice versa)  so with both transmitters operating correctly, the
> outputs will sum in the secondary.    you are adding teh currents
> delivered by each stage, from an equal voltage delivered by both.  
> You are, to all intents and purposes, connecting the transmitters in
> parallel to multiply the current delivered in total


> You can use the combining transformer for impedance matching if
> wanted - for example the Decca ones  have direct drive to the
> primary windings via a tank circuit and each of the ( In their case
> three) primary to single-secondary turns ration ratio sets the impedance transformation.


> Isolating combiners like Wilkinsons only come into their own
> properly when non-identical signals need to be summed while keeping
> the two sources isolated.    Intermod testing in receivers is a
> particular case in point, where the two test signals have to be
> summed without generating their own intermod products caused by  the
> output of one source leaking into the (probably non-linear) output
> of the other.  If that happened, self-generated IMPS that could swamp those being tested-for.



> Andy  G4JNT






--
Best regards,
 Chris                            mailto:dead_fets@gmail.com