as Dick, PA0SE, when raising his question, has mentioned my bipolar PA in the
latest version of the LF Experimenter's Handbook using complementary
darlington-pairs and that the emitter arrows in some of the transistor symbols
point the wrong way, I first had to clear with Dick if there really is a failure
in my circuit. I have not seen the latest edition of the LF Experimenter's
Handbook yet, I wanted to wait for the next Ham Radio to have a look at it and
presumably to buy it.
But Dick's scan of the circuit from the LF Experimenter's Handbook showed my
original circuit which indeed contained a failure; Tr4, a BD312, is a PNP
transistor, but erroneously I have used an NPN symbol for Tr4 in the circuit,
and I apologize for this failure. The position of all emitters and the
transistor types employed are all given correctly, however.
Concerning the design of my first LF PA, the main goal has been to generate 100
watts (the power level of the first temporary LF licences for DL in 1997) and
employ as few coils or transformers as possible. The only RF power coils are in
the output low-pass filter. Next, the sum of the power dissipation limits of all
power transistors (each transistor of the BD311-314 series has a Pt of 150 W)
should be at least as high as the total amplifier input (an old rule for me
which is also true for the data of typical RF power transistors and FETs and
which I also strictly obey for my HF PA designs). Furthermore I have not much
experience in power field effect transistors so far and therefore have preferred
to use epibase bipolar transistors with an Ft of about 4-8 MHz because they
would not oscillate on VHF (I have an older 160 m PA with such a transistor
<2N5068 with Ft 4 MHz> showing absolutely no harmonics above 8 MHz in the output
spectrum, the transistor is simply too slow to generate higher harmonics). On
the other hand I have been reported that power FETs in Hi-Fi-amplifiers, even in
circuits already sold on the market, may oscillate around 29 MHz, causing
interference to a nearby radio amateur. These are some of the disadvantages when
employing high frequency devices for low frequency purposes.
Compared to switching PAs the efficiency is rather low, about 50%, but the
amplifier has tolerated serveral times high VSWR or even not being connected to
the antenna without any damage. This reliability is what I want from all my PAs,
be it HF or LF; I am ranking efficiency secondary, I am not so often on the air.
It is my impression that the main problem with this dc coupled type of class B
or C amplifier to achieve higher efficiency is to drive one transistor of the
power pair really into cut-off condition while the other is conducting. This
condition can be met easier in transformer-coupled amplifier circuits.
When designing for higher power output I will not exclude also using FETs in the
future because for a single unit the available dissipation rating is higher for
FETs than for bipolars. But I guess I will have to study the problem of VHF
oscillations under all operating conditions, and add some protective means
against this. The centre tapped coil for instance, loaded by 15 ohms, located
between both power FETs in the Decca 5501 PA output, seems to me to be such a
In a certain way, my complementary bipolar PA is even showing similarities to
this basic Decca power amplifier block mentioned, such as two transistors in a
totem pole configuration, the output employing a series resonant circuit . . .
73 Ha-Jo, DJ1ZB
Dick Rollema schrieb:
To All from PA0SE
Most of the designs for high power transmitters for the LF bands I have seen
use power-FETs in the final. Examples are the designs by G0MRF and G3YXM in
The low frequency experimenter's handbook. But we all know from experience
that FETs are very vulnerable (extremely fast fuses). Would not power
bipolar transistors be more robust? What are the advantages of FETs over
FETs do not require real power to drive them. But for stability and maybe
other reasons rather low value resistors are often connected from gate to
common; this together with the higher voltage to drive a FET than a bipolar
transistor could well mean that the actual power to drive a FET amplifier is
higher than for a bipolar design.
The Handbook features a design by DJ1ZB using complementary darlington-pairs
(the emitter arrows in some of the transistor symbols point the wrong way!).
But they produce 100W only.
I think I have seen higher power bipolar transistor transmitters in Break-In.
Perhaps the ZL boys can comment.
If anyone has evaluated the pros and cons of field effect and bipolar
transistors for high power LF transmitters I would be grateful to hear them.
73, Dick, PA0SE