Received-SPF: neutral (gmail.com: 193.82.116.20 is neither permitted nor denied by best guess record for domain of owner-rsgb_lf_group@blacksheep.org)
From: MarkusVester@aol.com
Message-ID: <c3f.1399636.3228c714@aol.com>
Date: Thu, 31 Aug 2006 19:13:24 EDT
To: rsgb_lf_group@blacksheep.org
MIME-Version: 1.0
Subject: Re: LF: Why?
Content-Type: multipart/alternative; boundary="part1_c3f.1399636.3228c714_boundary"
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org
Status: O


--part1_c3f.1399636.3228c714_boundary
Content-Type: text/plain; charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable

Dear J. and LF,

... this is a very interesting question, exposing the heart of the matter. T=
o=20
my understanding, the presence of this little bypass capacitor decides=20
between two fundamentally different mdes of Class-D operation.

In "voltage-mode class-D" with the center tap RF-grounded, the drain voltage=
s=20
are ideally squarewaves between zero and 2x Vdd, with one transistor looking=
=20
like the mirror image of the other. The drain currents are half sines, just=20
like in class-B. The output filter has to present a high input impedance for=
 all=20
odd harmonics, so a T-type lowpass (or a series resonant circuit) should be=20
used. The transformer itself and the wiring must have very low stray inducta=
nce=20
to suppress high frequency ringing, caused by a common-mode parallel=20
resonance with the FET output capacitances. Stray inductance can be kept low=
 by=20
bifilar winding of the transformer primary. Alternatively, it can be complet=
ely=20
avoided in a single-ended pushpull (halfbridge) configuration, which does no=
t=20
require a balun transformer and shunts all even harmonics directly - this is=
 what=20
I have been using in my 200 W PA for six years now.

Without the bypass at the tap, you enter the "current-mode class-D" regime,=20
where the constant current from the choke is simply switched between the=20
transistors. Thus each FET delivers a squarewave current contribution. The o=
utput=20
filter has to shortcircuit the odd harmonics, as in a Pi or parallel tank=20
configuration. The off-state drain voltages are ideally halfcycle sinusoids,=
 peaking=20
at pi times Vdd. The common mode voltage developed across the choke consists=
=20
of even harmonics and looks like the output of a full wave rectifier. The=20
transformer's stray inductance is not a problem here as it is in series with=
 the=20
choke anyway, so bifilar winding is not needed.

Apart from these basic considerations, perhaps Jim or David could comment on=
=20
their views of the advantages of one mode or the other?

Kind regards,
Markus, DF6NM


In einer eMail vom 31.08.2006 19:47:07 Westeurop=E4ische Sommerzeit schreibt=
=20
vy1ja@northwestel.net:=20

> Thinkers,
>=20
> What are the reasons for not bypassing the center tap on the push-pull amp=
s=20
> apart from high bypass current?
>=20
> There are spikes on the rising edge near the positive and negative peaks i=
n=20
> the waveform of the output from the LPF from the amp I am working on, and=20
> bypassing the center tap to ground cleared up the spikes.  It also cleared=
=20
> up the bypass capacitor  :o).
>=20
> J.=20
>=20
>=20
>=20
>=20


--part1_c3f.1399636.3228c714_boundary
Content-Type: text/html; charset="ISO-8859-1"
Content-Transfer-Encoding: quoted-printable

<HTML><FONT FACE=3Darial,helvetica><HTML><FONT  SIZE=3D2 PTSIZE=3D10 FAMILY=
=3D"SANSSERIF" FACE=3D"Arial" LANG=3D"2">Dear J. and LF,<BR>
<BR>
... this is a very interesting question, exposing the heart of the matter. T=
o my understanding, the presence of this little bypass capacitor decides bet=
ween two fundamentally different mdes of Class-D operation.<BR>
<BR>
In "voltage-mode class-D" with the center tap RF-grounded, the drain voltage=
s are ideally squarewaves between zero and 2x Vdd, with one transistor looki=
ng like the mirror image of the other. The drain currents are half sines, ju=
st like in class-B. The output filter has to present a high input impedance=20=
for all odd harmonics, so a T-type lowpass (or a series resonant circuit) sh=
ould be used. The transformer itself and the wiring must have very low stray=
 inductance to suppress high frequency ringing, caused by a common-mode para=
llel resonance with the FET output capacitances. Stray inductance can be kep=
t low by bifilar winding of the transformer primary. Alternatively, it can b=
e completely avoided in a single-ended pushpull (halfbridge) configuration,=20=
which does not require a balun transformer and shunts all even harmonics dir=
ectly - this is what I have been using in my 200 W PA for six years now.<BR>
<BR>
Without the bypass at the tap, you enter the "current-mode class-D" regime,=20=
where the constant current from the choke is simply switched between the tra=
nsistors. Thus each FET delivers a squarewave current contribution. The outp=
ut filter has to shortcircuit the odd harmonics, as in a Pi or parallel tank=
 configuration. The off-state drain voltages are ideally halfcycle sinusoids=
, peaking at pi times Vdd. The common mode voltage developed across the chok=
e consists of even harmonics and looks like the output of a full wave rectif=
ier. The transformer's stray inductance is not a problem here as it is in se=
ries with the choke anyway, so bifilar winding is not needed.<BR>
<BR>
Apart from these basic considerations, perhaps Jim or David could comment on=
 their views of the advantages of one mode or the other?<BR>
<BR>
Kind regards,<BR>
Markus, DF6NM<BR>
<BR>
<BR>
In einer eMail vom 31.08.2006 19:47:07 Westeurop=E4ische Sommerzeit schreibt=
 vy1ja@northwestel.net: <BR>
<BR>
<BLOCKQUOTE TYPE=3DCITE style=3D"BORDER-LEFT: #0000ff 2px solid; MARGIN-LEFT=
: 5px; MARGIN-RIGHT: 0px; PADDING-LEFT: 5px">Thinkers,<BR>
<BR>
What are the reasons for not bypassing the center tap on the push-pull amps=20=
<BR>
apart from high bypass current?<BR>
<BR>
There are spikes on the rising edge near the positive and negative peaks in=20=
<BR>
the waveform of the output from the LPF from the amp I am working on, and <B=
R>
bypassing the center tap to ground cleared up the spikes.&nbsp; It also clea=
red <BR>
up the bypass capacitor&nbsp; :o).<BR>
<BR>
J. <BR>
<BR>
<BR>
<BR>
</BLOCKQUOTE><BR>
<BR>
</FONT></HTML>
--part1_c3f.1399636.3228c714_boundary--