DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=btinternet.com;
  h=DKIM-Signature:Received:X-Yahoo-SMTP:X-YMail-OSG:X-Yahoo-Newman-Property:Message-ID:From:To:References:Subject:Date:MIME-Version:Content-Type:Content-Transfer-Encoding:X-Priority:X-MSMail-Priority:X-Mailer:X-MimeOLE:X-Antivirus:X-Antivirus-Status;
  b=QFpu/PTG0P37RAe2d221vz7a4Rp+gra8GPIHNHaz34Tg0xQ4WOHj/NH+OLuXg+kT3v+5e42YbFjyB4pbYWBpM6STkDRePuc5qYJpqumu0xlCU7hasNXpHT+Ygq4J899SW94+WCofY0Z9cZ/jmnSRJ+1i98DcpRMcCxn54hnwpx4=  ;
Message-ID: <001c01cbad9d$19f6da70$4001a8c0@lark>
From: "Alan Melia" <alan.melia@btinternet.com>
To: <rsgb_lf_group@blacksheep.org>
References: <AANLkTim3_ub5p7DZfo=ZxSPy6BsX2TjBKsGbu14KedSZ@mail.gmail.com>
Date: Thu, 6 Jan 2011 12:17:53 -0000
MIME-Version: 1.0
DomainKey-Status: good (testing) 
Subject: LF: Re: 500kHz Classe-E
Content-Type: text/plain;
	charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org


Hi Andy, it should be possible. When I was playing I didnt have that much
current available at high volts. Finbar certainly had 550 W with 70v (ish)
power supply but we used two devices in parallel. I believe there may be
possible problems with paralleling. Finbat got a workable unit with two
devices but found three very twitchy ....I suspect parasitic
oscillations...... so we never got to a kW.  We then looked at Johan's
anti-phase drive to two units, this looks capable of giving a kW with two
pairs of devices and a little in hand......but no practcal results at that
level.

Suggestsed tuning a simple variometer (two telescoping coils) you dont want
a lot of variation on L2 and air cored coils are easy  (even at 136). I have
played with an L-Pi step-up on the output but I found the final C really
needed to be "variable" any reflected reactance can take you away from
"optimum" efficiency. The PA only needs tuning once then you tune the aerial
(at least that was my thought).

I used Wima PHE428 caps but RIFA do a similat pulse rated grade. It was
amusing using old HV caps from the valve era....these, whilst suitable
voltage working rating, would not take the RF current and dropped in
capacitance due, I think, to fusing the foil to lead-wire connection, as the
unit was tuned.

I did have an idea for a tuning aid consisting of a high-side DC current
sensor and dividing the output by the sensed RF current (in 50ohm, of
course) this would allow a "normal tune for a peak" rather than a more
difficult assessment of the FET drain-voltage. I was envisaging an
"analogue" divide with the RF current converted to PWM. and chopping the DC
current sensor output.

I would certainly be interested in the "stability".....we never lost any
FETs with basic Class-E stages......maybe lucky but all fails were due to
other reasons that mis-tuning or odd aerial impedances. The trip parallel
did in a few and Mike lost a couple by removing the drive to one of the fet
driver chips (the outputs go permanently high ....whoops)

Good luck with the build....I await the results with interest.

Alan
G3NYK

----- Original Message ----- 
From: "Andy Talbot" <andy.g4jnt@gmail.com>
To: <rsgb_lf_group@blacksheep.org>
Sent: Thursday, January 06, 2011 9:29 AM
Subject: LF: 500kHz Classe-E


> As I have all the bits available as surplus junk on the shelves, thought I
> would put together a QRO class-E  PA based around Alan's spreadsheet,
(with
> mod for matching built in).
> But before comitting to any particular power level / supply voltgage
> combination thought I'd ask what others have reliably achieved with a
> similar setup.
>
> Starting parameters :
>    Supply voltage anything from 28 - 55V adjustable.  25 A max
>    PA device  - IRFP452 (have a small handful, so can allow a few to go
> pop)  Rated 500V,   0.4 Rdson, 14A,  190 Watts
>    Drive, accurate 50% square wave from ICL7667
>    Resonant Inductor -  air wound with 2mm Litz wire.
>
> I reckon that at 55V 500 Watts might be a value to aim for with a single
> device, with the option to back off by lowering Vsupply.
> But before committing to winding inductors and making parallel combination
> of high voltage caps, does anyone have any practical experience?   Am I
> aiming for too high a power perhaps.
>
> Andy
> www.g4jnt.com
>