DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws;
  s=s1024; d=btopenworld.com;
  h=Received:X-Yahoo-SMTP:X-YMail-OSG:X-Yahoo-Newman-Property:Message-ID:From:To:References:In-Reply-To:Subject:Date:MIME-Version:Content-Type:Content-Transfer-Encoding:X-Priority:X-MSMail-Priority:X-Mailer:X-MIMEOLE;
  b=dtgUaZOdKrk9WgniTb6oESWs5H0h7ELEKkJsnmVOq36lc+VQG8mf3xsvWn0Yi32VFU+i3vty7s/A9oTVLPL6KwH+JaMVlzxYhkzCTXUYjTn9wwV0RXJRxdOd4nmbmUfLlwuDvx56MU1jz85U4AqPkX7vdTEv4TpxPyJ+sjwtt1A=  ;
Message-ID: <88FB4290CF234F66933047AD790B3719@JimPC>
From: "James Moritz" <james.moritz@btopenworld.com>
To: <rsgb_lf_group@blacksheep.org>
References: <38A51B74B884D74083D7950AD0DD85E828AC7C@File-Server-HST.hst.e-technik.tu-darmstadt.de>
In-Reply-To: <38A51B74B884D74083D7950AD0DD85E828AC7C@File-Server-HST.hst.e-technik.tu-darmstadt.de>
Date: Wed, 16 Dec 2009 13:31:27 -0000
MIME-Version: 1.0
DomainKey-Status: good (testing) 
Subject: LF: "Gain" between qrss3 and qrss10 / PA design
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


Dear Stefan, LF Group,

DK7FC wrote:
...> Yes, on 137kHz a half bridge is also a good idea since the losses are 
quite small. But, for a half or full bridge like in these decca 
transmitters, one needs an output ferrite core that can handle the whole 
output power. I think these are expensive (?), not easy to get and could be 
driven into saturation when dimensioned well....>

I have used ETD49-size SMPSU cores successfully for transformers at 136k at 
the 1kW level. These are readily available from big component distributors 
in this country such as RS Components and Farnell - the set of core halves, 
bobbin and clips costs about 5 - 10 GB Pounds. I suppose you should add some 
more for wire and insulation, but this is a lot cheaper than a suitably 
rated 50Hz transformer! In my experience, temperature rise due to hysteresis 
and other losses limits the power capability at these relatively high 
frequencies, core saturation is not the limiting factor.

In the overall cost of a QRO TX, the most expensive part is usually the DC 
power supply. So, if you are trying to achieve a low-cost TX design, and you 
already have a high power DC supply, it is sensible to design the TX to make 
optimum use that supply. If you don't already have a PSU, economy is the big 
advantage of a 330V DC off-line rectified supply. The TX output power is a 
function of the DC supply V, the load resistance, and the circuit topology 
used. The DC supply is fixed by the mains voltage if we use a direct 
rectified supply. If, for further economy, we try to design an output stage 
with no output transformer, or other impedance matching, then the load 
impedance is also fixed. So the power output is then fixed for a particular 
circuit configuration. For a push-pull or full-bridge design we really need 
an output transformer to drive a single-ended load, but this leaves the 
Class D half-bridge and Class E configurations, which are single-ended.

Using  the text-book formulae for Pout of an ideal circuit with 330V DC 
input and 50R load gives about 1.2kW for an "optimum" Class E, and 440W for 
the half-bridge Class D. The class E circuit is perhaps simpler, at least on 
paper, and has attractive output power, but further calculations show that a 
switching device handling peak voltage 1200V and peak current about 10.5A is 
required, which would be a problem. We could re-design the class D stage to 
produce 1.2kW by reducing the load impedance to 18.4ohm, which could be 
achieved using an LC matching network, if transformers must be avoided. Then 
the peak current and voltage in the two switching devices would be 330V , 
11.4A, which is a relatively easy requirement. So for this kind of power 
level, and off-line DC supply, the class D design is probably more practical 
. With current switching transistor technology, the 1200V peak voltage is 
the problem for any class E off-line design; but in the future it might not 
be, since there seems to be a lot of interest in developing high speed, high 
voltage switches based on GaN, SiC semiconductors at the moment.

But, in the off-line TX, another important reason to have an output 
transformer is to isolate the output RF ground and antenna system from the 
mains supply, partly for safety - but with a full-wave rectifier, the "0V" 
DC terminal, and so the RF ground of the output stage, is not at ground 
potential anyway. The same also applies to the gate drive input side. So, if 
you don't have a suitable DC PSU, an off-line directly rectified supply is 
attractive, but it will probably need RF transformers! In any case, an 
output transformer does give a lot more flexibility in the design.

Cheers, Jim Moritz
73 de M0BMU