Envelope-to: dave@picks.force9.co.uk
Delivery-date: Mon, 03 Oct 2005 23:10:49 +0100
Message-ID: <000701c5c867$3f480b80$cfcc8351@Main>
From: "Alan Melia" <Alan.Melia@btinternet.com>
To: rsgb_lf_group@blacksheep.org
References: <20051003181417.2795.qmail@web25415.mail.ukl.yahoo.com>
Date: Mon, 3 Oct 2005 23:10:14 +0100
MIME-Version: 1.0
Subject: Re: LF: Advice (again) please
Content-Type: text/plain; charset=iso-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org

Hi Brian, congratulations you have discovered Class-E. It is possible to
"mis-tune" the aerial so that the correct reactance is reflected back to the
PA to satisfy the conditions for Class-E.

Finbar EI0CF's early single ended PA ( LF Sourcebook ) stumbled on this and
since my initial experiements, I have found a book by Fritz Raab that
enumerates several different configurations of "switching" PA, that may be
"tuned into Class E" The salient point is the shape of the drain waveform.
The "switch" must switch on with no voltage across it, and must switch off
with no current flowing through it. These are the conditions equating to
minimum efficiency loss. The Nat Sokal QST article cover the conditions for
single ended stages. I think Fritz collaborated with Nat at one stage, and I
believe a lot of the ideas he produced are used in mobile phone PAs. The
condition may not lead to maximum "smoke" for a particular supply voltage,
but the PA will run cooler. In a class E stage the output circuit usually
finishes up being tuned around 10% low ...I think this means that it offers
the PA a slightly inductive load. I have an electronic copy of Nat's QST
article if you have not seen it (with the errors in the original article
corrected !)

I also desined a step-up L-Pi network ( circuit on my web site ) and found I
could get the same effect by tuning the output cap of that....that had to be
by switcing caps though as I used iron dust toroids for the coils, though
the values could be realised with air cored coils allowing a slight
adjustment my just offering a ferrite to the end. I think the trick may be
to avoid inserting the ferrite too deeply into the coil when losses may
escalate. I dont know whether they still have any but Mainline Surplus (web
site ) had some ferrite rods about 1inch in diameter and 7 inches long.
These had slots moulded in and it was suggested that these were to reduce
eddy currents when tuning (RF) dielectric heaters. If so these might be
quite good in this service.

Cheers de Alan G3NYK

----- Original Message -----
From: "brian hodgson" <g3ykb@yahoo.co.uk>
To: <rsgb_lf_group@blacksheep.org>
Sent: 03 October 2005 19:14
Subject: Re: LF: Advice (again) please


> Hi Chris -
>
> I've been dabbling with some PA circuits recently - the usual hard
switching, push-pull fets and transformer output follwed by a pi (C-L-C)
Lowpass filter.
>
> One trick I did discover was that by increasing the inductance of the low
pass filter (using the tried and tested technique of  poking a large ferrite
rod inside the coil) from 120 to 185 microhenries the voltage waveform on
the drains became a lot smoother - more like class E - and the overall
efficiency went up to better than 90% (over 200 watts out for 220 watts
input).
>
> Adjustment was quite sharp but not "critical". Watching the drain waveform
voltage on the scope was the best way to tweak but just tuning for minimum
supply current seemed almost as good.
>
> Brian
>
> "OSBORN, Chris" <Chris.OSBORN@astrium.eads.net> wrote:
> I'm generating 100 watts into my 50 ohm dummy load but
> only getting an efficiency of about 50%.
> The output transformer toroidal core is getting quite hot and I
> suspect that it's permeability is too low as the capacitance
> required
> to bring it to resonance is far higher than those of published
> designs
> in the LF Experimenter's handbook, using the same number of turns.
> These capacitors are running very hot too (as mentioned by PA0SE in
> his LFEH article).
>
>
> ---------------------------------
> Yahoo! Messenger  NEW - crystal clear PC to PC calling worldwide with
voicemail