Message-ID: <000f01c43ec9$9c9251c0$2e868751@rogersservices>
From: "James Moritz" <james.moritz@btopenworld.com>
To: rsgb_lf_group@blacksheep.org
References: <003201c43d13$c22fc500$2102000a@ibm>
Date: Fri, 21 May 2004 01:21:06 +0100
MIME-Version: 1.0
Subject: LF: Re: output transformer problems
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

Dear Dick, LF Group,

I have been playing around with different output circuit arrangements for my
G0MRF TX board - the simplest configuration which gives good results is like
this:-
-The 470nF capacitor C23 is eliminated completely.
-Inductor L4 is increased in value to 56uH (25 turns of 1mm enamelled wire
on T106-26 iron dust core) The value is not at all critical, provided it is
more than about 20uH and can handle the DC current. High Q is not required.
-Capacitors C22 and C24 are each replaced with a series combination of 4.7nF
+ 5ohms (2x 10ohm, 3W in parallel).
-The PA transformer uses an ex-monitor line output transformer core. The
secondary is 16turns of 1mm enamelled wire close-wound in the middle of the
round leg of the core. The primaries are 2x 4t bifilar wound directly over
the secondary. The wire used for the primaries was home-made "litz",
consisting of 3 groups of 4x 0.3mm enamelled copper wires, with additional
insulation of thin heat-shrink sleeving. I have now tried several ex-monitor
LOPT cores for the output transformer; they all work well.

The rest of the circuit is as shown in the original G0MRF design. The drive
circuit and output filter is unchanged. With this arrangement, and DC input
of about 37V, 12A, about 400W is obtained with efficiency something over
90%. The temperature rise was OK after 30 minutes continuous "key down"
operation. As far as the DC input goes, as with switching mode amplifiers in
general, the DC current is proportional to the DC voltage - in this case,
the amplifier behaves like a 3.1 ohm resistor as far as the PSU is
concerned. The power output varies as the square of the DC supply voltage,
down to very low values.

The waveforms in the circuit are less than ideal in comparison with a
"text-book" current-fed class D amplifier. It is possible to improve
efficiency and reduce harmonics somewhat by changing the PA output network
and by further improvements to the output transformer to reduce leakage
inductance. However the practical difference this makes is probably not very
great, unless you are really striving for perfection.

Cheers, Jim Moritz
73 de M0BMU

----- Original Message -----
From: Dick <knaap159@zonnet.nl>
To: <rsgb_lf_group@blacksheep.org>
Sent: Tuesday, May 18, 2004 9:06 PM
Subject: LF: output transformer problems


> Hello all,
>
> due to circumstances I'll have less time to work on the transmitter.
>
> Meanwhile I tried several other transformers, as well as some
> other things, like disconnecting the Scottky diodes and the 470nF
> C's .
> Now I have 6R8 from the TC4426 direct to the gates, which
> are grounded with 47k resistor.
> I tried all kind of things around the FET's with other C's to ground or
> with C-R to ground etc.
> Some progress has been made, but still difficult.
>
> The ouput filter seems to have a quite large influence on the over all
> drain current.
> Also the value of the capacitor that grounds the centre tap is important.
> In my case I found better results at 330nF.
>
> Several available cores tested, like 3C11 (white) and 3E1 (green) both
36mm.
> Also tried an T106-26 from an old computer SMPS.
>
> Wonder what current I should have when applying lower supply voltages
> to the FET's.
> For example. what currents to others have at 15V or 20V or so.
>
> 73, Dick, pa4vhf
>