This was the original issue

What I am seeing is each amp giving a nice (well, they seem nice to
me...) drain and gate waveforms individually,

but when combined both

waveforms deteriorate and the cold running FET's and output
transformers heat up quite dramatically

Based round, what appears to be  detuning ??  , what's   missing , is the  filtered sine wave  , then  the  combiner , 

 

the  most I have  seen  for HF is  the Racal  10 module  1Kw HF amp and , 16 in a  ' tree '  uhf wide band  system , using  4:1 combiners

 

73-G,.

 

 

--------------------------------------------------
From: "Alan Melia" <[email protected]>
Sent: Saturday, April 08, 2017 9:30 PM
To: <[email protected]>
Subject: Re: LF: Wilkinson combiners

> John this is a common misconception, these PAs produce something close to a
> sinewave not an RF square wave. I am not sure of Chris's design but Class E
> generally has a THD of about 7%.
> See Prof. Nat Sokal's original QST article. They are however sensitive to
> load reactance changes.
>
> There is another way of combining 2 Class E stages that was devised By Johan
> Bodin, but again I am not sure it applies to this design.
>
> Alan
> G3NYK
>
> ----- Original Message -----
> From: "John Fisher" <[email protected]>
> To: <[email protected]>
> Cc: <[email protected]>
> Sent: Saturday, April 08, 2017 6:58 PM
> Subject: Re: LF: Wilkinson combiners
>
>
>
> Hi Chris,
>
> To me, it doesn't sound like a good idea to combine two switching square
> wave signals together...Might be okay if they were synchronized...Better to
> filter them into sine waves first and then combine them...More expensive for
> double the number of iron powder toroids though...
>
> Jay is combining them ahead of the filter...That type of combiner must work
> okay that way.
>
> Cheers,
> John VA3VVV
>
>
> --------------------------------------------
> On Sat, 4/8/17, Chris Wilson <[email protected]> wrote:
>
> Subject: Re: LF: Wilkinson combiners
> To: "Markus Vester" <[email protected]>
> Received: Saturday, April 8, 2017, 12:42 PM
>
> Hello Markus,
>
>
> Thanks for the detailed reply,
> let me see if i have this straight
> though
> please, sorry if I appear a bit dense, it's probably
> because I
> am.... :)
>
> Right now I run one voltage mode Class D push
> pull amp into a T type
> low pass filter. All
> works great, nice gate and drain wave forms, and
> my Scopematch box shows perfect sinusoidal wave
> forms for voltage and
> current after the T
> type filter. Modest output transformer warming,
> each pair of FET's near as damn it cold.
>
> I add another duplicate amp
> and the Wilkinson combiner. The combiner
> feeds the T type filter. With either one or
> both amps running the gate
> and drain wave
> forms go to pot, the FET's run mad hot, as does the
> output transformer (and the power input choke
> gets hot, too, which it
> doesn't with a
> single amp straight into the filter).
>
> Are you suggesting I build a duplicate T type
> filter bank and put a
> filter between each
> amp and the combiner, with no filter on the output
> of the combiner, just run the antenna co-ax to
> the matching
> transformer and loading coil
> which are outside? I can do this, it
> means
> finding and buying some more big toroids, but it's
> perfectly
> possible....
>
> Is the W1VD combiner what people call a
> "broadband" combiner? And if I
> built this instead of the Wilkinson one would I
> be able to run it in
> front of my single T
> type filter? Or do I still need a pair of filters
> between it and the two amp's outputs? To
> remind, the W1VD one is here:
>
> http://www.w1vd.com/137-500-500WCombiner.pdf
>
> Thanks for your patience,
> and thanks Graham for your input too, just
> don't want to waste time and money
> building things that are incorrect
> due to my
> ignorance!
>
>
>
> > Hi Chris,
>
> > the PI-type combiner per se is not the
> problem. All I was trying to
> > say is
> that you have to prevent it from short-circuiting the
> > harmonics directly at the output of a
> (voltage-mode) class-D PA. In
> > that
> case, I would recommend a T-type low-pass filter between
> each
> > PA and combiner input, starting
> with a series inductor rather than a shunt capacitor.
>
> > Note that the desired
> harmonic load impedance depends on the type
> > of Class-D PA: Voltage-mode class D with
> rectangular voltage and
> > sinusoidal
> current should see a high impedance termination for
> > harmonics, i.e. an inductor or series
> resonant circuit. On the other
> > hand,
> current-mode class D designs would have sinusoidal voltage
> and
> > rectangular current, and require a
> low-impedance shunt for harmonics.
>
> > Best 73,
> > Markus
>
>
>
>
>
> --
> Best regards,
>  Chris
>  mailto:[email protected]
>
>
>
>
>